When a large company borrows a billion dollars, it pays a ratings agency like Moody’s for a letter grade on how likely it is to repay the money.

That grade matters because institutions depend on it to do business. For example, a pension fund may only be allowed to buy bonds above a certain rating.

Moody’s charges for the grade. Moody’s doesn’t sell research. It sells permission to do business.

Subscribe now

The moat is written into other people’s documents

Most moats are things a company owns — a better product, a bigger network, lower costs. A competitor can attack each by building something better.

Moody’s has a rarer kind of moat.

About ten firms are registered with the SEC as Nationally Recognised Statistical Rating Organisations, a designation that takes years of audited credibility to earn. In practice, the market is dominated by a Big Three: Moody’s, S&P Global and Fitch.

What makes this interesting is that the dependency has already been tested.

After the financial crisis, Congress told federal agencies to remove references to credit ratings from their rules and replace them with their own standards. They did. The Federal Reserve alone found 46 references, most in bank capital requirements.

The regulatory references changed. The commercial franchise didn't.

Fifteen years later, the same three firms still rate essentially everything that gets rated.

That tells you where the dependency actually lives. It was never one statute that could simply be repealed. It lives in insurance capital regimes, pension-fund mandates, index eligibility rules, and thousands of private contracts and internal policies.

In many cases, the relevant documents don't simply require independent credit analysis; they specify ratings from recognised agencies or use rating thresholds as part of the eligibility framework.

That creates an unusual kind of moat.

A product moat has to keep being better than the competition. Moody’s doesn’t necessarily have to be better. To displace it, someone has to change the rules, contracts and policies that other people rely on.

The moat isn’t in Moody’s documents. It’s written into everyone else’s.

That is the kind of moat worth learning to recognise: the one defended not by what the company owns, but by the cost of changing what everyone else has already written down.

Once you see the moat, the economics stop being surprising

The borrower pays, not the investor. The company being judged buys its own rating because it is the one that needs admission to the market.

And it buys that admission against a billion-dollar financing. A fee measured in basis points barely registers beside the cost of not being rated at all.

That is what pricing power can look like when the fee is tiny relative to the financing it unlocks.

It is also the model’s obvious weakness: the payer chooses the judge. In 2008, we saw what that can look like at its worst.

Buffett has owned Moody’s since it was spun out of Dun & Bradstreet in 2000. In 2010, the Financial Crisis Inquiry Commission subpoenaed him to explain why.

He had declined three invitations, doubting his testimony would be of much value. The commission disagreed and put the ratings agency’s largest shareholder under oath.

His answer went to the heart of the investment case:

“The single most important decision in evaluating a business is pricing power.”

A company that can raise prices without losing customers is a very good business, he said. A company that needs a prayer meeting before adding a tenth of a cent is a terrible one.

He was equally blunt about what he hadn’t needed to know. He had never visited Moody’s and didn’t even know where its offices were.

He wasn’t buying the managers.

He was buying the position.

The economics follow.

No factories. No inventory. The raw material is analysts, data and reputation, so very little additional capital is required to support each dollar of growth.

And demand doesn’t disappear when a bond is issued. Debt gets refinanced. Every maturity creates another opportunity for a fresh rating, before you even count whatever the world borrows next.

It is a toll on a road that keeps being rebuilt.

Last quarter shows what that model looks like in the accounts:

• Margin: adjusted operating margin of 55.3% on $2.19bn of revenue, up 15%.

• Earnings: adjusted EPS of $4.68, up 31%; full-year guidance of $16.50–$17.00.

• Durability: analytics recurring revenue of $3.66bn, up 9%, with 95% of customers renewing.

The point isn’t that Moody’s has found a clever way to make money from credit ratings.

The point is that once the market has decided a rating is the price of admission, the economics of selling that admission become extraordinarily good.

Moody’s (MCO): share price vs. my estimate of fair value, 2021–2026. The Conviction Play.

The flaw that is actually the feature

Here is where owning a business like Moody’s gets interesting on paper.

It needs so little capital that it can return most of what it earns to shareholders, primarily through share buybacks.

But there is an accounting oddity. When a company repurchases shares far above book value, it spends more cash than the accounting value of the equity it retires. Keep doing that for long enough and book value per share can actually go backwards.

That looks like a problem.

For a business like Moody’s, it is mostly an accounting problem.

A balance sheet records what a company’s assets cost. It has no line for decades of accumulated credibility, or for a name embedded in thousands of contracts and institutional policies.

Those are real economic assets. They simply aren’t recorded on the balance sheet.

For a business whose value sits largely in those unrecorded assets, book value doesn’t just understate the economics.

Eventually, it stops telling you anything useful.

What does reach the accounts is the earning power those assets produce. Over the same period, earnings per share compounded at 23.8% and free cash flow per share at 28.7%.

The shrinking book value isn’t evidence that the business is destroying value.

It is the accounting shadow of the mechanism creating it.

What the screen dislikes is, in part, exactly what Buffett bought it for.

So why has it fallen — and is the fear real?

The market has been marking down financial-data businesses on fears that AI will make parts of their information and analytics increasingly commoditised. Moody’s fell from a 52-week high of $546.88 to a low of $402.28 after S&P Global’s weak 2026 forecast in February, with Goldman Sachs highlighting AI disruption and rising costs as key concerns.

The fear is straightforward: if a language model can read a filing, analyse the numbers and answer the question, why pay for a platform that does the same thing?

It’s a good question. But it matters which business you are talking about.

For Analytics, the concern is real. It sells data, models and software, and some of that genuinely competes with capabilities that increasingly cheap AI can replicate. That is consistent with the recent numbers: Analytics grew 4% last quarter, versus 25% for Ratings. If AI-driven commoditisation accelerates, this is the part of Moody’s to watch.

Ratings is different.

It isn’t primarily a data business, and its moat isn’t technological. A language model can form a perfectly sensible opinion about whether a bond is creditworthy. What it cannot do is satisfy the mandate, because the mandate specifies the rating agency, not the quality of the analysis.

Better software doesn’t rewrite a contract.

And if Dodd-Frank couldn’t dislodge the market’s dependence on the major agencies, it is difficult to see how a cheaper analyst does. Morgan Stanley estimates that more than 70% of these firms’ profits sit in areas largely insulated from AI.

There is an irony here.

In July, Moody’s published research warning that the roughly $785 billion being spent on AI infrastructure this year — potentially reaching $1 trillion — is putting pressure on the credit quality of Microsoft, Amazon, Alphabet, Meta and Oracle.

Much of that debt still needs a rating.

Moody’s is being sold off on the same boom that is sending it more work.

The bear case

The strongest case against Moody’s isn’t AI.

It’s that some of the earnings behind today’s price may have been pulled forward — while the company is simultaneously buying back stock aggressively.

Management said activity was pulled forward into the second quarter without changing its full-year outlook. That means some of the 25% growth in Ratings came out of the second half, leaving investors with less visibility on the underlying run-rate.

More importantly, in the same release that raised buyback guidance to as much as $3.0 billion, Moody’s lowered free-cash-flow guidance to $2.7–$2.9 billion.

Repurchases are therefore now guided at or above the entire year’s cash-flow range, before the dividend.

That doesn’t make the buyback wrong. But it does sharpen the accounting question from the previous section.

Shrinking book value has to be explained by growing earning power. It can’t simply be excused by it.

And there is one final problem: valuation.

The discount in the share price is a discount to an estimate of intrinsic value. My fair-value line assumes the market continues to pay a premium multiple for Moody’s earnings. Value it instead as a financial cyclical, and the number falls to $422.00 — roughly 13% below today’s price.

That is the honest downside.

What would have to happen for me to be wrong — and when we’d know.

Great business, right price?

Admiring a business and acting on it are different disciplines, and the second is mostly arithmetic about the price you pay.

The multiple has already compressed: $484.96 buys about 31× trailing earnings, well below the peak. The question isn’t whether the multiple can compress again, but whether 31× is still too high. For that to happen in a lasting way, the ratings franchise would have to crack — or the market would have to decide that its economics deserve to be valued very differently. Neither looks likely, and neither is what the market is primarily worried about.

The shares sit about 14% below my fair-value line of $563.38 — a discount I read as mostly fear about Analytics, though the pull-forward is part of it too. It’s a discount, not a dislocation, and nothing about it makes today special.

At $485, Moody’s doesn’t need to become cheap for the valuation to remain interesting. The question is whether it can become cheaper without becoming worse.

— Harry, Independent Investment Researcher & Writer The Conviction Play

Part of The Conviction Play — Competitive Advantage series, on the structural barriers that protect a business’s economics for a decade.

A note on timing: this is an assessment of what is known today. Company reports bring genuinely new information, and a view that holds this quarter may not hold the next. Nothing here updates itself.

Disclosure: I hold a position in Moody’s Corporation (MCO), which I am building in stages, and may add to or reduce it without notice. This is personal research and opinion, not advice; it does not constitute investment advice. Capital is at risk and the value of equities can go down as well as up. Readers should do their own research or consult a licensed financial adviser before making any decision.

Subscribe now

The Conviction Play - Capital Allocation Series
In this article you’ll learn: why an investment cycle can make a healthy business look like a deteriorating one; how to separate demand risk from execution risk — and why execution is the more underwritable of the two; and the single question that keeps Oracle from being an unequivocal yes.

Every major investment cycle creates the same illusion. The spending appears first; the earnings arrive later.

Build a data centre today and depreciation begins almost immediately — the concrete, steel, power and GPUs all flow through this quarter’s accounts. Fill that same facility with long-term customer workloads over the following years and the economics emerge slowly. Accounting captures both faithfully. Markets often do not: they see free cash flow falling before they see utilisation rising, capital expenditure before operating leverage, today’s cost far more clearly than tomorrow’s return.

Amazon lived through exactly this while building its fulfilment network — years of apparently wasteful spending that turned out to be the most valuable capital it ever deployed.

Today, Oracle appears to be entering its own version of that transition. The question is not whether Oracle is spending extraordinary sums. It is. The question is whether the market has correctly underwritten the economics of that investment.

And the surprising conclusion from Oracle’s latest results is that demand no longer appears to be the primary uncertainty. Execution does. That is a very different — and far more tractable — investment problem.

And it is more tractable for a specific reason. Demand is uncertain because customers can choose not to buy. Execution is uncertain because management must deliver what customers have already bought. Those are very different risks, and the second is far more underwritable than the first.

Put plainly: the question is not whether AI demand exists. It is whether Oracle can convert demand it has already contracted into high-return, per-share value — before leverage and dilution overwhelm the economics. Is Oracle funding demand, or funding hope?

The thesis

Oracle is no longer trying to prove demand for AI infrastructure. It is trying to build infrastructure quickly enough to satisfy demand it has already contracted.

The prevailing narrative — exploding capex, weakening free cash flow, a leveraged balance sheet — is entirely legitimate. But the same evidence carries a second reading: that today’s investment is not a speculative bet on future demand, but the physical build-out required to serve demand that already exists. If that reading is right, the market is pricing the visible cost of the build far more heavily than the contracted earnings power it is meant to produce.

That possibility — not certainty — is where the opportunity lies.

Our framework asks five questions of any business before price even matters. Oracle passes four of them cleanly. In the end, this piece is about the fifth.

Oracle (ORCL): price vs. fair-value line, 2021–present.

Why Oracle?

Oracle is not an obvious AI winner, which is precisely why it is interesting. Unlike the many companies now bolting “AI” onto investor decks, it starts from one of the strongest enterprise franchises in the world: databases, applications and mission-critical systems embedded in the largest companies on earth over decades, with switching costs to match. Its ambition is not to become an AI start-up. It is to turn that franchise into a leading position in AI infrastructure.

The latest results suggest the transition is already underway. Total revenue grew 21% to $19.2 billion; cloud infrastructure grew 93%. And management guides for total growth to accelerate again next quarter, to 27–29% — an inflection that maps directly onto the datacentre capacity coming online right now. These are not the numbers of a mature software company fighting to stay relevant.

What the market believes

At its core, the market increasingly believes one thing: that Oracle is transitioning from a high-return software business into a lower-return infrastructure utility. Everything else in the bear case is evidence for that single claim. Capex has surged; free cash flow has turned deeply negative; debt has risen; interest coverage is now far less comfortable than its software peers’. On this view Oracle no longer deserves a premium software multiple — it is becoming a capital-intensive provider with lower returns, more leverage, and execution risk that may run for years. If that is right, today’s valuation is not a discount at all. It simply reflects a different, lower-quality business.

What the market may be underweighting

Reading the latest earnings call left a different impression. The striking thing was not what management emphasised — it was what they barely discussed. Almost no time went to convincing investors that demand exists. The entire conversation was power availability, construction schedules, megawatts and delivery timelines. Analysts kept asking how fast Oracle could build. Almost nobody asked whether the customers were there.

Businesses struggling to create demand rarely spend an earnings call explaining how quickly they can connect another gigawatt to the power grid.

And the operational detail proves the point mathematically. Oracle’s global GPU utilisation rate is 97.5%. When 35,000 GPUs across 59 customers came up for renewal, customers renewed 92% of them — and the small remainder was resold to other customers within the same quarter. You do not resell capacity inside a quarter in a market where demand is soft. Oracle delivered more than 1.2 gigawatts of capacity last year and expects to bring nearly another full gigawatt online in a single quarter — roughly the prior four quarters combined. The binding constraint is not demand. It is how fast Oracle can build.

So why hasn’t the market simply agreed?

If the transcript is this clear, the fair question is why the shares haven’t already re-rated. Sophisticated investors have read the same call. Three reasons are worth taking seriously. Free cash flow is still deeply negative — and negative cash flow is a fact today, while future returns are a forecast. The balance sheet genuinely worries people; leverage funding a technology build is a combination that has ended badly before. And AI-infrastructure economics remain unproven at scale — the market is being asked to believe a returns profile it has not yet been shown. In other words, the market is doing exactly what it should: demanding evidence of returns rather than promises of them. The opportunity, if there is one, exists precisely because that evidence has not yet arrived.

The structural evidence

Beyond utilisation, three things anchor the constructive case.

The backlog. Remaining Performance Obligations reached $638 billion, up around 360% year on year — contracted, not hoped-for, revenue. It is honest to note it is back-end-loaded: management expects only about 12% to convert within twelve months and another 34% within three years. The visibility is real; the conversion is a multi-year execution task, not a switch.

Who funds the build. A growing share is paid for by customers, not Oracle. Around $75 billion of contracts are now prepaid or built on customer-supplied hardware — at, management insists, the same or better margins. That attacks the scariest version of the bear case, in which Oracle finances the entire capital stack itself. It doesn’t. It still finances most of it — roughly 88% of the backlog is not customer-funded — but “most” is a materially safer proposition than “all.”

The quality of the earnings. The fair-value work here is done on an operating basis that strips out the one-off investment gains Oracle booked this year from stakes in other companies — the conservative read, not the flattering one.

Underwriting the business

Every company here goes through the same five questions — not a screen, but the underwriting tests that decide whether a business deserves capital at all, before price enters the conversation. Oracle clears four of them cleanly. It is the fifth that stops this being a simple story.

Does it have a real return engine? Yes — the legacy franchise has thrown off cash for decades, and management now puts steady-state return on invested capital for large infrastructure projects “in the high 20s.” That would be excellent. But it is management’s own back-of-envelope figure, on a generous definition, at a steady state not yet reached — a claim to test, not a fact to bank.

Does reinvestment compound? This is the heart of the case: capital poured into contracted demand at a high return compounds hard. The demand evidence is strong; the returns evidence is still arriving.

Is governance acceptable? Founder-controlled, two CEOs, a new CFO — who opened her first call by naming “disciplined capital allocation and preserving our investment-grade credit rating” as priorities. The right words from the person who has to raise the money. Adequate, worth watching.

Is there a credible path to per-share value? Credible, yes — earnings are growing near 19% and accelerating. On its own, this one passes.

Can it fund itself without chronic dilution? Here, and only here, the answer is no — and this is the single question keeping Oracle from being an unequivocal yes. Oracle cannot fund this build from its own cash flow. It plans to raise roughly $40 billion of debt and equity in the year ahead, including an at-the-market programme of around $20 billion that issues new shares straight into the market. And consider what that reverses: Oracle spent the last decade as one of technology’s most aggressive repurchasers of its own stock, steadily shrinking its share count. That era has paused. Capital that once bought shares back now pours into concrete and silicon, and the ATM issues stock on top. Every dollar may be well spent — but leverage and dilution both work against per-share value, and the returns are now leaning on the balance sheet rather than on internal cash. That is the watch, and it is the honest reason to be patient rather than emphatic.

Four questions pass. One is on watch. An underwriter’s job is to say which — out loud.

The bear case, taken seriously

The balance sheet is the real soft spot. Net debt-to-equity sits near 2.9×, interest coverage around 4.7× — fine for a software company, thin for one about to add $40 billion of funding. Management guides net cash capital spending of around $70 billion next year; reported capex is higher still, near $90 billion once customer prepayments and timing items are added. (The two numbers are worth keeping straight: ~$70bn is the cash Oracle itself lays out; the gap to ~$90bn is largely customer-funded.) Free cash flow stays negative through the build, and gross margin has already stepped down about five points as half-filled data centres carry cost before revenue.

Valuation is the subtler trap. The shares look inexpensive against Oracle’s own history only because the market is pricing in the earnings the build is supposed to produce. Value the company on what it earns today and it is not obviously cheap. The fair-value anchor leans on a premium multiple that may be too generous for a business now growing on debt and capex rather than internal cash; sober that multiple and today’s modest discount disappears. And a handful of counterparties carry a large share of the backlog — four customers each contracted for more than $8 billion this quarter.

At full strength the bear case says: a more leveraged, more capital-intensive, lower-quality Oracle, priced for a future it must deliver almost flawlessly. That deserves respect.

What would change my mind

Underwriting means naming, in advance, the evidence that would break the thesis. Four things would. If revenue growth falls below roughly 10% for two consecutive quarters, the demand story is broken, not paused. If interest coverage drifts toward 3×, the leverage is biting before the returns arrive. If GPU utilisation falls meaningfully from 97.5%, or renewals soften, the “capacity, not demand” thesis fails its central test. And if gross margin does not begin recovering as data centres fill, the unit economics aren’t what management claims — the single most important number to watch. Loss of the investment-grade rating would end the thesis outright.

Two dates put these to the test: the next earnings print on 10 September, and Oracle’s Investor Day on 28 October, where the updated long-term targets and the promised return economics will either firm up or fray.

Investment committee summary

The bull case: contracted, customer-validated demand at 97.5% utilisation, a $638 billion backlog, and a rising share of customer-funded capital at undiminished margins.

The bear case: a leveraged, cash-consuming build needing ~$40 billion of fresh funding, a still-falling margin, and a valuation that is cheap only if the promised returns arrive.

The swing factor is not demand — the evidence there is strong. It is execution and the balance sheet: whether Oracle can convert a back-end-loaded backlog into ramping, high-return revenue faster than leverage and dilution erode the per-share prize.

Implementation

A thesis this conditional does not deserve a full position on day one, and it isn’t getting one. When the same evidence supports “cheap” and “fairly valued” depending on an execution you cannot yet verify, the disciplined response is to scale in, not lump in: deploy in tranches, let price and evidence earn each addition, and keep enough powder that a lower price is an opportunity rather than a regret. It accepts that being early and being wrong look identical for a while, and it sizes accordingly.

I hold a position in Oracle and am building it deliberately, in stages — exposed enough to participate if the returns prove real, patient enough to benefit if the bear case plays out first. I would rather the market hand me a better price than chase an oversold bounce for one.

The five structural tests — four clear, one on watch:

Verdict

This is an underwritable position, not a table-pounding one. The demand is real enough to underwrite; the balance sheet is stretched enough to demand patience. So I own a starter position and add on Oracle’s terms, not the market’s mood — governed by the falsification tests above, not by a price target.

Investing often feels like forecasting the future. It is more often a test of how well you read the present. Oracle is spending extraordinary sums of money; that much everyone agrees on. The real question is whether that spending is deteriorating economics or the temporary accounting footprint of future earnings power. Markets will answer that eventually. Until then, the investor’s advantage lies not in predicting the future, but in correctly interpreting the present.

— Harry The Conviction Play

This article is part of The Conviction Play — Capital Allocation series, on businesses where the central question is whether management earns the right to reinvest.

Disclosure: I hold a position in Oracle (ORCL) and am building it in stages. I may add to or reduce it without notice. This is personal research and opinion, not investment advice. Capital is at risk. Please consult a qualified financial adviser before making any decision.

Conviction Play theme — Algorithmic Decision Infrastructure, within the Automation & machine-intelligence diffusion megatrend.

In this article you’ll learn the five questions my framework asks before it commits capital; how Netflix scores against each; and why the market’s caution is the opportunity here, not the warning.

Cheap stocks are everywhere. Great businesses trading below what they are worth are not. The entire art is telling the two apart — and it is where most value investing comes unstuck, because the market is usually cheap about something for a reason.

Netflix is the rare case where the reason looks wrong. Everyone knows the name; that was never the question. What the market has done is strip away most of the growth premium it once paid — increasingly pricing the world’s largest streamer as a maturing media business rather than the high-growth platform it was. The share price has fallen in line with that verdict. The economics have refused to cooperate.

The thesis in one sentence

Netflix is a structurally exceptional business that is still compounding, which the market has begun to price as though its best days are behind it — and the gap between that fading narrative and the improving numbers is the opportunity.

I don’t buy discounts — I buy businesses that pass five questions

A low price is the last thing my framework checks, not the first. Before capital is committed, a business has to clear five questions, and only when they line up does the price even come into it:

• Is the business structurally exceptional?

• Is its compounding durable?

• Are the earnings real cash, or accounting?

• Is the balance sheet resilient enough to survive being wrong?

• And only then — am I paying a price that leaves room for error?

Cheap-and-bad fails the first four and is a trap. Cheap-and-great passes them and is an opportunity. That distinction is the whole job. Scored across all five, Netflix grades out at an A — passing all five, with two of them (the balance sheet and capital allocation) good rather than great.

Netflix, question by question

It is structurally exceptional — and still growing. Global scale (well over 300 million households) funds a content budget no rival can match, which wins the subscribers that fund the next budget: a flywheel a decade in the making and genuinely hard to copy. Layered on top are three newer engines — an advertising tier turning a low price point into a fresh margin lever, the monetisation of shared passwords, and a push into live events that deepens engagement. Revenue is still growing about 16% a year and operating margins sit above 30%. That is pricing power, not a mature plateau.

The earnings are actually cash. The old knock on Netflix — profits that never became money because content ate every dollar — is gone. Free cash flow now runs ahead of reported profit, and the business yields close to 8% on a free-cash basis. This is the single most important quality test — are the profits real, or an accounting story? — and here they are real.

The balance sheet is resilient, if not a fortress. Debt is modest against the cash the business throws off: net borrowings are a small fraction of equity and less than a year’s free cash flow, with interest covered many times over. It is not debt-free, and cash sits a shade below long-term borrowings — so I’d call it solidly resilient rather than impregnable. Enough to survive a bad year; not itself a source of edge.

Capital allocation is maturing in the right direction. For years Netflix only knew how to spend. Now it generates surplus cash and has begun returning it — buying back stock while still reinvesting in content at high returns. It is early: no dividend, the buyback young. So I grade this good rather than great. But the direction is exactly what you want from a business crossing into cash generation, with each remaining share quietly owning a little more of the company.

And only now, the price. With four questions answered, price becomes the question — and Netflix is cheap on nearly every lens that matters. It trades around 20 times forward earnings, under 13 times free cash flow, and on a PEG below 0.7 — cheap on the measures the market itself sets. My own earnings-based framework points the same way, implying meaningful upside, though that estimate leans on a judgment about the right long-term multiple. Not one cheap number, but a stack of them pointing together.

One honest caveat, and only one. Reported trailing earnings were flattered by a non-recurring gain in the most recent quarter, which makes the earnings-based discount look a little wider than the underlying economics justify. Adjust for it and the discount narrows — but it does not disappear, and it barely touches the cash-flow lenses, which tell the same cheap story. The thesis was never built on one unusually strong quarter; it rests on Netflix compounding earnings and free cash flow over many years. The one-off nudges my conviction at the margin. It does not change the case.

Why the market is cautious

The market is not being irrational — it is being cautious for reasons worth taking seriously. Subscriber growth in wealthy markets is maturing, so more of the load now falls on price rises and the young ad tier. And the multiple has compressed hard: a business the market once paid nearly 40 times earnings for now changes hands nearer 20 times forward. That compression is the de-rating in plain numbers — the market quietly deciding to stop paying for much of Netflix’s future growth. If that re-rating is permanent — if Netflix is simply a mature media company now — then the historical anchor flatters the discount and fair value sits lower. The tape agrees for the moment: the shares are in a downtrend, under distribution and well off their highs.

That is the real bear case, and it deserves respect. But notice what it is: an argument about the pace of growth and the multiple, not about the quality of the business. No one is arguing the moat has cracked.

Why the opportunity exists

Here is the crux. The market has priced the conclusion — growth is over — while the evidence keeps improving: margins expanding, cash building, advertising scaling, buybacks starting. It is treating a business that is still getting better as though it has stopped. When the story and the numbers diverge like this, the numbers tend to win in the end.

Key questions

• Does the ad tier become a real profit engine? It is growing fast off a small base; whether it turns into a genuine second pillar decides how much growth is left.

• Does pricing power hold as subscriber growth matures? The thesis leans on raising prices in near-saturated markets without shedding members.

• Which multiple is the right anchor — the old ~39× or today’s ~20× forward? The size of the discount turns on this one question.

• Do the operating KPIs hold? Operating margin in the high-20s, subscriber additions, ad-tier pricing — the metrics that confirm the compounding is intact.

• Does free cash flow keep funding buybacks through the content cycle? The maturing capital-allocation story depends on it.

Answer these in Netflix’s favour and today’s price is a gift; answer them the bear’s way and it is a fair price for a slowing business. The next earnings print, due later this month, is the first clean read.

Great business, right price?

So — a great business, at what looks like the right price. That is the rare combination my framework exists to find, and Netflix clears the bar on the things that matter most: an exceptional franchise, real cash earnings, a sound balance sheet, and a genuine margin of safety. The only brake is the tape. With the shares still under distribution, the disciplined way to act on a view like this isn’t to swing at it all at once but to scale in as the turn confirms — patience on the timing, conviction on the business.

The screen flags it; the framework decides it. And on the framework, Netflix looks like opportunity, not trap.

— Harry, The Conviction Play

Disclosure: I do not currently hold a position in Netflix (NASDAQ: NFLX). The author may buy or sell securities mentioned without notice. This article reflects my personal views and the output of my own valuation system; it does not constitute investment advice. All investments carry risk, and the value of equities can go down as well as up. Readers should do their own research or consult a licensed financial adviser before making any decision.

If you enjoyed this, you can subscribe for future pieces on long-term investing, market mispricings, and capital allocation — written from a calm, conviction-driven lens.

June 2026, Conviction Play - Capital Formation Series

In this article, you'll learn what Uzbekistan's National Investment Fund is, why Franklin Templeton was chosen to manage it, how the fund aims to create value through privatisation, and whether the investment opportunity still exists after the discount to NAV has largely disappeared.

If you followed the Fondul Proprietatea story—a Romanian fund that owned stakes in state-owned utilities, a telecoms operator and an airport, all managed by Franklin Templeton—you watched something that looked like a curiosity compound into a multi-billion-dollar investment vehicle while delivering returns that outperformed much of the European equity market.

Franklin Templeton is attempting something remarkably similar.

This time the country is Uzbekistan, the vehicle is the National Investment Fund of Uzbekistan (LSE: UZNF), and the original discount that attracted investors at IPO has largely disappeared.

The question today is no longer whether the discount will close.

It’s whether Franklin Templeton can execute.

Are you willing to trust a manager’s estimate of what thirteen unlisted companies in Central Asia are worth before the market has had a chance to agree—or disagree?

Thesis: UzNIF is no longer a simple discount-to-NAV trade—that trade has largely played out. Today it is a bet that Franklin Templeton can convert privately appraised state assets into publicly traded companies, compounding intrinsic value faster than the market already expects.

Uzbekistan — the numbers that matter

37 million people. Comparable to Poland, and the most populous country in Central Asia—not a small frontier outpost.

$110 billion GDP, growing at 5–6% annually since 2017, making it one of the fastest-growing economies in the world.

The reform turning point: President Mirziyoyev came to power in 2016, ending 27 years of Soviet-era isolation. He opened the currency, invited foreign investment and began privatising state assets. The economy that had been largely closed to outside capital is now actively listing companies on the London Stock Exchange.

Why London? The government’s stated goal is to attract Western institutional investors while forcing governance improvements across state-owned companies. The National Investment Fund sits at the centre of that strategy.

How this type of fund works

A closed-end fund—the structure UzNIF uses—trades on the stock exchange like any ordinary share. You can buy it, sell it and watch its price move throughout the day.

What’s different is that the share price reflects what buyers and sellers think the fund is worth, which may be more or less than the estimated value of everything it actually owns.

That gap between market price and Net Asset Value (NAV) is the discount. Historically, closing that discount has been the classic investment thesis for vehicles like this.

The structure itself has a proven template. Three comparable vehicles illustrate different stages of the model:

Fondul Proprietatea

• Managed by Franklin Templeton

• Minority stakes in Romanian state-owned enterprises

• Government-led IPO pipeline

• Governance improvements plus privatisations created value

• Mature / end-stage example

Georgia Capital

• Self-managed

• Controlling stakes in portfolio companies

• Capital allocation drives value creation

• Mid-stage development

UzNIF

• Managed by Franklin Templeton

• Minority stakes in Uzbek state-owned enterprises

• Six IPOs mandated by Presidential Decree

• Governance improvements plus subsidiary IPOs expected to create value

• Early validation stage

The critical structural difference between Georgia Capital and UzNIF isn’t geography.

It’s ownership.

Georgia Capital owns controlling stakes and can dictate strategy directly.

UzNIF owns minority positions in state-owned companies and must influence outcomes through governance, voting rights and engagement.

That makes Franklin Templeton’s willingness to actively exercise those rights arguably the single most important variable.

The person leading UzNIF, Marius Dan, previously served as Deputy CEO of Fondul Proprietatea.

This isn’t merely inspired by the Romanian playbook.

It’s the same playbook, implemented by many of the same people.

What the fund owns—and what’s coming to market

UzNIF owns stakes in thirteen Uzbek companies.

None are mining or oil businesses.

Instead, the portfolio represents the essential infrastructure of a modernising economy:

• Power grids

• Railways

• Water infrastructure

• The national airline

• The largest telecoms network

• A major commercial bank

A Presidential Decree (No. 145, April 2025) requires six of these businesses to complete public listings.

This is what transforms the investment thesis from theory into reality.

Each IPO creates an observable market price for assets that currently exist only within valuation models.

Companies mandated to IPO

• Uzbekistan Airways — National airline (UzNIF stake: 25%)

• Uzbekhydroenergo — National hydropower operator (40%)

• Uzbektelecom — Telecoms backbone (30%)

• O’zbekiston Temir Yo’llari — National railway (40%)

• SQB Bank — Commercial bank (40%)

• National Electric Networks — National transmission grid (40%)

Airways is expected to be first.

Each listing converts an internally modelled valuation into an actual market price.

If those prices exceed today’s estimates, NAV rises from underneath.

The remaining seven investments—including electricity distribution, thermal generation, gas distribution, water utilities, insurance, the commodity exchange and urban transport—remain unlisted but are already included within NAV.

Why the manager matters

The incentives are aligned

Franklin Templeton’s fee structure rewards NAV growth.

• Base fee falls from 1.0% to 0.75% once fund value doubles.

• Performance fee equals 5% of portfolio dividends, but only if NAV grows at least 10% and reaches a new all-time high.

• IPO fee equals 4% of proceeds, while Franklin Templeton pays adviser costs itself.

The incentives encourage successful IPO execution rather than simply asset gathering.

They have already demonstrated independence

Before the fund had even fully launched, the Uzbek government proposed issuing new Uzbektelecom shares to another state agency.

The transaction would have diluted UzNIF’s stake from 30% to 10%.

Franklin Templeton voted against the proposal.

The proposal failed.

That is exactly the sort of minority shareholder activism that defined Fondul Proprietatea’s success.

Investment snapshot

• NAV per GDR: $32.35

• IPO price: $25.00

• Recent GDR price: approximately $32

• Audited unrealised gains (Dec 2024–Dec 2025): $138 million

• Largest contributors:

  • Uzbektelecom: +$93 million

  • National Railway: +$37 million

• Cornerstone investors:

  • BlackRock — $100 million

  • Allan & Gill Gray Foundation — $100 million

  • Franklin Templeton — $30 million

The risks—in order of importance

1. Every valuation is still an estimate

All thirteen holdings are valued using financial models rather than observable market prices.

The first subsidiary IPO will provide the first independent reality check.

2. Energy valuations depend heavily on regulatory reform

Approximately 35% of NAV assumes Uzbekistan adopts a new RAB regulatory framework around 2028.

If implementation slips to 2030, the prospectus sensitivity analysis suggests some regulated utility equity values could approach zero.

The government announced early implementation in April 2026.

That is encouraging.

It is not yet the same thing as legally implemented regulation.

3. There is no contractual discount floor

Unlike Fondul Proprietatea, UzNIF has no mandatory buyback mechanism if shares trade at a large discount to NAV.

With today’s discount largely eliminated this is less significant than at IPO, but it remains an important structural difference.

4. Capital allocation remains unclear

Suppose Airways IPOs above current valuation.

What happens next?

Will Franklin Templeton:

• distribute proceeds?

• repurchase shares?

• reinvest capital?

The prospectus does not provide a definitive answer.

Execution alone is not enough.

Capital allocation after successful IPOs may ultimately determine shareholder returns.

Can I actually buy this?

Unlike many frontier-market opportunities, investing in UzNIF does not require opening a brokerage account in Tashkent.

Its Global Depositary Receipts (ticker: UZNF) trade on the London Stock Exchange.

They are:

• Dollar denominated

• FCA approved

• Supported by standard London settlement infrastructure

• Accessible through most UK investment platforms offering international equities

That accessibility matters.

Many frontier-market reforms become investable only years after the underlying opportunity emerges.

Here, international investors have been able to participate from day one.

Evidence Tracker

Confirmed

✅ BlackRock ($100m) and Allan & Gill Gray Foundation ($100m) invested as cornerstone investors

✅ Franklin Templeton (Fondul Proprietatea team) appointed with contractual minority rights

✅ Franklin Templeton successfully blocked the Uzbektelecom dilution proposal

✅ London Stock Exchange listing with FCA-approved prospectus

✅ Presidential Decree No. 145 mandates six subsidiary IPOs

Still to be proven

◻ Airways IPO completed

◻ RAB regulatory framework formally implemented

◻ NAV validated through observable public market prices

◻ Capital allocation policy following subsidiary IPOs clarified

The investment thesis advances as the outstanding items move from expectation to confirmation.

Final Thoughts

The original investment case was simple.

Buy strategically important state-owned assets at a meaningful discount to NAV and wait for that discount to narrow.

That opportunity has largely passed.

Today’s investment case is considerably more demanding.

It requires believing that Franklin Templeton can replicate the governance and privatisation playbook it developed in Romania—this time in Uzbekistan—and compound intrinsic value faster than the market already expects.

That may still prove to be an exceptional investment.

But investors are no longer buying a discount.

They’re buying execution.

— Harry, The Conviction Play

This article is part of the Capital Formation series, which examines markets at the moment they become investable for global institutions.

Disclosure: I do not currently hold a position in UZNF. I may initiate or exit positions in securities mentioned without notice. This article reflects personal research and opinion only and should not be considered investment advice. Capital is at risk. Please consult a licensed financial adviser before making investment decisions.

UzNIF Official Website, www.uznif.com/en-uz

In this article you’ll learn what actually protects a software company when anyone can build software in an afternoon; why Microsoft’s real moat sits above the code; and how to tell an overheated price correcting from a business being permanently marked down.

Conviction Play theme: Algorithmic Decision Infrastructure, within the Automation & Machine Intelligence Diffusion megatrend.

Everyone knows Microsoft. Far fewer can tell you what they’d actually be buying. Not Windows, not Office — those are the storefront. The business underneath is something stranger and harder to copy: the identity, the data and the distribution that a vast share of the working world runs on. The rails along which corporate work actually moves.

And right now the market is pricing that machine as if the AI wave is about to wash it away. The shares are down close to a third from their high, and nearly a fifth below an overheated peak at the start of June. The mood says: if artificial intelligence can write software for nothing, what protects a software company?

It is exactly the right question — with a more interesting answer than the share price implies.

The mispricing, if there is one, sits in a precise place: not whether Microsoft is a great business, but what its earnings are worth in a world that has stopped trusting software moats.

Our thesis:

The market is selling Microsoft as a software company the AI wave might disrupt. It is really the trust-and-permission layer that the AI wave has to run on — a toll now offered at a discount, if the discount is real.

What I look for first in any company is whether its earnings are durable by construction — whether they keep arriving because of how the business is built, not because of a kind economic cycle or a cheap-money mood.

Microsoft passes that test on the business comfortably.

The entire argument is about the second half of that sentence: the price.

What actually protects a software company

In January this year, the question stopped being theoretical.

AI tools that build working software from a plain-English description — “vibe coding” — arrived in force, and in a single session the market wiped hundreds of billions of dollars off software stocks.

The logic was brutal and simple:

If an AI agent can write the app, reach the data directly and skip the dashboard entirely, what exactly are you paying a software subscription for?

For a great deal of software, that is a fair fright.

The thin layers — the convenience tools, the simple workflow apps, the businesses whose only moat was being the screen people happened to log into — are genuinely exposed.

But the panic confuses the code with the moat, and the two were never the same thing.

Code has been getting cheaper for thirty years; it has never been the thing that stopped a rival.

What stops a rival is everything around the code:

• The data a company holds and you don’t

• The network of users that makes leaving costly

• The regulatory and security approvals that take years to earn

• The sheer trust required to be allowed near a company’s most sensitive systems at all

You can vibe-code a copy of almost any product in a weekend.

You cannot vibe-code your way into ten thousand enterprises’ identity systems, their compliance attestations, or their willingness to let you act on their behalf.

Those are the moats that survive the AI era.

The rest is wrapping paper.

Why Microsoft is the machine, not the casualty

It owns the permission layer

The thing an AI agent needs most is not cleverness — it is the right to act inside a company:

To read the files. Send the mail. Move money.

All under a named identity an auditor can later check.

That layer — corporate identity, permissions and the audit trail — is one of the hardest things in software to build and the last thing a business hands to a stranger.

Microsoft already is that layer for much of the corporate world.

The agent revolution doesn’t bypass it.

It depends on it.

Data gravity and distribution

Work already happens inside Microsoft’s email, documents, meetings and directory — the system of record for how organisations actually run.

An agent that automates your work has to reach into that estate.

And the estate is Microsoft’s.

The switching cost isn’t the software.

It’s everything wired around it.

The agent layer taxes work, it doesn’t replace it

The fear is that if agents do the work, companies need fewer seats and Microsoft’s revenue shrinks.

But agents don’t replace the substrate of work — the documents, the data, the identities.

They run on top of it.

Microsoft’s plan is to charge for that agent layer on top of the seat you already pay for, and in time for what the agents consume.

Work doesn’t leave the estate.

It gains a more expensive layer.

The company that owns where work happens can tax its automation, not be disintermediated by it.

The financials underneath are unglamorous in the best way:

• Earnings per share near $16

• Roughly 25% EPS growth over the past year

• Around 20% annualised EPS growth over the last three years

• Revenue still growing approximately 18%

• Cloud and AI services leading growth

• Fortress balance sheet with almost no net debt

A staircase, not a rollercoaster.

Earnings that have climbed without a meaningful step backwards — precisely the profile the AI panic insists should no longer be possible.

Microsoft (MSFT): price vs. earnings-based fair value, 2021–present. The Conviction Play.

The bear case: a re-rating, not a wobble

The honest risk here is not that Microsoft stops growing.

It is that the price the market will pay for that growth has permanently fallen — and that the fall isn’t finished.

For years Microsoft changed hands at around 31 times earnings.

Today it trades nearer 23 times trailing profit, and under 20 times next year’s.

It has already been marked down hard.

The case that the markdown is deserved is a serious one.

Microsoft has guided to roughly $190 billion of capital spending in 2026 — up about 60% in a single year.

The heaviest capital intensity it has carried since the dot-com build-out.

Roughly two-thirds of that spending is on chips and short-lived equipment that depreciates quickly.

That spending turns an asset-light compounder into something heavier and less certain.

And the payoff rests on AI demand that is still being proven.

Azure’s growth, though still running near 40%, has to slow eventually.

The Copilot add-on meant to monetise the build-out has so far reached only a low-single-digit share of business customers.

Stack those together and the bear’s case is coherent:

31x was a cheap-money, scarcity-premium multiple.

23x is not a trough to buy.

It’s the market correctly repricing a business that will be more capital-hungry, lower-margin and slower-growing than its past.

On that reading, anchoring fair value to the old multiple is just nostalgia.

Re-rate to today’s multiple and the “fair” price falls to almost exactly where the shares already trade.

The discount disappears.

There is a subtler worry too:

That the centre of gravity of work drifts to an agent layer someone else controls, leaving Microsoft a data source rather than the place work happens.

That is a headwind to one vector, not a crack in the moat.

The trust layer is far harder to dislodge than a chat window.

But it is worth watching over years, not quarters.

The questions that decide it

A position like this turns on a handful of open questions, not a single number.

The ones I’m watching:

• Does the market ever pay a premium multiple for Microsoft again?

• Does the heavy capex earn its keep?

• Does depreciation from fast-ageing chips arrive before the AI revenue?

• Does the agent layer actually monetise beyond today’s low-single-digit Copilot adoption?

• Does work stay inside Microsoft’s estate?

• Can Azure hold near its current growth long enough to justify the build-out?

Answer those in Microsoft’s favour and the discount is real.

Answer them the bear’s way and today’s price is roughly fair.

The useful part is that most of them get clearer with the next two or three earnings prints.

Which is exactly why this is a wait-for-evidence call, not a leap of faith.

Great business, right price?

So everything hinges on a multiple I’ve just admitted I no longer fully trust.

The honest way through that is to stop arguing about the multiple at all and run the valuation in reverse.

Instead of picking a fair multiple, discount the future earnings back at a normal rate and solve for the growth today’s price already implies.

Do that, and the answer is sobering.

The price bakes in only single-digit annual earnings growth for years to come.

Against a business still compounding north of twenty percent.

On that test a discount survives even if 31x never returns.

The market is paying for a Microsoft that has already half-stopped growing.

The catch — and it is the whole catch — is that the heavy capex is exactly what could make that pessimism come true.

Depreciation from those data centres lands in future profits whether the AI demand shows up or not.

And if it lands faster than the revenue, the pessimistic growth the price implies becomes the realistic one.

That makes this a discipline problem, not a courage problem.

The business is still compounding and throws off more cash than almost anything its size.

The moat that matters — identity, permissions and trust — is exactly the part the AI wave can’t route around.

And the bull case now rests on one question:

Does the market ever pay up for this again, or has that era quietly closed?

I don’t own Microsoft today, and I’m in no hurry to start buying into a falling price on a multiple I’m no longer sure of.

What would move me is not a number on a screen but evidence.

A double-digit discount to a fair value I actually trust.

One that survives the heavier capex.

And proof that earnings can keep compounding through it rather than being slowly eaten by it.

Admire the machine.

Respect the re-rating.

The toll booth is still collecting.

The only open question is what the market will agree to pay to own it.

— Harry, The Conviction Play

Disclosure

I do not currently hold a position in Microsoft Corporation (NASDAQ: MSFT). The author may buy or sell securities mentioned without notice. This article reflects my personal views and the output of my own valuation system; it does not constitute investment advice. All investments carry risk, and the value of equities can go down as well as up. Readers should conduct their own research or consult a licensed financial adviser before making any investment decision.

If you enjoyed this, subscribe for future pieces on long-term investing, market mispricings and capital allocation written from a calm, conviction-driven lens.

In this article you’ll learn why a great company and a great stock can part ways; how AI threatens not consulting’s survival but the economics that made it exceptional; and what to watch to tell a genuine re-rating from a passing dislocation.

Conviction Play Theme: Workforce Productivity & Augmentation
Megatrend: Labour Scarcity & Demographic Strain

You know Accenture.

The name on the largest technology and transformation programmes in the corporate world — the firm a Fortune 500 board calls when it has to change something enormous and cannot do it alone.

What you may not know is the machine underneath the name.

And right now the market is worried that machine is being quietly rebuilt from the outside.

The shares have fallen more than half in six months, to around 7× trailing free cash flow against a long-run average closer to 20×. The reflexive read is that a quality compounder has been marked down too far.

The more interesting question isn’t whether the discount is wide.

It’s whether the old multiple still describes the business at all.

Accenture (ACN): shares have fallen roughly 58% from their recent peak

Subscribe now

The Thesis in One Sentence

The market isn’t pricing “Accenture dies.” It’s pricing “Accenture’s economics become ordinary” — and you can wreck a stock without wrecking the company.

That distinction is the whole piece.

A company can keep growing revenue, keep its clients, keep its place on every shortlist — and still deserve a lower multiple than it once commanded, because the thing that made its profits exceptional has become less scarce.

That is the first thing I look for in a name like this:

Not “is it cheap?”

But “is the moat strengthening or weakening?”

What Consulting Actually Sells — And What AI Takes

For decades the large advisory firms — Accenture, the Big Four, McKinsey, BCG — ran the same powerful model.

Expertise was scarce and expensive; coordinating a complex programme at scale required a large organisation; so partners leveraged pyramids of junior staff, selling the same expertise repeatedly while associates did the underlying work.

Call it the old equation:

Knowledge → Humans → Billable Hours

Artificial intelligence attacks that equation at its base.

• Research.

• Analysis.

• Modelling.

• Benchmarking.

• Documentation.

• Coding.

• Presentation building.

The tasks that once required teams of analysts increasingly arrive through software at near-zero marginal cost.

The emerging equation looks different:

Knowledge → Software → Outcomes

For any people-intensive business, that is a dangerous transition.

It does not necessarily destroy demand.

It erodes the scarcity the firm used to monetise.

The leverage that built the firm can run in reverse.

A consultant historically joined Accenture because they needed the platform — its talent, methodologies, reputation, and client access.

Give one capable operator a stack of frontier models — AI analysts, researchers, coders, and project managers — and they can do work that once took dozens of associates.

That pushes the industry toward fragmentation:

• Boutiques.

• Independents.

• Small specialist firms.

• Not elimination.

• Fragmentation.

The leverage shifts away from the institution and toward the individual and the platform.

The deepest threat isn’t that AI helps Accenture — it’s that AI reduces the need for it.

Today’s consulting narrative is straightforward:

Clients need our help implementing AI.

That is true.

The harder question is the one a decade out.

Why does the client still need the intermediary once the analysis and the build are commodities?

When the value migrates to the platform, the orchestrator’s cut inevitably comes under pressure.

Where the rent survives is accountability, not analysis.

This is where the bear case is usually overstated — and where I’d push back on my own thesis.

AI doesn’t hit consulting evenly.

It guts the research, benchmarking, slide-building, and coding layer.

It presses on architecture and domain redesign.

It barely touches enterprise transformation, governance, programme leadership, executive alignment, and accountability.

The hard part of a transformation is rarely knowing what to do.

It’s getting fifty thousand people to actually do it.

And carrying responsibility if it goes wrong.

A board often hires Accenture for a reason closer to:

“If this goes sideways, I want them standing beside me.”

than:

“We lack the analysis.”

Knowledge can become free.

Accountability cannot.

The sharper analogy may not be the research analyst whose work AI now replicates.

It may be the auditor.

Clients don’t hire auditors because information is scarce.

They hire them because someone must sign off, absorb risk, coordinate stakeholders, and assume responsibility for the outcome.

In that world, Accenture becomes less intellectual and more operational.

The numbers from my own dashboard:

• Valuation: On my own numbers, about 7× trailing free cash flow per share — a cash-flow multiple, not a reported P/E — against a long-run average near 20× on the same basis, and roughly 57% below its 52-week high. A violent de-rating, not a gentle drift.

• Earnings: Per-share earnings are still growing — up roughly a quarter on a year ago. The collapse is in the multiple, not the cash flows.

• The Fair-Value Gap: My system’s fair-value line sits far above today’s price — but that line is drawn off the historical ~20× multiple, which is precisely what’s in question. The system itself flags the anchor as low-confidence, and I read it the same way.

Accenture (ACN) multiple compression? Price vs. free-cash-flow-based fair value.

The Bear Case

The real risk isn’t revenue collapse.

It’s margin compression.

Premium economics came from three things:

• Scarce expertise

• Leverageable labour

• Capabilities clients couldn’t easily replicate

AI compresses all three at once.

Revenue may survive.

Economic rents may not.

Even if the top line grows, returns on human capital can fall — and that alone can justify a lower multiple.

The market may simply be saying:

Accenture will still be here in 2036, but it won’t deserve the same premium.

That is a far milder claim than:

Consulting is dead.

The genuinely dangerous scenario is the innovator’s dilemma in its purest form.

The bear case isn’t that Accenture fails to adopt AI.

It’s that it adopts AI perfectly and discovers that perfect adoption reduces the number of billable humans the business needs.

The technology that makes you more productive can destroy the scarcity that made you valuable.

The debate isn’t whether AI hurts consulting.

It’s whether 20% of the value stack disappears or 70% does.

Today’s compression suggests the market is increasingly willing to entertain the larger number.

A re-rating that merely normalised the multiple toward the low-to-mid teens would still leave the stock a long way below where this business has historically traded.

Accenture (ACN): The Conviction Play framework read — quality gates, valuation, and stance. Valuation on a FCF basis.

Great Business — Still An Exceptional One?

This is the discipline beat, and for once it isn’t really about price.

Accenture is cheap against its own history.

The catch is that “its own history” is the assumption under attack.

I’m not interested in the headline discount when the anchor it’s measured against may be migrating beneath it.

So I’m not acting on the gap.

What would change my mind isn’t another AI-partnership press release.

It’s evidence that the rent is holding:

• Pricing power

• Margin resilience

• Client retention

• Returns on capital

that refuse to fade as the technology diffuses to everyone.

The future of consulting probably won’t be decided by who adopts AI fastest.

It may be decided by whether clients ultimately pay for intelligence — or for accountability.

If they pay for intelligence, AI is deflationary and the rent erodes.

If they pay for accountability, the moat may prove far more durable than the market is currently willing to assume.

That is the question.

And I don’t yet know the answer.

Which is exactly why I’d want to be paid for the uncertainty before forming a view with my capital.

— Harry, Independent Investment Researcher & Writer

Disclosure

I do not currently hold a position in Accenture. The author may buy or sell securities mentioned without notice. This article reflects my personal views and the output of my own valuation framework; it does not constitute investment advice. All investments carry risk, and the value of equities can go down as well as up. Readers should conduct their own research or consult a licensed financial adviser before making investment decisions.

If you enjoyed this article, consider subscribing for future research on long-term compounders, market mispricings, capital allocation, and emerging investment themes through the lens of The Conviction Play.

Subscribe now

In this article: what a “pre-profit” business is and why even a brilliant one defeats valuation; what SpaceX actually earns its money doing; and why the day a great company finally becomes buyable is so rarely the day worth buying.

Subscribe now

Watch a Falcon booster fall back from the edge of space, slow itself on its own engines, and land upright on a drone ship in the Atlantic, and you’re watching something that looked like science fiction a decade ago. SpaceX has now done it hundreds of times. It has made orbit cheaper than anyone thought possible, strung thousands of satellites into a constellation beaming internet to nearly nine million households, and bolted on an artificial-intelligence arm besides. As a feat of engineering and ambition, there is nothing else like it on public markets.

And as of 12 June 2026, it is on public markets: after years as the most valuable private company on earth, SpaceX listed on the Nasdaq as SPCX. Here is what a disciplined investor has to say out loud, though — I can stand in awe of the machine and still have no honest idea what it’s worth. Those are two different questions, and merging them is how good investors lose money on great companies.

The thesis in one sentence

SpaceX may be the most remarkable company of its generation — but a business without a stable earnings line is one no disciplined framework can price, and the long history of IPOs says the moment everyone can finally buy is rarely the moment worth buying.

That’s not a verdict on the rockets — it’s a verdict on the information. My process begins with what a business earns and what the market has historically paid for it; take the earnings away, and the anchor everything else hangs from isn’t there.

What a “pre-profit” business is, and why it breaks valuation

Most companies I write about are mature compounders — years of steady profit, a track record, a normal range of prices the market has paid. From that you can estimate fair value and judge whether today’s price offers a margin of safety.

A whole class of business denies you that — the cash-burning startup, the just-listed IPO, the conglomerate mid-transformation — whose earnings are negative, erratic, or simply too short to mean anything. There’s no settled profit to anchor to, so any “fair value” is a forecast dressed as a measurement: a story about year ten, discounted back and presented as fact.

That’s the trap. Not that such businesses are bad — some become the investments of a generation — but that they invite you to guess precisely, and the guess does all the work. Before any company earns a place in my book it must answer five plain questions: does it earn real returns, can it reinvest and compound them, can it fund itself without endlessly issuing shares, is it sensibly governed, and is there a credible path to growing value per share? A company with no public track record can’t yet answer one of them — not “fails,” can’t answer. Admission waits on evidence.

SpaceX is, right now, exactly that kind of name — a shame, because the business underneath is fascinating.

What SpaceX actually is

Strip away the romance and SpaceX is three very different businesses bolted together.

The rockets are the moat. Reusable launch is a structural cost advantage no rival has yet matched at scale — the picks and shovels of the space economy. Whoever needs a satellite in orbit or a crew at the station, the odds are SpaceX is the cheapest ride. Launch and crew brought in about $4 billion in 2025, while the company ploughed nearly $3 billion into its next-generation rocket, Starship. Increasingly that demand is sovereign — national-security launches, allied governments, and a defence-oriented Starshield network — making the capability strategic infrastructure as much as a commercial service.

Starlink is the cash machine — the part most investors underrate. Satellite internet is a recurring-subscription business: sell the dish once, bill the household every month. It has scaled ferociously, from 2.3 million subscribers in 2023 to 8.9 million by the end of 2025, turning over $11.4 billion in 2025 — about 61% of the company — at roughly $4.4 billion of operating profit. The core of SpaceX is no money-loser; it’s a highly profitable connectivity utility wrapped inside a rocket company.

The AI arm is the wild card. SpaceX has absorbed an artificial-intelligence business that booked $3.2 billion of revenue in 2025 — and a $6.35 billion operating loss. That single line turned a profitable enterprise into a loss-making one.

The IPO prospectus holds the whole tension in three figures:

• Top line: $18.7 billion of revenue in 2025, led by Starlink.

• Split personality: Starlink earned ~$4.4bn operating profit; the AI arm lost ~$6.35bn; the group swung from a ~$791m net profit in 2024 to a $4.9 billion net loss in 2025.

• Honest core: adjusted EBITDA — a rough gauge of operating cash generation — of about $6.6 billion, so the underlying business throws off real cash before the AI build-out.

This is no vaporware IPO; there’s a dominant, cash-generative engine at its heart. But there’s also no clean, repeatable bottom line — a profitable utility and a loss-making moonshot sharing one income statement — which is exactly why no earnings-based fair value will hold still.

The bear case: the price you can’t check

Lead with the real risk, and here it’s the price. SpaceX priced at $135, valuing it near $1.8 trillion; within days the market had pushed it past $2 trillion — bigger than Tesla, among the largest companies in America — on $18.7 billion of revenue and a reported loss. That’s roughly a hundred times sales: not a measure of what the business earns, but a down-payment on a future that has to go almost perfectly right.

History is unkind to that bet. An old trading-desk joke has it that IPO really stands for It’s Probably Overpriced — and the evidence rhymes. Insiders choose when to sell, and, as Warren Buffett puts it, that timing is “unlikely to be a time that’s good for you.” Jay Ritter’s classic study found newly listed firms badly underperformed established peers over the three years after listing — worst of all for the most hyped offerings in the hottest markets. None of that is destiny, but it’s the base rate, and the base rate deserves respect.

Two more flags, named plainly. Governance: Elon Musk’s super-voting shares give him roughly 85% of the votes on about 42% of the equity — public shareholders are passengers on mergers, capital and strategy. And the AI arm will keep clouding the very earnings line a framework like mine needs until its losses narrow or it’s valued separately.

Great business, no price

Here admiring the machine and acting on it part ways. SpaceX clears the awe test effortlessly and the valuation test not at all — not because it failed, but because it can’t yet sit the exam. No settled earnings line means no fair value, no margin of safety to demand, no level at which the price is plainly cheap. The honest answer is the boring one: do nothing, and wait for evidence.

What would change my mind isn’t a better story — it’s better information. A few quarters of reported results that let the business be judged on what it earns rather than what it might. A clean line between the profitable Starlink core and the cash-burning AI bet. And a price that asks for a margin of safety rather than demands faith. Until then, the most exciting company on the market is, for me, a watch-list name and nothing more. The rockets are extraordinary. The discipline is just arithmetic — and the arithmetic isn’t there yet.

— Harry, The Conviction Play

Sources: financial figures are drawn from SpaceX’s IPO prospectus (Form S-1) filed with the SEC, and 2025 results disclosed at its 12 June 2026 Nasdaq listing.

Disclosure: I do not currently hold a position in SpaceX (Nasdaq: SPCX). The author may buy or sell securities mentioned without notice. This article reflects my personal views; it does not constitute investment advice. All investments carry risk, and the value of equities can go down as well as up. Readers should do their own research or consult a licensed financial adviser before making any decision.

If you enjoyed this, you can subscribe for future pieces on long-term investing, market mispricings, and capital allocation — written from a calm, conviction-driven lens.

In this article you’ll learn how to recognise a “cyclical” business and why they catch out even seasoned investors; what Thermo Fisher actually does and why it has grown for decades; and how to tell a genuine bargain from a value trap.

One of the best businesses in the world has just fallen more than 25% from its high. Its sales have softened, its profits have dipped, and the mood around it has turned gloomy. To a bargain-hunter, that sounds like an invitation; to a sceptic, a warning. Which it is depends almost entirely on the kind of business you are looking at.

The company is Thermo Fisher Scientific, and it is an unusual case: part dependable, decades-long compounder, part something far trickier — a cyclical. Before we can judge whether today’s lower price is a gift or a trap, we need to understand what a cyclical is, why it catches people out, and how much of one Thermo Fisher really is.

Subscribe now

First: what is a cyclical business?

Most companies grow more or less steadily. A cyclical does something different — its fortunes rise and fall in waves, tied to the health of the wider economy or its own industry’s boom-and-bust rhythm. When times are good, demand surges and profits balloon; when the cycle turns, orders dry up and profits collapse. Carmakers, airlines, steel and chemical producers, housebuilders, oil and gas, semiconductors, and to a degree defence — these are the classic cyclicals. Think of the oil price doubling and then halving, or the “chip cycle” that swings semiconductor makers from glut to shortage and back.

Cyclicals can be wonderful investments — if you buy near the bottom, when things look worst and prices are lowest, and ride the recovery up. But they are genuinely treacherous, for one counterintuitive reason that the great investor Peter Lynch spent pages warning about: with a cyclical, the usual valuation signals run in reverse.

Normally, a low price-to-earnings ratio — the share price divided by yearly profit, a rough gauge of how cheap a stock is — signals a bargain. With a cyclical, it often signals the opposite. At the top of the cycle, profits are at a record high, so even a steep share price looks “cheap” measured against that inflated profit — and that low ratio is precisely the danger sign, because the next move is down. At the bottom, profits have collapsed, so the stock looks “expensive” on a high ratio (or has no earnings to measure at all) — and that is often when the real opportunity is greatest. As Lynch put it, buying a cyclical after years of record earnings, just as its valuation has fallen to a tempting low, is a well-worn way to lose half your money.

The lesson most investors should draw from this is humility. Timing a cyclical means knowing its industry well enough to read the turn, and getting it wrong is costly. For the less experienced, the safest instinct is to treat an obviously “cheap” cyclical with suspicion rather than excitement — and to be certain you can tell a temporary dip from a permanent decline before committing a penny.

Which brings us to Thermo Fisher — because it is not a pure cyclical at all.

What Thermo Fisher actually is

If modern science and medicine are a gold rush, Thermo Fisher sells the ‘picks and shovels’. It is the largest supplier in the world of the everyday equipment that makes research and drug-making possible: the chemical reagents and disposable plastics a laboratory burns through daily, the precision instruments on its benches, the kits that run medical tests in hospitals, and — increasingly — the outsourced factories and clinical-trial services that drug companies hire to develop and manufacture their medicines. Whoever’s new drug succeeds, the odds are that Thermo Fisher supplied some of the tools that made it.

That reach is enormous. With $44.6 billion of sales in 2025, Thermo Fisher is bigger than any direct rival — a position built over twenty years of steady acquisitions. And, crucially, roughly 82% of those sales are recurring: not one-off machines, but the consumable supplies and ongoing services customers reorder again and again. It is the “razor-and-blade” model of science — sell the instrument once, sell the supplies forever.

Why does it keep growing? Because the forces beneath it are some of the most dependable in the economy. The world is ageing and spending ever more on healthcare; drug discovery is increasingly a science of biology, with more experimental medicines, more clinical trials and more complex therapies, each demanding more tools and testing; and drug companies are handing more of their manufacturing and research to specialist partners — exactly the business Thermo Fisher has spent tens of billions building. Over time, that has made it a quiet compounding machine. Stripping out the accounting noise from all those acquisitions and measuring the business by the hard cash it produces, Thermo Fisher has grown its free cash flow per share — the surplus cash each share is entitled to — at roughly 11% a year for more than fifteen years, from about $3 in 2009 to nearly $18 today.

The twist: a compounder with a cyclical heartbeat

So if the demand is so durable, why has the stock fallen so far? Because layered on top of that steady growth is a cyclical pulse — and lately it has been beating hard.

Thermo Fisher’s customers are research labs, biotech firms and drug manufacturers, and their spending rises and falls with funding, confidence and the wider economy. When money is pouring into biotech and research budgets are generous, they buy instruments and supplies freely; when it tightens, they wait. On top of that ordinary rhythm came the single biggest demand shock the industry has ever seen: the pandemic.

COVID was a mega-cycle in miniature. The world suddenly needed staggering quantities of testing, vaccines and bioproduction, and Thermo Fisher — a key supplier of all three — saw demand explode. Customers, desperate not to run short, ordered far more than they needed. Then the emergency passed and they were left with stockpiles to work through, so they stopped buying. This “destocking” — running down a glut of inventory instead of placing fresh orders — weighed on the whole sector for nearly two years. You can watch it in Thermo Fisher’s cash: free cash flow per share spiked above $20 in the boom, then sagged back toward $16 as the destocking bit. Add a soft patch in China and tighter research budgets in the United States, and you had a great business that, for a stretch, looked decidedly ordinary — and a share price down 26%.

This is the key to the whole puzzle. Thermo Fisher is not a steelmaker whose profits vanish in a downturn; its vast base of recurring supplies meant its cash dipped, not collapsed. It is a structural compounder with a cyclical overlay — durable enough that the lights never flickered, cyclical enough to go on sale.

Trough or trap?

So we return to the question we started with, better equipped to answer it. Is this a temporary dip in a great business — a trough — or the start of a lasting decline — a trap? Three questions settle it: is the lost demand merely delayed or genuinely gone; is the company’s dominant position intact; and is the cash still flowing?

On all three, the weight of evidence points to trough. The destocking was customers using up supplies they had already bought — demand postponed, not destroyed; the underlying science never paused. By late 2025 the signs of a turn were clear: orders from drug and research customers began to normalise, the suppliers hit hardest by the destocking were back to around 10% annual growth, and Thermo Fisher told investors to expect up to 6% revenue growth in 2026 — and raised its dividend, something management tends to do only when it can see recovery in the order book. Its dominance never slipped, and its credit rating was actually upgraded. Tellingly, while Thermo Fisher’s cash flow per share has shrunk about 4% a year over the last two years — the dip — it has still grown about 8% a year over three — the trend. The bear says the two-year figure is the new reality; the bull says it is the cycle. So far, the recovery data favours the bull.

The road ahead

Look past the cycle and the long-term drivers are as durable as any in investing: an ageing, health-spending world; an explosion in biological drug discovery; and the steady shift of drug-making and research into the hands of outsourced specialists like Thermo Fisher. The recovery is already showing up in the cash — free cash flow is expected to reach $6.8–7.3 billion in 2026 — and management keeps buying back around $3 billion of shares a year, so each remaining share quietly owns a little more of the business. None of this needs a bold forecast; it needs only that the world keeps funding science and making medicines.

The catch

Two honest risks remain. The first is the bears’ best point: that part of the weakness isn’t cyclical at all. Pressure on US government and university research funding — tighter grants, budget uncertainty — could be a lasting drag on the slice of Thermo Fisher’s business tied to academic and public labs, even as biopharma recovers. It is a real structural blemish on an otherwise cyclical story, and the thing most worth watching.

The second is price — and here the cyclical lesson pays off. Because Thermo Fisher’s cash never collapsed, its valuation isn’t sending the deceptive low-ratio signal that traps buyers of pure cyclicals; it trades at about 26 times free cash flow, modestly below its long-run norm near 30. That puts fair value somewhere around $529 against a recent price near $473 — cheaper than usual, but only by about a tenth. It also carries meaningful debt from its years of dealmaking — comfortably investment-grade, and falling, but a reminder this is a leveraged compounder, not a fortress. If the soft patch proved structural rather than cyclical, a more cautious multiple would drag fair value into the high $440s, and a harsher one toward the high $390s. The downside is real if the trough-or-trap call goes the wrong way.

Thermo Fisher (TMO): price vs. free-cash-flow-based fair value, 2021–present. The Conviction Play.

So — bargain or trap?

Almost certainly a bargain in the making rather than a trap — but “almost certainly” is doing real work in that sentence, and that is exactly why patience matters here.

A pristine compounder asks one question of a buyer: the price. A cyclical compounder asks two — that the recovery prove itself, and that the price offer a real margin of safety. On the first, the trend is encouraging but not yet confirmed; another quarter or two would lay the structural-softness fear to rest. On the second, a 10% discount is a start, not the cushion a careful investor should demand of a leveraged, cyclical name; the deep discounts this business hands out only in moments of panic are where the great opportunities have always been.

So the honest verdict is neither “buy” nor “avoid,” but “watch closely.” Thermo Fisher is a genuinely wonderful business, caught in a cycle that has made it cheaper than it usually gets. Admire it now, and act when the recovery confirms itself and the price does the work. The trough is the opportunity — just make sure it really is a trough.

— Harry, The Conviction Play

Disclosure: I do not currently hold a position in Thermo Fisher Scientific (NYSE: TMO). The author may buy or sell securities mentioned without notice. This article reflects my personal views; it does not constitute investment advice. All investments carry risk, and the value of equities can go down as well as up. Readers should do their own research or consult a licensed financial adviser before making any decision.

If you enjoyed this, you can subscribe for future pieces on long-term investing, market mispricings, and capital allocation — written from a calm, conviction-driven lens.

Subscribe now

In this article you'll learn what makes MSCI worth owning — and how to recognise any business built to compound its earnings, year after year.

MSCI is recognised everywhere in investing, usually as a label on the products people buy rather than as a company in its own right. Think MSCI World, or MSCI Emerging Markets — the benchmarks that sit beneath thousands of index funds and ETFs. Far fewer investors stop to look at MSCI Inc. itself as a business you can own. Do that, and one of the purest structural compounders in the public market comes into view: consistent earnings growth year after year, a deep competitive moat, and profit margins most companies can only envy.

This is the opposite of the usual under-the-radar story. MSCI is famous, and it is richly valued. The mispricing, if there is one, isn’t in the name — it’s in how rarely people appreciate the compounding machine beneath it. MSCI gets paid because markets exist, and because the world’s participation in those markets keeps growing, year after year.

That is the whole investment case in a sentence: an earnings engine that compounds quietly, consistently, and for a very long time.

Subscribe now

What MSCI actually is

MSCI began life inside Morgan Stanley — the name is short for Morgan Stanley Capital International — and was spun out as a standalone company in a 2007 listing on the New York Stock Exchange. Since that debut it has been one of the best-performing shares in the US market, returning north of 2,600%: roughly every $1,000 invested at the IPO became more than $23,000, compounding at about 20% a year for nearly two decades.

Strip away the share price and the business is elegantly simple. MSCI builds and licenses the indices that the investment industry uses as its yardsticks. When a pension fund decides it wants exposure to global shares, or emerging markets, or developed Europe, the benchmark it reaches for is very often an MSCI one. Around $18 trillion of assets worldwide are measured against MSCI indices, and more than $2.3 trillion sits in ETFs that track them directly. Index licensing is the crown jewel — roughly 55% of revenue but around 70% of operating profit — and around it MSCI has built three further businesses: the analytics and risk tools institutions run their portfolios on, sustainability and climate data, and a growing private-markets arm spanning real estate and private capital.

What ties them together is the quality of the revenue. Most of MSCI’s income is recurring — clients pay every year for continued access, rather than once for a product — and they renew at a remarkably high rate. Trailing earnings per share have climbed from about $10.50 to $17.85 over the past four years: a steady upward staircase, with no meaningful setback along the way. That consistency is the thing to hold on to. It is the signature of a business worth understanding thoroughly before any price is even discussed.

What makes a business a structural compounder

The single most important number in investing is a company’s earnings — what it genuinely makes for its owners — and the surest way to lose money is to lose sight of whether those earnings are real, durable, and growing.

A structural compounder is simply a business whose earnings you can rely on to keep rising because of how it is built, not because of luck or a kind turn in the economic cycle. A few plain tests set them apart. The revenue should be recurring — money that arrives again next year without having to be re-won from scratch. The company should be able to reinvest its profits at high rates of return, so each dollar it keeps tends to throw off more profit later. It should grow the value of each share without constantly printing new ones — issuing lots of new shares quietly shrinks every existing owner’s slice of the company, so a business that funds its own growth is worth far more over time. And it should be run by managers who behave like owners.

Most companies fail at least one of these tests. MSCI passes all four — and the reason it passes comes down to the nature of its moat.

The moat: a toll booth on global investing

A competitive moat is whatever stops rivals from competing a company’s profits away. MSCI’s is unusually wide, and the cleanest way to picture it is a toll booth.

As more of the world’s savings flow into index-tracking funds, the assets benchmarked to MSCI’s indices rise — and a slice of MSCI’s fees rises automatically alongside them, at almost no extra cost to the company. MSCI doesn’t have to pick winners or predict markets; it simply takes a small, growing cut of the fact that investing happens at all. And it sits above the industry’s fee war rather than in it: while ETF providers compete ferociously to charge investors ever less, MSCI sells them all the same essential ingredient — and is paid whichever provider wins. In 2025 alone, ETFs tracking MSCI indices pulled in a record $204 billion of new money.

The second strand is switching costs. Changing the benchmark beneath a fund sounds trivial; in practice it rewrites performance records, client reporting, fund documentation, risk systems and regulatory filings all at once. Once an institution has built itself around an MSCI benchmark, leaving is expensive and disruptive — so almost no one does. Client retention runs in the mid-90s percent, year in and year out.

The third strand is industry structure. Indexing is effectively an oligopoly: MSCI, S&P Dow Jones Indices and FTSE Russell — the “Big Three” — together control more than 80% of global index revenue, and in Europe MSCI alone benchmarks close to half of all passive equity assets. Being the standard is itself a moat: the more managers, products and traders quote an MSCI index, the more everyone else has to. It is a network effect wrapped around a trusted brand wrapped around decades of proprietary data.

The financial signature of all this is exactly what you’d expect of a well-run toll booth:

• Earnings: trailing EPS of $17.85, up around 14% on the year and compounding close to 15% a year over three years.

• Quality: operating margins above 55%, more than 90% of profit converting into free cash, and client retention around 95%.

• Growth: revenue rising in the low-to-mid teens, across all four segments.

The road ahead

What makes this more than a snapshot is that the toll road keeps getting busier. The long migration from active stock-picking to low-cost index funds is still in full flow, and every fresh wave of assets widens MSCI’s base of fees. Global participation in markets keeps deepening — more savers, more countries, more product. Newer currents add to the flow: demand for sustainability and climate data, the gradual opening of private markets to index-style measurement, and the rise of customised and “direct” indexing — all of which MSCI is positioned to charge for. Its total run-rate of recurring revenue passed $3.1 billion during 2025 and was still growing at around 10%. None of this depends on a heroic forecast. It needs only that the world carries on investing, and carries on wanting to measure how it’s doing.

The catch: price, not business

If there is a risk here, it isn’t the business — it’s the price. Genuinely wonderful companies are rarely allowed to look cheap, and MSCI is no exception. The shares change hands at roughly 34 times earnings: broadly in line with their own long-run average, but around 74% above the ~20 times its industry has averaged over the past decade. The market understands the quality, and charges for it.

Current valuation vs estimated fair value anchor (as of 16 June 2026)

That premium already assumes years of continued growth, and if growth slowed sharply — or a prolonged bear market shrank the pool of index-linked assets — the rating could compress and the shares could fall well before the business did. But this is a reason for patience, not avoidance. A great company at a full price is still a great company; it simply sets a higher bar for when to buy. The moat is not in doubt. Only the entry point is.

Great business, right price?

Which leaves the one question that truly matters with a company like this: not whether MSCI is excellent — it plainly is — but at what price excellence becomes a good investment.

On a sober estimate, fair value sits somewhere around $642 a share against a recent price near $599 — a discount, but a slim one, and well short of the margin of safety a careful buyer would insist on. History suggests MSCI rarely offers a deep discount; the moments it has — a market panic, a passing growth scare — have tended to be precisely the moments to act. The discipline, as always, is to separate admiration from action: to know exactly what you would be glad to own, and then to wait, patiently, for the market to offer it at a price that does the work for you.

The strongest moats, in the end, sit not in the companies competing within a market but in the infrastructure the whole market quietly depends on. MSCI is that infrastructure. What stands between it and a high-conviction holding is not the business — it is a price. And prices, eventually, move.

— Harry, Investment Researcher & Writer

Disclosure: I currently hold shares in MSCI Inc. (NYSE: MSCI). This article reflects my personal views; it does not constitute investment advice. All investments carry risk, and the value of equities can go down as well as up. Readers should do their own research or consult a licensed financial adviser before making any decision.

If you enjoyed this, you can subscribe for future pieces on long-term investing, market mispricings, and capital allocation — written from a calm, conviction-driven lens.

Subscribe now

Peter Lynch ran the Magellan Fund at Fidelity and compounded returns at 29% per year for over a decade — one of the best records in investing history.

His core idea was simple.

Earnings drive stock prices.

Not immediately. Not cleanly. But over time, they do.

“Earnings” just means profit — what a company actually makes. And Lynch’s point was straightforward: in the short run, markets can price stocks on excitement, fear, or narrative. But over time, price and profit reconnect.

This article tests that idea using a real stock: AeroVironment (AVAV).

AeroVironment makes small defence drones used by the US and its allies — an area that has seen rising investor interest in recent years.

Over the last six months, AVAV has fallen nearly 50% — from narrative-driven highs back to where its earnings actually justify. That move isn’t random.

A Simple Way to Estimate Fair Value

There’s a simple way to anchor a stock to its earnings — used by Lynch and described in his book One Up on Wall Street — called the Earnings Line.

It works like this:

• The company’s profit over the last year (called trailing twelve months earnings per share, or TTM EPS)

• Multiply it by the typical valuation the market has historically given that business

That gives you a rough fair value.

We’ll call it the Earnings Line (EL):

EL = Earnings × Typical P/E multiple

P/E (price-to-earnings) simply means how much investors are willing to pay for each $1 of profit.

For AVAV today:

• Earnings (TTM EPS): $3.01

• Typical valuation: ~63× earnings

• Earnings Line: ~$189.63
  

In other words:

Investors have historically paid about 63 times what AVAV earns each year.

The EL is not a prediction. It’s a reference point — a way of asking:

What would this business be worth if it were priced normally?

Price Reverting to Earnings Line — AeroVironment (AVAV) Price vs Earnings Line (2021-2026)

What Happened to AVAV

In late 2025, AVAV wasn’t trading anywhere near that level.

The stock ran sharply higher, driven by strong sentiment around defence technology and growth expectations.

At the peak:

• Price reached over $400

• Earnings Line was roughly $190
  

That’s a premium of more than 100%.

In simple terms:

The market was pricing AVAV as if its earnings would need to roughly double — at the same valuation — to justify the price.

That’s not a small stretch. It’s a dislocation.

What Happened Next

Over the following months:

• The stock fell 47.5% over six months

• It is now trading at $189.26

• The Earnings Line sits at $189.63
  

Price didn’t fall because the business deteriorated.

In fact:

• Earnings are growing (+44.0% YoY)

• Revenue growth is +143.4% YoY
  

What changed was the relationship between price and earnings — not the business itself.

And eventually:

Price returned to where earnings said it should be.

This Isn’t a One-Off

Looking at every time AVAV moved more than ±25% away from its Earnings Line over the past decade:

• 19 dislocation events

• 84.2% reverted back to fair value

• In the post-2022 regime:
  
  → 7 out of 7 reverted (100%)
  

These reversions did not happen immediately.

On average:

• Reversion occurred over weeks to months

• Overvalued periods corrected faster

• Undervalued periods recovered more gradually
  

The key insight is this:

The Earnings Line doesn’t predict timing. It predicts destination.

What This Means in Practice

This is not a trading signal.

It’s a positioning tool.

When price is far above earnings (e.g. +25% or more):

• Expectations are ahead of reality

• Risk of medium-term decline increases

When price is far below earnings (e.g. -25% or more):

• The business is being discounted

• Recovery tends to happen — but more slowly

When price ≈ earnings (like today):

• Price and fundamentals are aligned

• Future returns depend on earnings growth, not revaluation

When This Framework Breaks

There is one important exception.

If earnings themselves change materially — for example:

• A major shift in growth

• A structural change in profitability
  

Then the “anchor” moves.

In those cases:

Price does not revert — because fair value has changed.

How to Use This

You don’t need a complex model to apply this idea.

A simple approach:

• Look at what a company earns over the last year (EPS — earnings per share)

• Check what valuation the market has typically applied (P/E — price-to-earnings multiple)

• Compare it to today’s price

If the price is far above what earnings justify, be cautious. If it’s far below, be patient and consider adding gradually.

If you want to go deeper, Peter Lynch’s One Up on Wall Street remains one of the clearest guides to thinking about stocks this way.

Final Thought

Peter Lynch put it simply:

“The stock market is filled with individuals who know the price of everything and the value of nothing.”

AVAV is a recent example of what happens when price runs ahead of value — and then comes back.

Earnings didn’t move. Price did.

The Conviction Play.

This article reflects personal research and is for informational purposes only. It is not financial advice. All investments carry risk.

Subscribe now

Between June 2025 and February 2026, Grindr’s (GRND) share price fell 61%.

Over the same period, revenue grew 28%, the company became GAAP profitable for the first time, and guidance was raised repeatedly.

The market wasn’t reacting to the business. It was reacting to a margin call — a forced liquidation triggered when lenders demand repayment after the value of pledged shares falls below required collateral levels.

Understanding why requires a brief look at the ownership structure. Grindr’s two largest shareholders — James Lu, through his vehicle Longview Holdings, and George Zage — together controlled more than 60% of the company’s shares. Public filings also disclosed that substantial portions of those holdings had been pledged as collateral against personal loans.

In a company with a relatively thin public float, that structure created a specific vulnerability: if the share price fell far enough, lenders could compel forced sales regardless of the shareholders’ views on the business. Those sales would then hit a market with limited capacity to absorb them.

That is exactly what happened.

For investors who understood the mechanics, the situation created two distinct opportunities: first, the snapback as forced selling exhausted itself; then a more fundamental re-rating once the market began to focus again on what the business was actually worth.

This is the story of how it unfolded, what the signals were, and what the numbers suggest from here.

What Forced Selling Does to a Price

Forced selling is one of the most misunderstood dynamics in markets, because on the surface it is nearly indistinguishable from genuine deterioration. A falling stock, negative sentiment, no obvious floor — it looks like the market knows something you don’t.

The mechanics are straightforward. A large holder who has borrowed against their position faces a margin call when the value of their collateral falls below the lender’s threshold. They are required to sell — regardless of price, regardless of their view on the business. The selling pressure drives the price lower. The lower price deepens the collateral deficit, which forces more selling. A self-reinforcing spiral, driven entirely by financial structure rather than business reality.

What makes it an opportunity is precisely its mechanical nature. It has a definable cause. It has a traceable endpoint — the moment the supply exhausts itself. And the price it creates is entirely disconnected from intrinsic value.

The Telltale Signs — All Present in Grindr

These situations don’t announce themselves, but they do leave tracks. In Grindr’s case, every key signal was visible in public data for those looking.

Concentrated ownership with disclosed leverage. With Lu and Zage together holding over 60% of shares, and both having disclosed pledged collateral in SEC filings, the vulnerability was structural and on the record. Lu’s position through Longview was so large relative to the tradeable float that any forced liquidation would overwhelm demand almost regardless of how carefully it was managed.

A short report that publicly named the mechanism. In September 2025, a published short report explicitly identified the pledged collateral and warned that a price decline could trigger a forced liquidation cascade. This was not a generic short attack on the business. It targeted a specific structural vulnerability — and it was widely covered.

Elevated short interest confirming the bet. Short interest built visibly through autumn, consistent with institutional short sellers pressing the margin call thesis. The short side was not speculating on business deterioration — it was pressing a known structural pressure point.

Form 4 filings confirming steady liquidation. From October 2025, Lu filed SEC Form 4 disclosures within days of each sale. The overhang was traceable in near-real-time: tranche size, price achieved, position remaining. You could watch the supply shrink week by week.

Analysts maintaining price targets throughout. When a business is genuinely deteriorating, analysts cut targets. Throughout this episode they did the opposite. The average analyst target held at $21–22 while the stock traded at $13, then $11, then sub-$10. A 70–100% implied upside gap from institutional analysts is a clear diagnostic: the price action is being read as technical, not fundamental.

The Catalyst

In September 2025, a published short report flagged that insiders had pledged the majority of their Grindr shares as personal loan collateral and warned explicitly that a price decline could trigger a forced liquidation cascade. The report was widely covered and not obscure. Short interest built from there, applying selling pressure that deepened the collateral deficit for leveraged holders.

The key insight — relevant to anyone trying to build a framework from this case — is that none of this required access to private information. The pledged shares were in the public filings. The liquidation tranches would be filed on Form 4 within two days of each sale. The mechanism was in a report covered by financial media. The opportunity was a matter of reading what was already public.

The Fundamentals Said the Opposite

While the stock declined through autumn, the business was delivering its best results since listing. On November 6, Grindr reported Q3 earnings that beat on every metric: revenue of $115.8M against a $113.5M consensus, EPS of $0.16 against $0.12 — a 33% beat — with quarterly revenue growth of 30% year-on-year. Full-year EBITDA guidance was raised. The business and the stock were moving in opposite directions.

That divergence — price falling, business accelerating, analysts not cutting — is the fingerprint of forced selling. It is not the fingerprint of fundamental deterioration.

The Forensic Timeline

Every significant development was visible in public filings within days of occurrence. The key beats:

Jun 2, 2025 — GRND 52-week high: $25.13

Sep 9 — Published short report names pledged collateral, warns of margin cascade. Stock −2.4% on day. ~$15.31

Oct–Nov — Lu begins selling in steady tranches via Form 4. Q3 earnings beat on every metric — revenue +30% YoY, EPS +33% vs estimate. Stock ignores it.

Nov 26 — $18/share take-private proposal from Lu and Zage withdrawn. The overhang becomes the entire story.

December — Zage buys in the open market, absorbing ~40% of Lu’s sales. Price holds. Then the earnings blackout period begins. Zage can no longer legally trade as an insider. The one consistent buyer disappears. Price accelerates from $13 toward $11.

Feb 4–6 — THE LOW — Lu’s largest single tranche: 1.45M shares in three days at ~$10.07 average. $9.73 intraday low.

Feb 14 — Q4 13F filings published. Institutional accumulation revealed throughout the forced selling period: Deutsche Bank +1,984%, Woodline Partners new $7.4M position, SG Americas +140%.

Feb 26 — Q4 earnings beat. FY net income $103M. Fourth consecutive GAAP-profitable quarter. Expanded buyback authorised.

Early March — ~20% recovery from lows within two weeks. ~$11.24

Opportunity 1: The Snapback

The 20% recovery from $9.73 happened before Q4 earnings were even reported. The 13F filings published on February 14th revealed that Deutsche Bank, Woodline Partners, SG Americas and others had been accumulating through the forced selling period — institutional buyers positioning ahead of the print rather than reacting to it. When Q4 confirmed the business narrative on February 26th, the resilience made sense: the repricing had already begun.

The first opportunity was a flow opportunity. The thesis was not about the business — it was about the mechanics. Forced supply exhausted itself, equilibrium reasserted, and price reverted toward fair value. The Zage blackout dynamic — a single identifiable buyer who could no longer legally buy — even provided a rough timing framework for the capitulation. When the tranches got largest, the end was closest.

Opportunity 2: The Fundamental Re-Rating

The second opportunity is different in character. With the forced selling chapter largely closed, the market is now confronted with a more straightforward question: what is Grindr worth on its own merits?

Goldman Sachs maintains a Buy at $17. JMP Securities sits at $21. The average analyst consensus is $18–22. Institutional coverage is also expanding, with Morgan Stanley recently initiating coverage following the forced-selling period. Against a current price of approximately $11.24, that implies 50–95% upside from analysts who held their targets through the entire decline — and set them before Q4 confirmed the business was tracking ahead of expectations.

The valuation is the anchor. With an enterprise value of approximately $2.35B and 2026 guided EBITDA of greater than $217M, the stock currently trades at roughly 11x forward EBITDA. Profitable subscription platforms growing revenue at 20%+ with 44% EBITDA margins do not typically trade at 11x. The multiple compression was a consequence of the forced selling event, not a reflection of the business. The question is how quickly the market rebuilds conviction once the noise clears.

The FY2025 results that were hiding behind the overhang:

Revenue: $439.9M (+28% YoY) Net income: $103M (first full profitable year) EBITDA margin: 44.5% Q4 revenue: $126M (beat $122M estimate) Shares repurchased: 25.1M 2026 revenue guidance: >$528M

The EPS Growth Case

What makes the re-rating argument analytically interesting is that the earnings per share growth case rests on multiple independent drivers — each meaningful in isolation, compounding together.

Buyback programme. In Q4, the company authorised an additional $400M in share repurchases, bringing total remaining repurchase capacity to approximately $450M across authorised programmes through March 2029. The company already repurchased 25.1M shares in 2025 at prices well above current levels. At current prices, that remaining capacity could retire a substantial portion of the float — a mechanical EPS tailwind that compounds even without any revenue growth.

Pricing and product strategy. Grindr is repricing subscription tiers upward across its product range. Revenue per paying user is already expanding. The company appears to have resolved a structural tension that limited earlier monetisation efforts: historically, aggressive paywalling risked degrading the free experience and weakening the network. The current strategy inverts this — premium tiers generate revenue from power users, and those proceeds fund improvements to the free product. Premium users effectively subsidise the network, strengthening the engagement and network effects that make the platform durable.

New products and operating leverage. The EDGE premium tier — an AI-native product for power users — is in active testing, excluded from 2026 guidance pending visibility. Two years of AI infrastructure investment are beginning to pay back: AI generated roughly 60–70% of new code in 2025 and increased engineering productivity by ~1.5×, per management. As fixed costs scale more slowly than revenue, margin expansion compounds.

Marketing optionality. Grindr has grown almost entirely organically. Paid marketing, particularly in under-penetrated international markets, represents an untapped revenue accelerator — upside the current valuation does not require to justify the thesis.

User base quality. The MAU decline that concerned analysts through 2025 was largely attributable to bot removal and platform cleanup, not genuine user churn. The reported base is cleaner and more monetisable than the headline numbers implied.

Q1 and guidance structure. Management indicated that Q1 2026 revenue growth and EBITDA margin are expected to pace ahead of full-year guidance. Several revenue initiatives were deliberately excluded from the 2026 outlook pending clearer visibility. If any are incorporated through the year, consensus estimates will need to move.

A New Chapter

The bear thesis that dominated sentiment through the liquidation period — declining users, margin call contagion, credibility questions — has been addressed by the results. Revenue grew 29% in Q4 alone. Net income for the full year reached $103M. Grindr is guiding to $528M+ in 2026 revenue, with expanding margins and a substantial buyback programme providing mechanical EPS support regardless of top-line growth.

The business that emerges from this period is profitable, cash-generative, and compounding across multiple independent levers. The price at which it is currently available was set not by fundamental analysis but by a margin call cascade. The gap between those two things is the analytical opportunity.

The Broader Lesson

Forced selling situations occur wherever concentrated, leveraged ownership meets a thin public float. The specific mechanics vary; the fingerprints are consistent. Price detaches from value. Public filings tell the story in near-real-time. Analysts hold their targets. The business keeps printing. And the gap between price and intrinsic worth grows until the supply exhausts itself.

Two opportunities emerged from one dislocation. The first required nerve — buying into the final stages of a forced selling cascade with no guarantee of the exact bottom. The second requires conviction — building a position in a profitable, cash-generative platform at a price created by mechanics rather than fundamentals.

When forced selling sets the price, markets briefly stop behaving like auctions of opinion and begin to resemble liquidation sales. The merchandise is the same. The price is not. And the filings tell you exactly when the sale is ending — for anyone willing to read them.

Tracking forced selling in real time requires nothing proprietary. Set up Google Alerts for “[ticker] Form 4” and “[ticker] insider selling.” The key diagnostic at each filing: are the tranches getting larger or smaller? Larger tranches under falling prices signal mounting margin pressure. When they shrink, then stop — that is the exhaustion signal. For any concentrated-ownership stock, the first thing to check is whether major holders have disclosed pledged shares in SEC filings. That single data point is the precondition for everything described in this article.

This article reflects personal investment analysis and opinion and is not financial advice. Investors should conduct their own research before making any investment decisions. All data sourced from publicly available SEC Form 4 filings, 13F disclosures, company earnings releases, and market data.

A couple of weeks into the year, one of the better-performing stocks in the market isn’t AI, defence, crypto, or anything remotely futuristic.

It’s pancakes. Coffee refills. Rocking chairs. Southern comfort.

Cracker Barrel (NASDAQ: CBRL) is up roughly 37% year-to-date, quietly outperforming many of the narratives that dominate market conversation. That contrast is what makes it interesting — not because it’s a prediction about the future, but because it’s a reminder of how markets actually behave.

I first paid attention to Cracker Barrel last year for an unusual reason. It briefly became a national talking point after a proposed rebrand sparked backlash so intense that even President Donald Trump weighed in. When the company reversed course, he publicly applauded the decision, saying:

“Congratulations ‘Cracker Barrel’ on changing your logo back to what it was. All of your fans very much appreciate it.”

That moment stuck with me. Not because of politics, but because it revealed something deeper: this wasn’t a forgotten, irrelevant brand. It was a business people felt personally attached to. The reaction wasn’t indifference — it was emotion.

From a distance, the episode looked like a misstep. From closer up, it looked like a stress test that revealed how much the brand still mattered.

By the end of last year, Cracker Barrel’s share price had been pushed down hard. Sentiment was poor, expectations were low, and the valuation compressed to levels that implied a business in terminal decline. Some of that reflected genuine challenges. But some of it was mechanical.

As the calendar year ends, stocks that have disappointed are often sold aggressively — tax-loss harvesting, portfolio clean-ups, and general “clearing of the decks.” It’s not a story unique to Cracker Barrel. You see it across markets every December. Then January arrives, the tax year resets, capital flows return, and prices often rebound — especially where pessimism overshot reality.

That doesn’t mean every January bounce is justified. But it does mean that sentiment can push prices well past what the underlying business deserves — in either direction.

Cracker Barrel is not a company of the future. It doesn’t need to be. It’s a large, cash-generative business with hundreds of locations, deep cultural roots, and a customer base that clearly doesn’t want it to disappear. Management is attempting a turnaround, and whether that ultimately succeeds remains to be seen. But the idea that the business had suddenly become worthless never really added up.

One small, slightly absurd detail put that into perspective for me. At the lows late last year, Cracker Barrel’s market value had fallen to around half a billion dollars. The company has roughly 365,000 Instagram followers. In theory — very much not as a serious exercise — if each follower had chipped in a little over a thousand dollars, they could have bought the whole company.

That’s not an investment thesis. It’s just a human way of illustrating how far expectations had collapsed.

When I recently posted a light-hearted image of Cracker Barrel alongside its share price, the response wasn’t analytical at all. It was nostalgic. Messages came in from Americans saying they used to go there with their parents. Blueberry pancakes. Family road trips. Sunday breakfasts. That kind of reaction doesn’t show up on a balance sheet — but it matters.

None of this is about declaring Cracker Barrel a winner, or pretending pancakes have replaced technology as the engine of progress. It’s simply a reminder of something easy to forget in markets obsessed with the future:

prices move on sentiment as much as stories, and value often appears where expectations have quietly collapsed.

Sometimes that looks like a new technology. Sometimes it looks like free coffee refills.

And sometimes, two weeks into a new year, it looks like both humility and surprise served on the same plate.

Harry — The Conviction Play

Disclaimer: Not financial advice.

Markets do not only misprice weak businesses.

Sometimes they misprice excellent ones — not because something breaks, but because too many people owned them for too long, expected too much too quickly, and eventually grew tired of waiting.

Evolution AB (EVVTY) increasingly looks like one of those cases.

As of late 2025, the company remains highly profitable, structurally important to its customers, and firmly cash generative — yet it trades near multi-year valuation lows after several years of share-price stagnation. Nothing dramatic has gone wrong. There has been no collapse in demand, no balance-sheet stress, no loss of relevance.

Attention has simply moved elsewhere.

That combination — a durable business, a decaying narrative, and exhausted shareholders — is often where mispricings quietly begin.

From momentum darling to prolonged indifference

For investors who bought Evolution near its 2021 peak, the experience since has likely been frustrating. The share price is down roughly 50% from those highs. Rallies have struggled to hold. Each recovery has felt less like a new chapter and more like an opportunity for someone else to exit.

The chart tells a psychological story as much as a financial one: a long, grinding unwind of enthusiasm.

This distinction matters. Sentiment-driven drawdowns behave very differently from fundamental ones. They rarely end in a single dramatic collapse. Instead, they fade slowly — through boredom, disappointment, and quiet abandonment.

Evolution does not look broken.
It looks like a stock the market has simply lost patience with.

And impatience, more than fear, is often what creates durable mispricings.

What Evolution actually does (and why it is misunderstood)

Evolution is not a consumer gambling brand.

It does not advertise.
It does not acquire players.
It does not compete for attention.

Instead, it supplies live-casino infrastructure and digital casino games to regulated online casino operators. This includes live dealers, game formats, studios, streaming technology, and compliance-ready systems that power what players ultimately see on screen.

That distinction is critical.

Evolution is not betting on gamblers.
It is betting on operators needing reliable, regulator-approved, always-on content.

Once an operator integrates Evolution’s platform, switching is neither quick nor trivial. Games are embedded into the user experience and tied to regulatory approvals, operational workflows, and player expectations. Over time, this creates real friction — and pricing power.

Viewed through this lens, Evolution looks less like a volatile gambling stock and more like infrastructure inside a regulated digital ecosystem.

When sentiment turns, however, nuance is rarely rewarded.

The economic engine: cash generation first, growth second

At its core, Evolution is a cash-generating business.

Even after significant reinvestment, the company converts a large proportion of revenue into free cash flow. Working capital requirements are modest. Incremental content carries high margins once studios are built. Growth is largely self-funded.

This matters because it reframes almost every debate around the stock.

Evolution does not need external capital to grow.
It does not need aggressive leverage.
It does not need heroic assumptions.

It needs time.

Margins in context: dilution, not deterioration

One of the most frequently cited concerns is margin compression.

On paper, margins are lower than at pandemic-era peaks. That is true. But peaks reached during extraordinary operating conditions are a poor benchmark for long-term quality.

As Evolution expands into more tightly regulated markets — particularly the United States — its cost structure naturally changes. Certain jurisdictions require state-by-state studios, often co-located within physical casinos. That is structurally more expensive than serving entire regions from centralised hubs.

In Q3 2025, Evolution reported an adjusted EBITDA margin of 66.4%, comfortably within management’s 66–68% full-year guidance — despite these mix effects.

Margins are compressing at the group level not because the business is weakening, but because the revenue mix is changing. What matters is whether incremental revenue remains attractive. By most observable measures, it does.

Expansion-driven dilution is not the same thing as competitive decay, even if markets occasionally treat them as interchangeable.

Geographic volatility is a feature, not a flaw

Evolution operates across jurisdictions with very different regulatory maturity, enforcement intensity, and market structure. Volatility is not an anomaly — it is part of the operating reality.

Asia has been the most visible pressure point recently. In Q3 2025, reported net revenue declined modestly year-on-year, while constant-currency growth remained positive. The gap reflects geography and regulatory friction more than demand.

Management has been explicit that recent disruption stems from heightened cybercrime countermeasures, regulatory transitions, and temporary over-corrections in certain markets. In several cases, Evolution prioritised compliance and platform integrity over short-term revenue optimisation.

That choice weighs on near-term optics. It also reinforces the durability of the business model.

Compounding businesses are rarely smooth quarter to quarter.

Capital returns as signal, not spectacle

One of the least discussed — and most telling — aspects of the past two years has been Evolution’s capital allocation.

Despite slower growth and increased regulatory scrutiny, the company has returned more than €900 million to shareholders through dividends and share buybacks, while maintaining a net cash position and substantial balance-sheet flexibility.

These buybacks are not cosmetic.

They suggest that management sees no need for defensive hoarding, no urgency for large acquisitions, and no better use of excess capital than returning it to shareholders at current valuations.

In a compounding business, capital discipline is often the clearest signal of confidence.

A second engine hiding in plain sight

While Live Casino remains the core, Evolution’s RNG segment has quietly grown into a meaningful secondary engine.

In Q3 2025, RNG revenue grew low-single digits year-on-year, outpacing Live Casino growth for the first time. While smaller in absolute terms, RNG is less capital-intensive and broadens the monetisation base.

It is not a new story.
It is simply no longer being listened to.

What this is — and what it is not

This is not a turnaround.
It is not a momentum trade.
It is not a regulatory arbitrage bet.

It is a case study in how strong, cash-generative businesses can become mispriced when narratives decay faster than fundamentals.

As Warren Buffett once observed, “The stock market is a device for transferring money from the impatient to the patient.”

The appropriate posture here is not urgency, but observation.

Let valuation do the work.
Let the chart stabilise.
Let uncertainty compress before capital is committed.

The opportunity is not that Evolution suddenly becomes great again.

It is that the market quietly realises it never stopped being good.

Final thought

Markets love novelty. They hate familiarity. And they punish yesterday’s winners far longer than seems reasonable.

Evolution feels like a business the market has walked away from — not because it failed, but because it became inconvenient, complex, and unfashionable.

Those are not guarantees of upside.

But for patient observers, they are often where it begins.

Source: Evolution AB Investor Relations

Subscribe now

Disclaimer: This article reflects personal views only and does not constitute investment advice. All investments carry risk, and the value of equities can go down as well as up. Readers should conduct their own research or consult a licensed financial adviser before making investment decisions.

What if one of Europe’s most consistently profitable banks wasn’t based in a major financial centre, but in a fast-modernising country of 3.7 million people in the South Caucasus — with a second engine of growth emerging in nearby Central Asia?

TBC Bank Group, headquartered in Tbilisi and listed in London, is one of Georgia’s two leading banks. It rarely appears in mainstream European banking discussions. Yet it consistently delivers returns on equity above 20%, operates one of the region’s most advanced digital ecosystems, and maintains capital and liquidity ratios comparable with much larger institutions.

Despite this profile, the stock trades at valuation levels more typical of lower-quality or higher-risk banks. It is a disconnect worth examining.

A Leading Franchise in a Concentrated Market

TBC is one of two dominant players in Georgia’s banking sector, a market defined by rational competition, conservative supervision, and relatively high financial penetration. This structure gives scale institutions a durable advantage.

The bank serves retail, MSME, and corporate clients, supported by a digital-first strategy that has reshaped customer behaviour. Georgia now ranks among the region’s leaders in digital banking adoption, with TBC at the centre of that evolution.

This combination of market concentration, operational discipline, and digital capability underpins TBC’s long-term resilience.

Digital Leadership and Regional Ambition

Over the past decade, TBC has built one of the strongest mobile banking platforms in the Caucasus. A growing share of customer activity now runs through digital channels, allowing the bank to scale efficiently while improving service quality.

Beyond Georgia, TBC is expanding in Uzbekistan—a significantly larger market with rapid digital transformation and low banking penetration. Uzbekistan’s population is roughly ten times that of Georgia, yet formal financial usage is still early in development.

If TBC succeeds in executing its digital-first model in a market of this scale, Uzbekistan could evolve into a second engine of growth rather than a side business.
This optionality remains underappreciated in current valuations.

Financial Performance: High Returns and Operational Discipline

TBC’s financial profile is notable for its consistency. Across recent reporting periods, the bank has delivered:

• ROE of 20–25%
• Strong net interest margins, supported by disciplined pricing and a stable deposit base
• A lean cost structure, with cost-to-income ratios near 34%
• Prudent credit quality, with conservative NPL and coverage ratios

TBC has compounded book value per share at mid-teens rates for much of the past decade, supported by steady double-digit loan growth and durable margins. If sustained, this alone can generate attractive long-term shareholder outcomes, even without multiple expansion.

Capital and Liquidity Strength

TBC maintains strong CET1 capital ratios, comfortably above regulatory minimums. CET1, a core measure of loss-absorbing capacity, provides resilience and strategic flexibility.

Liquidity is equally robust, supported by:

• a diversified depositor base
• conservative maturity management
• access to international financial institution funding

This foundation has enabled the bank to remain stable through periods of external volatility.

Valuation: A Disconnect Between Price and Fundamentals

Despite its performance, TBC trades at valuation levels that understate its structural quality:

• P/B near 1.0x
• P/E materially below many European peers

Historically, the bank has traded closer to 1.1–1.7x book, even while generating similar returns. Today’s valuation appears driven more by geography, index exclusion, and perceived geopolitical risk than by fundamentals.

At these levels, investors are effectively paying near book value for a business generating 20–25% ROE—implying high single-digit to low double-digit annual returns even without a re-rating.

A higher multiple is possible over time, but not required for long-term value creation.

Context: Why This Bank Is Often Overlooked

Georgia is a small but fast-growing economy with a reform-oriented policy environment and a well-regulated banking system. Still, several factors limit TBC’s visibility among global investors:

• modest market capitalisation
• limited index representation
• lower liquidity
• geopolitical perception rather than economic reality

These characteristics can screen the bank out of institutional mandates, but they do not diminish its operating strength or balance-sheet quality.

Risks: Real, But Manageable

Key considerations include:

• Macro sensitivity: Georgia’s open economy is exposed to external shocks.
• Currency volatility: GEL movements matter for foreign investors.
• Regulation: Generally strong, though changes could affect profitability.
• Geopolitical perception: Sentiment risk often outweighs fundamentals.
• Uzbekistan execution: Scaling in a new regulatory environment carries complexity.
• Liquidity: Frontier-market listings trade with modest daily volume.

These risks are typical of well-run emerging-market banking systems and warrant monitoring rather than alarm.

Conclusion: A Quiet Compounder Worth Watching

TBC Bank may not be widely recognised beyond the region, but its fundamentals are difficult to ignore. It is consistently profitable, operationally efficient, well-capitalised, and digitally advanced. Its expansion into Uzbekistan adds long-term optionality, while steady book-value growth provides a reliable engine for compounding.

Whether the market re-rates the stock or not, a bank compounding intrinsic value at high rates tends to reward patient shareholders.

In a sector where many institutions struggle to deliver sustainable double-digit returns, TBC stands out as a quietly compounding franchise in an overlooked geography.

Written in Tbilisi, Georgia
— Harry, Independent Investment Researcher & Writer

Learn more here

🔗 Read TBC Group’s latest annual report here.

Disclaimer

This article is for informational purposes only and does not constitute investment advice. All investments carry risk, including the loss of principal. Readers should conduct their own research or consult a licensed financial adviser before making investment decisions.

Georgia Capital (LSE: CGEO) is a UK-listed investment holding company with majority stakes in several of Georgia’s most established private-sector businesses spanning banking, healthcare, insurance, pharmacy retail, education, and renewable energy. These are market-leading, cash-generating operating companies — not early-stage or speculative ventures.

At FY-2024, Georgia Capital reported net asset value (NAV) of GEL 3.6bn (≈£1.02bn), equal to GEL 95.95 (£27.14) per share. NAV per share has compounded at approximately 33% over three years, 29% over five years, and 18% per year since inception.

Despite this track record, shares continue to trade at an estimated 30–40% discount to independently assessed NAV, meaning public-market investors are effectively paying 60–70p for every £1 of underlying economic value.

Recent results suggest the discount remains active rather than historical: in Q3 2025, NAV per share grew 7.9%, supported by approximately 30% EBITDA growth across large private-portfolio holdings. The central question is therefore not whether value is being created, but why public markets have yet to reflect it.

Subscribe now

1) Platform Structure and Value Creation Model

Georgia Capital operates as a permanent-capital business-building platform, distinct from closed-end private-equity vehicles and passive conglomerates. The model centres on:

• Targeting scalable, capital-light businesses capable of reaching GEL 300m+ equity value within 3–5 years

• Taking majority ownership and board control

• Professionalising governance, reporting, strategy, and capital allocation

• Retaining full flexibility over exit timing

The company is UK-domiciled, audited by PwC, reports under IFRS, and uses Kroll for independent third-party valuations. NAV is formally assessed annually, with consistent methodology applied to interim checks.

2) Portfolio of Market-Leading, Cash-Flowing Businesses

The portfolio combines a listed valuation anchor with majority-owned private operating companies in essential-service sectors characterised by recurring demand, pricing power, and high entry barriers.

Listed Holding

Lion Finance Group (formerly Bank of Georgia) — 17.8% stake (as of Sep-2025)

• H1 2025 profit: GEL 1.03bn (+28% YoY)

• ROE: ~28%

• Market share: ~38% loans | ~41% deposits

• ~86% of retail loans originated digitally (Q2 2025)

• Loan-book growth guidance: ~15% annually

This holding provides transparency and an externally observable valuation reference for NAV.

Private-Portfolio Highlights

Healthcare Services

• National network of hospitals, clinics & diagnostics

• +20% revenue, +46% EBITDA, >19% margin (Q3 2025)

• Leverage reduced from 5.0× → 3.8×

Pharmacy & Health Retail

• 438 stores, ~36% organised-market share

• +7.5% retail revenue, +6.6% same-store, +33% wholesale

• +18% EBITDA (Q3), >90% cash conversion

• Leverage ~1.3×

Insurance (Aldagi, Imedi L, Ardi)

• Leading P&C and health insurers

• 9M 2025: +9% revenue

• P&C ROE >40%, Health ROE ~38%

Renewable Energy & Education

• Hydropower assets and private-schools platform

• Steady, demand-aligned expansion

Across the portfolio, shared attributes include recurring demand, margin improvement, strong cash-flow conversion, and prudent leverage.

3) Macro Backdrop: Growth With Stability

Georgia is a small, open, reform-oriented economy located between Europe, Central Asia, and the Middle East, combining sustained growth with fiscal discipline and broad trade access.

• Real GDP: +9.4% (2024) | +7.9% (8M 2025)

• Infrastructure capex: ~7.2% of GDP (2024)

• Inflation target: 3% (latest ~4.8%, Sep-2025)

• Public debt: ~36% of GDP

• Free-trade access: ~2.8bn consumers

• Tourism revenue: $3.6bn (9M 2025) | 141% of 2019 level

The key constraint is capital-market depth, not economic fragility.

4) Why the Discount Persists

The discount appears structural rather than fundamental:

• Legacy regional-risk perception

• Frontier-market classification

• Liquidity constraints for institutional investors

• Local value creation vs offshore price discovery

• Multi-asset-holding complexity

• Global capital narrative focused on AI, defence, and mega-caps

This suggests a pricing inefficiency, not a business-quality concern.

5) Potential Catalysts

Re-rating is typically driven by evidence of value realisation, not storytelling. Potential catalysts include:

• Partial exits or asset monetisation at or above carrying value

• Sustained NAV per-share compounding

• Continuation of discounted buybacks

• Improved liquidity and potential index eligibility

• Clearer segment-level disclosure

• Selective portfolio simplification

Common theme: visibility, comparability, and liquidity — not new narratives.

6) Capital Allocation & Balance-Sheet Discipline

Georgia Capital has shifted from expansionary investment to disciplined capital return, improving balance-sheet strength and increasing per-share value creation:

• Reduced capital commitments as the portfolio matured

• Net-cash position with ~US$77m liquidity

• Share count reduced to ~35m (down >25% from post-demerger peak)

• Capital-return programme (GEL 700m) prioritising buybacks at a discount

Why it matters:
Cash flow is now increasingly returned to shareholders, not recycled into dilution-risk projects. At a 30–40% NAV discount, repurchases are mathematically value-accretive.

7) Risks

• Regional geopolitical sentiment

• GEL/GBP currency translation effects

• Single-economy concentration

• Private-asset valuation periodicity

• Limited share liquidity

These factors justify some discount — the debate is how much, not whether.

Conclusion

Georgia Capital offers majority-controlled exposure to essential-service, cash-generating businesses at a 30–40% discount to independently assessed NAV, supported by sustained operating performance, disciplined capital allocation, and a shrinking share base.

For patient investors able to tolerate liquidity and sentiment cycles, the thesis is simple:

Georgia Capital appears to trade at a discount to intrinsic value despite owning high-quality, cash-generative businesses that deliver >20% ROE; management compounds NAV through disciplined capital allocation, essential-sector exposure, and aggressive share buybacks that concentrate value for long-term holders.

Written in Tbilisi, Georgia
— Harry, Independent Investment Researcher & Writer

Disclaimer: Thanks for reading. If you enjoyed this piece or have your own perspective on under-the-radar opportunities, I’d be keen to hear from you.

You can subscribe 📩 for future articles on long-term investing, market mispricings, and capital allocation — all written from a calm, conviction-driven lens.

This article reflects my personal views only and does not constitute investment advice. I currently hold shares in Georgia Capital (LSE: CGEO). All investments carry risk, and the value of equities can go down as well as up. Readers should conduct their own research or consult a licensed financial adviser before making investment decisions.

• Georgia Capital PLC — Annual Report 2024. Published March 20 2025. Available from the company’s website.

• Georgia Capital PLC — 3Q 2025 & 9M 2025 Results / Earnings Call. Published October 27 2025.

What if one of Europe’s most profitable banks was based not in London, Paris, or Frankfurt, but in a small country in the Caucasus region — and it was hiding in plain sight?

Bank of Georgia (LSE: BGEO) is not a household name. Based in Tbilisi, it quietly delivers returns on equity close to 29%, has grown its customer base and loan book year after year, and leads its region in digital banking adoption. Yet its valuation remains modest by international standards. Listed on the FTSE 250 under the holding name Lion Finance Group PLC, it trades at just over one times book value and at a price-to-earnings ratio below 5x.

That combination—strong fundamentals, steady execution, and low valuation—makes it worth a closer look.

A Digital Leader in a Fast-Growing Market

Bank of Georgia serves over 1.7 million customers — a remarkable footprint in a country of just under 4 million people. Operating across retail, SME, and corporate banking, it is known locally for its user-friendly digital platforms. Today, more than 95% of retail transactions are now conducted digitally — the result of years of deliberate investment in technology aimed at boosting convenience and driving down costs. Here CEO, Archil Gachechiladze, describes the banks focus on AI as a driver of modernisation:

“We’re delighted that Global Finance recognises our commitment to AI innovation. From risk management to personalised offerings, AI is integral to our operations… As we scale generative AI, we remain focused on evolving our capabilities to meet diverse client and business needs.” — CEO Archil Gachechiladze, Bank of Georgia (Lion Finance Group PLC)

In 2024, the bank took a significant step toward regional diversification by acquiring Ameriabank in neighbouring Armenia. While this introduces complexity — including cross-border integration and a new regulatory environment — it positions the bank for long-term growth beyond its home market.

“Our successful expansion into Armenia [via Ameriabank] is a historic milestone… despite the holding company’s new name, the identity and operations of Bank of Georgia and Ameriabank will remain unchanged, focusing on empowering customers and contributing positively to their communities.” — CEO Archil Gachechiladze, Bank of Georgia (Lion Finance Group PLC)

Financial Performance: Strong and Consistent

The numbers from Q1 2025 are impressive:

• Return on Equity (ROE): 28.7% — a measure of how efficiently the bank uses shareholder capital. For context, many large European banks target 10–15%.

• Net income (excluding one-off items): ₾513 million, up 39% year-on-year.

• Net interest income: ₾684 million, up 56%.

• Cost-to-income ratio: 35% — indicating a lean and efficient operation.

• Loan book: ₾34.1 billion, up 23.2% year-on-year.

• Book value per share: ₾170.99, up 25.8%.

Despite this, the stock trades at a price-to-book (P/B) ratio of just 1.1x and a price-to-earnings (P/E) ratio below 5x — far below typical levels for profitable banks. For context, European banks with solid fundamentals often trade at P/B multiples of 1.5x to 2.0x, and P/E ratios of 8–10x. These figures are more commonly associated with low-growth or distressed institutions — not with banks delivering nearly 29% returns on equity.

Lion Finance Group: 500%+ Growth, Hidden in Plain Sight (BGEO)

Stock Chart June 2020-June 2025 over 500% growth (not inclusive of dividends)

Context: Why Might This Be Overlooked?

Georgia is a small country, often misunderstood or simply overlooked by international investors. While it borders Russia, Turkey, Armenia, and Azerbaijan, it has consistently followed a pro-European, market-oriented trajectory in recent years.

Economic growth has been strong: averaging 7% annually since 2021. Public debt remains relatively low, and the banking sector is well regulated. Yet investor awareness remains limited, and liquidity in the stock is thin, with average daily volumes under 25,000 shares.

This isn’t necessarily irrational. For larger institutional investors, the lack of trading volume, geographic unfamiliarity, and perceived geopolitical risk can be disqualifying. But for long-term retail investors or smaller funds, these same dynamics can occasionally create opportunities.

Risks: Real, But Manageable

No investment is risk-free, and Bank of Georgia is no exception. Key considerations include:

• Currency exposure: The Georgian lari (GEL) can fluctuate meaningfully against GBP or USD.

• Political perception: Regional tensions and local protests can affect sentiment, even when the bank’s operations remain unaffected.

• Execution risk: Integrating Ameriabank in Armenia brings complexity.

• Liquidity: The stock is lightly traded. This is not a high-turnover name.

That said, the business has operated through volatile periods before, and has remained profitable and disciplined.

Conclusion: A Quiet Compounder Hiding in Plain Sight?

Bank of Georgia may not be on most investors’ radar, but its fundamentals are hard to ignore. It is consistently profitable, operationally lean, digitally advanced, and modestly valued. Its home market is growing, and its early steps into regional expansion are promising.

Whether it earns a re-rating remains to be seen. But at today’s valuation, it offers something rare: a high-return, well-run bank trading at a discount more typical of troubled institutions. For long-term investors, this disconnect between price and performance could present a case of asymmetric upside grounded in fundamentals.

That said, price and value can diverge — and any investment decision should rest on a well-reasoned view of intrinsic value, coupled with the discipline to act only when price aligns with conviction.

In a market where many stocks are priced for perfection, Bank of Georgia stands out as something rarer still: a quietly compounding business hiding in plain sight.

Written in Tbilisi, Georgia
— Harry, Independent Investment Researcher & Writer

🔗 Read Lion Finance Group’s latest annual report here: Lion Finance Group Investor Relations Site

Disclaimer: Thanks for reading. If you enjoyed this piece or have your own perspective on under-the-radar opportunities, I’d be keen to hear from you.

You can subscribe 📩 for future articles on long-term investing, market mispricings, and capital allocation — all written from a calm, conviction-driven lens.

This article reflects my personal views only and does not constitute investment advice. I currently hold shares in Georgia Capital (LSE: CGEO). All investments carry risk, and the value of equities can go down as well as up. Readers should conduct their own research or consult a licensed financial adviser before making investment decisions.

Subscribe now