The Best Buys come from The Mispriced Research Index — our internal ranking system used to compare investment opportunities across the same framework.

We’re focusing on the top 25% of the Index by QPOS Score — currently, every opportunity scoring 79/100 or higher.

Subscribe now

Let’s go!

7. Autotrader Group

Auto Trader Group operates the UK’s largest automotive marketplace, connecting car buyers with retailers and private sellers. Its platform combines vehicle listings with pricing data, advertising tools, retailer software, and digital retailing products designed to move more of the car-buying process online. The business benefits from strong network effects: a large buyer audience attracts sellers, while broad vehicle inventory reinforces consumer traffic.

Investment Snapshot

• Market Cap: $4.93B

• Revenue: $826.3M

• P/E: 13.5x

• P/FCF: 12.6x

• ROIC (5Yr Avg): 48.4%

• 5Yr Base Case Fair Value: £8.65

• Implied Upside: +88.5%

• Buy Zone: £4.30 - £4.75 → 🟡 INSIDE

• QPOS Score: 79/100

6. Intuit

Intuit is a financial technology company serving consumers, small and mid-sized businesses, and accounting professionals. Its main products include QuickBooks for accounting, payroll and payments, TurboTax for tax preparation, Credit Karma for personal finance, and Mailchimp for marketing automation. Its business is largely built around recurring software subscriptions, transaction-based services, and financial-product referrals.

Investment Snapshot

• Market Cap: $73.72B

• Revenue: $21.45B

• P/E: 16.8x

• P/FCF: 8.5x

• ROIC (5Yr Avg): 14.7%

• 5Yr Base Case Fair Value: $525

• Implied Upside: +91%

• Buy Zone: $285 - $360 → 🟡 INSIDE

• QPOS Score: 80.1/100

5. Wavestone

Wavestone is a global consulting firm focused on business and technology transformation. It advises large companies on areas such as strategy, IT transformation, cybersecurity, data and AI, operational change, and risk management. The firm operates across 17 countries with more than 6,000 employees and generates most of its business from helping clients manage complex, large-scale transformation programs.

Investment Snapshot

• Market Cap: $835.35M

• Revenue: $1.1B

• P/E: 8.9x

• P/FCF: 5.7x

• ROIC (5Yr Avg): 15.8%

• 5Yr Base Case Fair Value: €63

• Implied Upside: +111%

• Buy Zone: €35 - €40 → 🟢 BELOW

• QPOS Score: 82.3/100

Read more

• Introduction

• The risk is being in the wrong place

• The index is no longer a neutral decision

• From a valuation bubble to an earnings question

• Liquidity matters more than it used to

• Gold, and the leverage inside the miners

• Crypto is resetting

• I am not investing for a black swan

• What matters is already in the price

Let’s go.

Subscribe now

Introduction

For most of my investing life, falling interest rates were simply part of the background.

For most investors, they were even more than that.

From the early 1980s through 2021, the world experienced one of the great secular declines in interest rates.

The exact causes changed over time, but the direction was remarkably persistent. Lower inflation, globalization, demographics, central-bank credibility, technology, and eventually quantitative easing all contributed to a world where capital became progressively cheaper.

That cycle broke in 2022.

I don’t know if we are at the beginning of a forty-year cycle in the opposite direction.

Nobody does.

But I think it would be a mistake to build a portfolio that requires the old world to return.

When the risk-free real return available to investors rises, everything else has to compete with it.

Long-duration assets become more sensitive to assumptions.

Weak businesses have to refinance at higher costs.

Growth financed by cheap money becomes harder to justify.

Most importantly, the hurdle rate for investing goes up.

The risk is being in the wrong place

I don’t think the most useful question today is whether the market is going up or down.

The better question is:

That leads to a very different portfolio

The biggest risk, in my view, is not necessarily a broad market crash.

It is owning the wrong businesses, sectors, or assets in a regime where the cost of capital is structurally higher, inflation remains less predictable, and performance becomes increasingly dispersed.

That last point is important.

After years in which owning the index was enough, I think the payoff from selection is becoming more interesting again.

George Noble recently made a similar observation: dispersion is increasing and corrections are appearing underneath the surface of the market. His conclusion is considerably more bearish than mine, particularly toward consumer equities, but the underlying observation is useful.

I am not bearish on equities. I am becoming more selective about which equities deserve capital.

The index is no longer a neutral decision

Passive investing is one of the greatest financial innovations ever created.

I don’t think indexing is dead.

I do think there are periods when the expected payoff from simply buying the index becomes less attractive relative to doing the work underneath it.

This may be one of them.

As of today, the ten largest companies represent roughly 38% of the S&P 500. The largest constituent alone — — represents almost 8%.

That does not make the index bad.

It means buying the index today is making a larger implicit bet on a relatively small group of companies than many investors realize.

And some of those companies are undergoing a fundamental change.

Technology companies are increasingly starting to look like industrial companies.

Not because software is disappearing, but because AI is incredibly capital intensive.

The AI revolution is creating enormous physical infrastructure requirements. Recent reporting suggests the financing needs are becoming large enough that even corporate bond investors are starting to differentiate between traditional issuers and AI-related borrowers.

This changes the question I want to ask.

The question is not: “Will AI change the world?”

I think it probably will.

The question for an investor is: “Will the return on the capital invested in AI justify the amount of capital required?”

Those are very different questions.

A technology can transform society and still produce disappointing returns for some of the investors financing it.

From a valuation bubble to an earnings question

This is where I think the debate around AI becomes more interesting.

Everyone talks about valuation.

I am increasingly interested in the possibility of an earnings bubble.

Not necessarily because current earnings are fake, but because today’s earnings expectations may implicitly assume that extraordinary capital spending will generate extraordinary future economic returns.

That assumption deserves to be tested.

Growth creates value when the return on incremental invested capital exceeds the cost of that capital. When ROIC falls below the cost of capital, more investment can actually destroy value. That principle is simple, but it becomes much more important when both the amount of investment and the cost of financing it increase.

Free cash flow matters for the same reason. Growth capex can be extremely valuable, but only if the investments ultimately create economic value. Spending more is not the same thing as becoming more valuable.

Who captures the economics: model providers, chip companies, hyperscalers, utilities, data centers, software companies, or ultimately the customer?

I don’t know the answers yet.

That uncertainty is exactly the point.

If the market price assumes a very favorable answer, the burden of proof becomes higher.

Graham understood this decades ago: the more growth a valuation requires, the more sensitive the investment becomes to small errors in the growth assumption.

Price is not separate from quality.

Price determines the odds.

Liquidity matters more than it used to

There is another variable I am spending more time studying: liquidity.

Modern financial markets depend not only on the level of rates, but on the availability and movement of balance-sheet capacity.

Michael Howell’s Capital Wars framework looks at the relationship between the stock of debt and global liquidity. His argument is that a heavily indebted financial system continuously needs liquidity because so much existing debt must be refinanced.

When liquidity grows faster than refinancing needs, financial assets tend to benefit.

When liquidity becomes scarce relative to debt, stress rises.

I don’t treat any single macro indicator as a trading system.

But the framework is useful.

It helps explain why liquidity can move markets even when the fundamental story has not changed much.

It also reinforces something I increasingly believe:

Cash is not an admission that I don’t know what to buy. Cash is optionality.

Its value is not just the yield I receive while holding it.

Its real value is the ability to deploy capital when future expected value becomes materially better.

If markets become more dispersed, that option becomes more valuable.

Gold, and the leverage inside the miners

Gold plays a different role.

It doesn’t need earnings growth to perform its portfolio function. It can benefit when confidence in fiscal discipline, monetary stability, or fiat purchasing power deteriorates.

That does not mean gold only goes up.

Cathie Wood currently argues almost the opposite of my base case: it sees technology-driven productivity, lower inflation, and potentially a stronger dollar creating conditions in which gold could fall materially.

That is a scenario worth respecting.

My interest in gold is not based on certainty that currencies will be debased. It is based on the payoff if that risk becomes more important.

Gold miners are a different bet.

They introduce operational, political, management, cost, and execution risk, but they can also provide operating leverage to the gold price. In Q1 2026, the World Gold Council reported that average gold-producer margins rose far faster than the gold price as higher realized prices outpaced rising mining costs.

So I think of the two exposures differently.

Gold is the hedge. Miners are the leveraged equity expression of the thesis.

Sizing should reflect that difference.

Crypto is resetting

Crypto sits at the other end of the portfolio.

Bitcoin has recently shown renewed strength after a long cooling-off period, reaching an eight-month high in September.

I am also watching improving traction beneath Bitcoin, particularly among higher-quality protocols already on our watchlist, including NEAR, Uniswap, Raydium, and Jupiter.

I don’t interpret this as permission to chase.

Quite the opposite.

My personal view is to use periods of short-term overheating to reduce risk or, selectively, express tactical downside through small leveraged short positions, while using meaningful weakness to accumulate crypto assets where I think the long-term asymmetry remains attractive.

The word daily matters. Leveraged inverse products can behave very differently from a simple long-term short because of daily resetting, path dependency, volatility, and compounding. For me, this is a tactical instrument, not a long-term holding.

Crypto itself remains a high-volatility allocation.

That means the thesis can be right and the position can still be wrong if the sizing is wrong.

I am not investing for a black swan

It is easy to construct a portfolio around everything that could go wrong:

• debt crisis;

• currency crisis;

• AI bubble;

• war;

• inflation;

• recession.

There is always another disaster to prepare for.

I don’t think that is a useful way to invest.

Black swans are, by definition, difficult to forecast. Building the entire portfolio around predicting one can be just another form of market timing.

I would rather think probabilistically.

The useful exercise is not pretending we know which future will happen, but mapping several plausible futures, assigning rough probabilities, and asking how our decision behaves across them.

I remain optimistic about technology, entrepreneurship, productivity, and long-term economic progress.

But optimism is not a valuation method.

And diversification does not mean owning fifty things that all require the same macro environment.

The optimal portfolio should not require low interest rates, record margins, expanding valuations, abundant liquidity, and exceptional AI growth to all happen at the same time.

It should have several ways to work.

Quality businesses that can compound capital.

Selective equities where expectations are low enough to create asymmetric outcomes.

Cash to give optionality.

Gold to provide a different monetary return driver.

Gold miners when the payoff justifies the additional operating risk.

Crypto where the upside remains large enough to justify volatility.

And, importantly, the ability to pass.

What matters is already in the price

I don’t know where the S&P 500 will trade next year.

I don’t know whether the 10-year Treasury yield will be 3% or 6%.

I don’t know where Bitcoin will finish this cycle.

And I don’t know whether today’s enormous AI investment cycle will produce extraordinary returns on capital or merely extraordinary capital expenditures.

Fortunately, I don’t think I need to know.

The objective of this project is not to predict the future exactly.

It is to understand what future the current price already requires.

Then compare the upside if reality is better with the downside if reality is worse.

The opportunity I see today is not necessarily in predicting the next bull or bear market.

It is in the widening gap between assets that need almost everything to go right and assets where very little success is already priced in.

Dispersion is the opportunity.

And if the world really has entered a regime where capital has a meaningful cost again, I expect that distinction to matter much more than it did during the last decade.

— Luca

Paychex —

0. Topics Covered

• Why We Are Watching It

• What It Does

• What We Like

• What Scares Us

• Quality Score

• Fair Value Range

• Scenario Range

• Buy Zone

• Margin of Safety

• Expected Value

• QPOS Score + View

• Final Takeaway

Let’s go.

1. Why We Are Watching It

We are watching Paychex because:

• it operates a recurring, mission-critical business with high switching costs;

• it has strong margins, strong cash generation, and an asset-light model;

• the business quality looks better than the current investment setup.

Paychex serves around 840,000 clients and says it pays roughly 1 in 11 U.S. private-sector workers. The model is built around payroll, HR, compliance, benefits, insurance, and PEO services.

2. What It Does

Paychex is a payroll and human-capital-management business serving mainly small and mid-sized companies.

It helps employers manage work they cannot afford to get wrong: payroll, taxes, HR, benefits, compliance, insurance, and employee administration.

It makes money through:

• recurring payroll and HCM software/services;

• HR outsourcing, PEO, and insurance;

• interest earned on client funds held before payroll and tax payments.

The main growth drivers are client growth, pricing, revenue per client, cross-sell, PEO worksite employees, interest rates on client funds, and acquisitions.

3. What We Like

The attractive parts:

• Durable moat: payroll and compliance are deeply embedded in customer workflows. Switching is possible, but it creates friction, execution risk, and administrative work.

• Strong recurring economics: Paychex benefits from recurring revenue, pricing power, high retention, and mission-critical services. The business does not need heavy physical reinvestment to grow.

• High cash generation: nine-month FY26 operating cash flow is about $2.0 billion while capex only about $169 million. That is the kind of capital-light structure we want to see.

• Strong margins and operating leverage: Q1 FY27 revenue increased 6%, operating income increased 14%, and adjusted EPS increased 10% to $1.34. Adjusted operating margin improved to 42.0% from 40.7%.

• Capital allocation: the company returns meaningful capital through dividends and buybacks. The issue is that the payout also limits the amount that can be reinvested organically, making pricing, cross-sell, and disciplined M&A more important to the long-term compounding story.

4. What Scares Us

The risks:

• Paying too much: this is the main risk today. Paychex can remain an excellent company while producing mediocre shareholder returns if the starting multiple is too high.

• Paycor integration: the ~$4.1 billion Paycor acquisition expands Paychex further upmarket, but it also introduces integration risk, additional debt, amortization, and the possibility that expected synergies disappoint.

• Moderate organic growth: Paychex is not a hyper-growth business. Management's FY27 guidance points to total revenue growth of 5–6% and adjusted EPS growth of 7–9%.

• Competition and AI: ADP, Paylocity, Intuit, Workday, and others compete across payroll and HCM. AI may help Paychex improve efficiency, but it may also lower barriers, improve competitor products, and increase pricing transparency.

5. Quality Score

Business Quality — 4.5/5

• Durable recurring revenue

• High switching costs

• Strong market position

• Mission-critical product

• Pricing power

Financial Quality — 4.5/5

• Strong cash conversion

• Capital-light model

• High margins

• Strong operating leverage

• Solid balance-sheet capacity, although Paycor increased leverage

Management Quality — 3.5/5

• Consistent execution

• Meaningful dividends and buybacks

• Disciplined operating cost management

• Paycor creates a new capital-allocation test

• Organic reinvestment runway is good, but not exceptional

Total Score

12.5/15

6. Fair Value Range

Our rough fair value range:

Read more

Baseline estimates now point to roughly $7.6 trillion of AI-related CapEx between 2026 and 2031, spanning compute, data centers, networking, and power infrastructure.

Annual spending alone could rise from roughly $765 billion in 2026 to more than $1.6 trillion by 2031.

Those numbers are impressive.

But as investors, the more interesting question is not how big they are.

It is what they mean.

Because the largest technology companies are no longer only selling software and scaling digital products at very low marginal costs.

They are buying GPUs.

Building data centers.

Securing energy.

Expanding networks.

And committing enormous amounts of capital to physical infrastructure.

AI is starting to look much more like an industrial business.

And that changes the investment equation.

Subscribe now

Growth is not enough

I recently watched this interview with Aswath Damodaran, and one idea stayed with me:

It sounds obvious.

Yet it is probably one of the mistakes investors make most often.

We find an extraordinary company.

A huge market.

A technology that could change the world.

Impressive growth.

And slowly, almost without noticing, we stop asking how much we are paying for it.

But a great company and a great investment are not the same thing.

That distinction matters even more in AI.

It is easy to say: “AI is going to be huge.”

It is much harder to say: “This specific company will create significant value for its shareholders from AI.”

Between those two statements sits almost everything that matters.

How much of the market can the company capture?

How much pricing power will it have?

What will margins look like?

How much competition will enter?

How much capital will be required?

And what return will the company earn on that capital?

A huge TAM is only the beginning of the analysis.

Not the end.

AI is becoming more capital intensive

For years, many of the world’s largest technology businesses had an attractive economic characteristic.

They could grow without needing an equivalent increase in physical capital.

Software is powerful partly because the next customer often costs very little to serve.

AI infrastructure changes that equation.

Training and running large models requires chips, servers, data centers, networking equipment, cooling systems, and huge amounts of electricity.

These are real assets.

And they are expensive.

That means investors may need to start thinking about some AI companies less like pure software businesses and more like capital-intensive industrial companies.

The question is no longer only: How fast can revenue grow?

It is also: How much capital needs to be invested to produce that growth?

That distinction matters.

A company can grow revenue quickly and still create very little shareholder value if generating that growth requires enormous amounts of incremental capital at mediocre returns.

Growth only creates value when the return on the capital invested to generate that growth is attractive.

This is why I think ROIC will become increasingly important in analyzing AI.

Revenue growth tells us how quickly the business is getting bigger.

ROIC tells us something more important: whether getting bigger is actually creating value.

The hidden question: how long will the assets last?

There is another issue.

Asset life.

A data center built today may remain physically usable for many years.

The accounting statements may depreciate the equipment over a relatively long period.

But economic life and accounting life are not always the same thing.

Especially in technology.

A GPU does not need to stop working to become economically obsolete.

A new generation of hardware may simply offer much better performance, lower energy consumption, or better economics.

If that happens faster than expected, companies may need to reinvest sooner than investors assume.

And that has major consequences.

Imagine a company spending tens of billions of dollars on AI infrastructure.

If those assets remain economically productive for a long time and generate high returns, that CapEx may create enormous value.

But if hardware needs constant replacement while competition pushes down prices and margins, the economics look very different.

This is why headline EPS may not tell us enough.

Even free cash flow needs context when growth CapEx becomes extremely large.

We need to understand the reinvestment requirements behind the growth.

Two companies can report similar revenue growth while producing completely different outcomes for shareholders.

One may require very little incremental capital.

The other may need to reinvest almost everything it earns just to remain competitive.

Same growth.

Very different economics.

You can be right about the trend and wrong about the investment

This is one of the most useful lessons in investing.

You can be completely right about the future of an industry and still lose money investing in it.

In 1999, you could have been right about the internet.

The internet did change the world.

Probably even more than many people expected.

That did not mean every internet company was a good investment.

The same can be true with AI.

AI can become one of the most important technologies of our lifetime.

The market can grow enormously.

Some companies can become much larger than they are today.

And at the same time, many companies can destroy capital.

These things are not contradictory.

The problem is that history makes the winners look obvious.

We see Amazon.

We forget the hundreds of companies that disappeared.

We see Nvidia’s past returns and think the opportunity should have been easy to recognize.

It rarely was.

That is survivorship bias.

We see the tree that became an oak but we forget all the seeds that never grew.

From regret to FOMO

Damodaran makes another interesting point.

Before FOMO — Fear Of Missing Out — there is often something else: ROMO: Regret Of Missing Out.

You watched Bitcoin rise without owning it.

You missed Nvidia.

You watched Amazon compound for years.

Then another exciting opportunity appears.

This time, you tell yourself, you will not miss it.

The internal logic changes from:

“What return can I reasonably expect at this price?”

to:

“I cannot afford to miss the next Amazon.”

That is a dangerous shift.

Because price slowly becomes secondary.

You are no longer buying because the expected value looks attractive.

You are buying insurance against future regret.

Those are very different decisions.

An investor should be able to say:

Not because the business is bad.

Not because the technology is unimportant.

But because the price may leave no margin for error.

The higher the expectations embedded in the price, the more things need to go right.

Stories matter. But they eventually need to become numbers.

Investors often fall into two camps.

One group looks almost entirely at numbers: multiples, margins, earnings, cash flow.

The other focuses heavily on the story: TAM, technology, visionary founders, adoption curves, future growth.

Both approaches can miss something important.

Numbers without a story can underestimate change.

Stories without numbers can justify almost anything.

A serious investment process needs both.

The story has to eventually become numbers.

Otherwise, it is still only a story.

This does not mean we need to predict the future perfectly.

We can’t.

The real purpose of valuation is not to calculate exactly what a stock will trade at five years from now.

It is to understand what the current price already assumes.

If today’s valuation requires a company to capture a huge percentage of a trillion-dollar market, maintain extraordinary margins, reinvest at high returns, and defend its position against aggressive competition for decades, that tells us something.

It tells us what we are betting on.

And then we can ask the question that matters: how reasonable are those expectations?

Value investing was never about low multiples

This is also why I think value investing needs to be understood correctly.

For decades, investors often associated “value” with a very specific kind of company:

• low P/E;

• low price-to-book;

• tangible assets;

• slow growth;

• simple businesses.

But those characteristics are not the underlying principle.

The principle is much simpler:

A bank trading below book value can be expensive such as technology company trading at a high multiple can be cheap.

It depends on the future cash flows, the risks involved, the capital required to generate them, and the expectations already reflected in the price.

Modern companies make this analysis harder because much of their economic value comes from things traditional accounting does not capture particularly well: software, research and development, intellectual property, networks, brands, data.

But the fundamental question has not changed.

The AI metrics I think matter more now

This is where the AI investment debate becomes particularly interesting.

Revenue growth will still matter.

Margins will still matter.

EPS will still matter.

But they may not be enough.

I increasingly want to understand four things.

1. Incremental ROIC

For every additional dollar invested into AI infrastructure, how much additional operating profit can eventually be generated?

2. Reinvestment intensity

How much capital must the company continually reinvest to maintain its competitive position?

3. Economic asset life

How long will today’s chips, servers, and data centers remain economically productive?

4. Expectations embedded in the price

How much future success are investors already paying for?

This brings everything back to the framework we use when studying investments:

• Quality: Is this a business worth owning?

• Price: Am I being paid for the risks I am taking?

• Odds: What has to happen for today’s valuation to work?

• Sizing: How much capital does the uncertainty justify?

AI does not make those questions obsolete. It makes them more important.

The hardest part: changing your mind

There is one final lesson from the Damodaran interview that matters more than any valuation model.

We will be wrong. Not occasionally. Regularly.

That is part of investing.

The real problem is not buying something that later falls.

The problem starts when our identity becomes attached to the decision.

We stop updating the thesis.

We defend the position.

We search for evidence that confirms what we already believe.

We ignore evidence that contradicts us.

Sometimes we double down simply because admitting a mistake has become psychologically expensive.

But the market does not care what we believed six months ago.

The only relevant question is:

AI will produce extraordinary companies. It will probably also produce extraordinary capital destruction.

Both can be true.

The job of the investor is not to predict which technology will change the world.

It is to understand the economics behind that change.

Applovin —

1. Why We Are Watching It

We are watching AppLovin because:

• The business is growing at a rate that is difficult to ignore.

• Its economics are strong: high margins, low capital intensity, and substantial free cash flow.

• The market debate around AI may be framing the company incorrectly.

The simple thesis:

AppLovin is not a software feature waiting to be replaced by AI. AI is already part of the product. The real question is whether its performance advantage can remain durable inside an advertising ecosystem controlled by much larger platforms.

The key question:

Can Axon — the AI-powered advertising engine that powers AppLovin Ads — keep producing meaningfully better advertising outcomes as competitors improve their own AI, data, and distribution?

2. What It Does

AppLovin is an advertising technology platform that uses AI to help advertisers acquire customers and publishers monetize their inventory.

The core engine is Axon, AppLovin’s AI-powered advertising system.

Its moat depends on performance: better data, better models, better matching, and better returns for advertisers.

If the model produces better returns, customers have a reason to keep spending. If it stops producing better returns, the moat becomes much less interesting.

3. What We Like

The attractive parts:

• Exceptional operating economics: Q2 2026 revenue reached $1.924 billion, up 53% year over year. Adjusted EBITDA reached $1.614 billion, up 58%.

• Very high margins: 84% Q2 Adjusted EBITDA margin. That is incredible profitability for a business still growing this quickly.

• AI appears to be an accelerator today: Q2 did not show the pattern we would expect from a product losing relevance: collapsing growth, pricing pressure, and margin compression. Instead, revenue and EBITDA continued to grow rapidly.

• The balance sheet is no longer the main risk: the latest update showed roughly $3.05 billion of cash against $3.52 billion of long-term debt.

• Per-share economics are getting support from buybacks: AppLovin repurchased 3.3 million shares for roughly $1.5 billion during the first half of 2026.

4. What Scares Us

The risks:

• Platform dependence: Apple and Google control important parts of the mobile ecosystem. Changes to privacy rules, identifiers, app distribution, or data access could reduce the quality of the signals AppLovin uses.

• The moat is performance-driven, not structural: advertisers can use several platforms at once. Switching costs exist, but they are not comparable to an ERP, payments network, or mission-critical database that is deeply embedded inside a customer’s organization.

• AI could commoditize part of the advantage: the risk is that Meta, Google, Amazon, or another advertising platform builds models that close the performance gap while controlling larger data sets or distribution networks.

• Expansion outside gaming still needs proof: e-commerce creates a large potential runway, but success in mobile gaming does not automatically guarantee the same competitive advantage in other advertising verticals.

• Regulatory and data-practice risk remain: a business dependent on data access and advertising infrastructure deserves a higher risk discount than the reported margins alone might suggest.

5. Quality Score

Business Quality — 4/5

• Strong market relevance

• Powerful data and model feedback loop

• Significant growth opportunities

• Very strong current competitive performance

Financial Quality — 5/5

• Very high margins

• Strong free cash flow

• Very low capital intensity

• High returns on capital

• Limited current balance-sheet stress

• Significant buybacks

Management Quality — 3/5

• Execution has been excellent

• Capital allocation has supported per-share value

• Buybacks still need price discipline

Total Score

12/15

6. Fair Value Range

Our rough fair value range:

Read more

The strategic thesis has not materially deteriorated.

Debt remains high. Money supply remains elevated. Central-bank demand is still strong. Gold continues to behave like a monetary asset in an environment where confidence in fiat currencies and fiscal discipline remains an open question.

But the investment setup has changed.

Higher real yields increase the opportunity cost of owning an asset that produces no cash flow. At the same time, the higher gold price leaves less room for error.

The result is a useful distinction:

That is where gold sits today.

1. The setup: structural strength, weaker asymmetry

The most important change since our previous review is not inflation, the dollar, or the Federal Reserve balance sheet.

It is real yields.

The U.S. 10-year Treasury yield has moved above 5%, while the 10-year real yield has reached 2.615%.

Historically, that is a difficult environment for gold.

Yet gold has not corrected materially.

That resilience is significant.

It suggests that structural monetary demand remains strong enough to offset part of the pressure coming from higher real rates.

Central banks remain an important part of that support. The World Gold Council data cited in our model shows 289 tonnes of central-bank purchases in Q2 2026.

But there is another side to the story.

Gold ETF flows were approximately -44.8 tonnes in Q2.

Financial demand is therefore not confirming a fully developed bullish scenario.

The picture is mixed: structural demand remains strong, but financial conditions have become less supportive.

2. The Gold Thesis

The strategic gold thesis remains intact.

The asset continues to provide exposure to several long-duration monetary risks: high sovereign debt, persistent fiscal deficits, elevated money supply, currency debasement risk and continued reserve diversification by central banks.

None of those variables has changed enough to invalidate the thesis.

What has changed is the tactical setup.

Gold is now more expensive while real yields are higher.

The underlying asset has not become worse.

The price-to-risk relationship has.

Mispriced Analysis

The market may still be underestimating the long-term consequences of fiscal stress and reserve diversification.

That remains the central bullish argument.

But today’s price already reflects more of that thesis.

Gold’s estimated market capitalization has increased from roughly $28.6 trillion to $30.0 trillion, while U.S. M2 has remained around $23.2 trillion.

That moves the Gold/M2 ratio from approximately 1.23x to 1.29x.

In other words, gold has become roughly 5% more expensive relative to M2 since the previous review.

At the same time, higher real yields reduce our fair-value assumptions.

Our updated scenarios are:

The base case remains above the current price.

But the asymmetry is no longer particularly attractive.

The probability-weighted expected value is approximately +5.5% in total.

Spread over a three- to five-year horizon, that falls to roughly 1%–2% annualized.

Under our framework, that is below the return hurdle required for new capital.

The important point is therefore not that gold is “overvalued.”

It is that the expected payoff at today’s price is too small relative to the uncertainty and downside.

Our View

• Thesis status: INTACT, tactically weakened

• Current price: ~$4,286

• Attractive entry zone: below ~$4,100

• Higher-conviction entry zone: below ~$3,750

• Base scenario: $4,750

• Bear / Base / Bull: $3,050 / $4,750 / $7,200

• Probabilities: 35% / 50% / 15%

• Sizing score: 2.25/5

3. What we are watching

The most important variable remains the U.S. real yield.

If real yields stop rising while gold remains resilient, that would strengthen the case that structural monetary demand is overpowering traditional rate pressure.

A decline in price below roughly $4,100 would also improve the expected-value profile.

Below approximately $3,750, the margin of safety would become materially more interesting, assuming the strategic thesis remained intact.

We would become more cautious if real yields moved sustainably toward 2.75%–3.00%, the dollar strengthened materially above current levels, central-bank purchases weakened, or persistent ETF outflows suggested that financial demand was deteriorating.

The thesis would also require reassessment if fiscal risk began to normalize in a credible and durable way.

Those are the variables that matter.

Most daily price movements do not.

4. Conclusion

Gold remains a strong monetary asset.

That does not automatically make it an attractive allocation at every price.

The strategic case remains intact: fiscal risk is elevated, central-bank demand remains strong and gold continues to hold up despite unusually high real yields.

But the market is no longer offering the same asymmetry:

• the price is higher

• real yields are higher

• the fair-value range is slightly lower

and the probability-weighted return is now below our hurdle for new capital.

Takeaways

• The long-term gold thesis remains intact.

• Higher real yields are the main tactical risk.

• Central-bank demand remains structurally supportive.

• ETF demand has not yet confirmed a stronger bull case.

• Gold has become more expensive relative to M2.

• The updated base case is approximately $4,750.

• Below $4,100, the setup becomes more interesting again.

Until next time,

Luca

The Shiller CAPE ratio reached 41, the highest level since the dot-com era, roughly the 99th percentile of the U.S. market’s historical valuation range, and more than double its long-term average of roughly 17x.

That does not mean the market is about to crash.

CAPE is not a timing tool.

But it does tell us something important: the price investors are paying for normalized earnings is unusually high.

And price matters.

One of the simplest lessons from value investing is that a great business can still be a poor investment if the price already assumes a great future.

High valuations leave less room for disappointment.

Lower valuations can create a margin of safety—but only if the underlying business is sound.

That second part matters.

Cheap alone is not enough.

Some stocks deserve low valuations because their businesses are deteriorating.

The better hunting ground is the intersection of quality and price: businesses with solid economics that the market may be valuing too pessimistically.

The S&P 500 is not “the market.”

It is a capitalization-weighted collection of companies. When its largest constituents command high valuations, the index can look expensive while individual companies elsewhere in the market trade at very different expectations.

So instead of asking: “Is the market expensive?”

A more useful question may be: “Where are expectations still low enough to create an asymmetric opportunity?”

Here are ten U.S.-listed companies from one of our current universe trading below 15x earnings.

Subscribe now

1. Deckers Outdoor —

P/E: 11.6x

Deckers owns a portfolio of footwear brands including HOKA, UGG and Teva. The company sells globally through wholesale, owned retail and direct-to-consumer channels.

The research question is whether current expectations underestimate the durability of its brands—or correctly anticipate a normalization in growth.

2. Abercrombie & Fitch —

P/E: 12.6x

Abercrombie & Fitch is a global apparel retailer operating online and through more than 750 stores across North America, Europe, Asia and the Middle East.

The company has gone through a significant brand and operating transformation.

The question now is less about the turnaround itself and more about durability: how much of the improvement in margins and brand relevance can persist through a normal retail cycle?

3. Cirrus Logic —

P/E: 14.7x

Cirrus Logic develops high-performance, low-power semiconductor solutions used in smartphones, wearables, laptops and other consumer devices. Its portfolio includes audio components, camera controllers, haptic technologies and other mixed-signal chips.

The key research issue is concentration and product dependence versus the value of Cirrus Logic’s intellectual property and ability to expand into additional applications.

4. H&R Block —

P/E: 7.9x

H&R Block is best known for tax preparation, but the business also includes digital tax products, small-business services through Block Advisors and Wave, and financial products.

The business operates in a mature category with recurring demand and a recognizable brand. The question is whether the market is underestimating the durability of that cash flow—or correctly pricing long-term disruption from software and changing consumer behavior.

5. Crocs —

P/E: 9.6x

Crocs is a global casual-footwear company built around one of the most recognizable products in the category.

The company reported record second-quarter 2026 revenue and said its core Crocs brand exceeded $1 billion in quarterly revenue for the first time.

Is Crocs a durable global consumer franchise temporarily priced like a fad?

Or is the low multiple correctly anticipating that today’s margins and demand will be difficult to sustain?

6. Hamilton Lane —

P/E: 14.6x

Hamilton Lane is an investment management firm focused on private markets.

Its business is built around providing institutions and other investors with access, portfolio construction and investment solutions across private-market strategies.

The interesting part is the business model.

Private markets have benefited from long-term growth in institutional allocations, but asset managers ultimately depend on fundraising, investment performance and client retention.

The research question is whether the valuation adequately compensates investors for those cyclicality and market risks.

7. Korn Ferry —

P/E: 14.6x

Korn Ferry is a global consulting firm working across leadership, hiring, compensation, organizational design and workforce strategy.

The company is often associated with executive search, but its business today is broader.

That matters because a more diversified consulting platform may deserve different economics than a pure recruiting company.

The key issue is cyclicality. Hiring and consulting budgets can weaken quickly when corporate confidence falls.

8. Kontoor Brands —

P/E: 14.0x

Kontoor Brands owns Wrangler, Lee and Helly Hansen, giving it exposure to denim, lifestyle apparel, workwear and outdoor clothing.

The research question is whether Kontoor can translate brand durability, direct-to-consumer expansion and international opportunities into consistent free cash flow without overpaying for growth.

9. Cal-Maine Foods —

P/E: 11.1x

Cal-Maine is the largest egg producer in the United States.

Its business spans conventional and specialty eggs, with a growing presence in prepared egg-based foods. The company also operates with a vertically integrated production network.

The obvious complication is cyclicality.

Egg prices, feed costs, disease outbreaks and supply conditions can cause earnings to move sharply.

That makes the headline P/E particularly important to interrogate: are today’s earnings representative of normalized earning power?

A low multiple on peak earnings is not necessarily cheap.

10. Stride —

P/E: 11.6x

Stride provides technology-enabled education services across K–12 education, career learning, professional training and talent development.

The company serves learners across all 50 U.S. states and internationally.

Fiscal 2026 revenue reached $2.52 billion, while operating income increased to $450.8 million.

The question is whether online and career-focused education can continue taking share—and whether Stride can convert that demand into durable economics while navigating regulatory and execution risks.

Cheap is the beginning of the research, not the conclusion

None of these companies belongs in a portfolio simply because its low P/E.

A low multiple tells us something about Price.

It tells us very little by itself about Quality, Odds or Sizing.

That is why we use the QPOS framework:

• Quality. Is this a business we actually want to own?

• Price. What expectations are already embedded in the valuation?

• Odds. What happens under bear, base and bull scenarios?

• Sizing. How much risk would the thesis deserve if it survives the research?

The goal is not to find the statistically cheapest stocks.

It is to find situations where the gap between price and underlying reality may be large enough to matter.

Especially when the broader index gives us very little margin for error.

Why The Mispriced Research exists

The Mispriced Research is built around one question:

We look for liquid opportunities where expectations may be wrong, research the downside before the upside, and invest with a long-term horizon.

We publish the portfolio and put real capital behind the research.

You make your own decisions.

We do the research.

If that approach is useful to you, you can follow the full research process with The Mispriced Pro.

Pro includes:

• Best Buys — 1/month: our monthly shortlist of the highest-ranked opportunities we are researching across quality, valuation and expected payoff.

• Research Notes — 1/week: one focused deep dive each week covering the thesis, valuation, scenarios, risks and what would change our mind.

• Portfolio Access: full access to holdings, weights, performance and portfolio history, plus an alert whenever we make a change.

Use code MISPRICED30 for 30% off your first year.

Get 30% off for 1 year

See you inside

Yelp —

1. Why We Are Watching It

We are watching Yelp because:

• It is a highly cash-generative, asset-light business trading at a depressed valuation.

• Its five-year average ROIC is around 16%, while free cash flow margins remain attractive.

• The market appears to be pricing in stagnation or structural decline, creating potentially interesting asymmetry if the franchise proves more durable than expected.

The simple thesis:

Yelp may be worth considerably more than the market price if its current cash generation survives the transition in local search.

The key question:

Is Yelp building a second growth engine, or simply using new products to offset the slow erosion of its core business?

2. What It Does

Yelp is a local discovery platform that connects consumers with local businesses through reviews, search, leads, and transactions.

It makes money mainly through:

• Performance-based advertising from local businesses

• Services-related lead generation

• Newer businesses including Hatch, Yelp Host, API and data licensing, and food-ordering products

Advertising still represents roughly 95% of revenue, so this remains primarily an advertising business today.

The newer products matter because Yelp needs another source of growth.

Hatch, AI tools, restaurant software, and data licensing could become that second leg.

But they have not proven it yet.

3. What We Like

The attractive parts:

• Strong cash generation: Yelp generated about $324 million of free cash flow in 2025 on roughly $1.47 billion of revenue.

  The business is highly asset-light, with gross margins around 90% and free cash flow margin around 20%.

• A real, although imperfect, moat: Yelp still owns a large review corpus, valuable local-intent data, a recognized consumer brand, and a marketplace connecting consumers with businesses. More reviews improve discovery. Better discovery attracts users. User intent creates value for advertisers.

  The network effect is real.

• Aggressive share reduction: management has used significant free cash flow for repurchases. Yelp spent roughly $292 million on buybacks in 2025, reducing the share count including equity awards by roughly 10%.

  At depressed prices, continued repurchases can materially increase owner earnings per share if the underlying cash flow remains durable.

• Founder alignment: Founder and CEO Jeremy Stoppelman owned roughly 5.5% of the company. Compensation also includes performance measures tied to revenue, adjusted EBITDA, and shareholder returns.

4. What Scares Us

The risks:

• Google and AI disintermediation: Yelp does not fully control its customer acquisition. Google remains an important source of web traffic, while Maps, AI-generated answers, and changing search behavior could reduce the need for users to visit Yelp directly.

• The core advertising engine is barely growing: in Q2 2026, advertising revenue declined 3%. Services was roughly flat, Restaurants, Retail & Other declined 10%, and total ad clicks fell 5%. Better monetization can offset lower traffic for a while but it cannot do it forever.

• Capital allocation is becoming more important: Yelp spent roughly $264 million on the Hatch acquisition during the first half of 2026 while also continuing large buybacks. Part of the transaction was financed through a revolving credit facility. The risk is not that Yelp cannot generate cash.

  The risk is that management uses that cash to buy growth at lower returns than the original business produced.

5. Quality Score

Business Quality — 3/5

• Real brand

• Proprietary review content

• Local-intent data

• Network effects

Financial Quality — 4/5

• Capital-light business

• High ROIC

• Healthy balance sheet

Management Quality — 2.5/5

• Founder ownership

• Share repurchases

• Reducing stock-based compensation

Total Score

9.5/15

6. Fair Value Range

Our rough fair value range:

Read more

Today, we’re adding 10 stocks to the investable Universe. Five of them are also being added to The Mispriced 25, our research index.

The Universe is the starting point of the research process: the set of assets that have passed a minimum threshold of acceptability and are worth following more closely.

What is the Universe?

The Universe is the investable opportunity set behind The Mispriced research.

It currently includes roughly ~250 assets, mostly public equities, alongside a smaller number of selected digital assets.

An asset entering the Universe does not mean it is cheap, attractive today, or suitable for the Index.

It means it has passed a basic filter.

The business, asset, liquidity, structure, and risk profile are good enough to justify further research.

From there, the harder questions begin:

• Is the quality high enough?

• What is already reflected in the price?

• Are the odds attractive?

• And, if the thesis is strong enough, how much risk does it deserve?

In other words: Quality. Price. Odds. Sizing.

The Universe gives us somewhere to look.

The Mispriced 25 is where select ideas earn a place.

🆕 New additions

1. Brookfield Asset Management

• Market Cap: $78.06B

• ROIC (5Yr Avg): 16%

• P/E: 28.4x

Brookfield Asset Management is one of the largest alternative asset managers globally, with exposure across infrastructure, real estate, private equity, renewable power, and credit.

The core attraction is the combination of a scaled asset-management platform, recurring fee-related earnings, long-duration capital, and exposure to secular growth in private markets.

Added to both the Universe and The Mispriced 25.

2. Nvidia Corporation

• Market Cap: $5.45T

• ROIC (5Yr Avg): 100.3%

• P/E: 28.6x

Nvidia has become one of the most important infrastructure businesses behind modern AI computing.

Its position extends beyond GPUs. The broader ecosystem includes networking, software, development tools, and an installed base that creates meaningful switching costs.

The business clearly passes the quality threshold. The harder question is price: how much future growth is already embedded in expectations?

Added to both the Universe and The Mispriced 25.

3. Cboe Global Markets

• Market Cap: $31.16B

• ROIC (5Yr Avg): 11.8%

• P/E: 23.3x

Cboe operates financial exchanges and market infrastructure across options, equities, futures, foreign exchange, and other markets.

It is an interesting picks-and-shovels business: rather than taking directional market exposure, Cboe earns from trading activity, data, access, and market infrastructure.

Added to the Universe for further research.

4. Crocs

• Market Cap: $5.48B

• ROIC (5Yr Avg): 32.2%

• P/E: 9.9x

Crocs is a consumer business built around an unusually recognizable product, strong brand economics, and an asset-light operating model.

The debate is less about whether the core Crocs brand has been successful and more about durability: how long can its economics persist, and how much skepticism is already reflected in the valuation?

Added to both the Universe and The Mispriced 25.

5. Cummins

• Market Cap: $77.61B

• ROIC (5Yr Avg): 12.9%

• P/E: 28.8x

Cummins is a global power technology company with deep positions in engines, power systems, components, and related technologies.

It combines an established industrial franchise with exposure to the long transition in transportation and power generation.

Added to the Universe.

6. Elmos Semiconductor $ELG

• Market Cap: €2.46B

• ROIC (5Yr Avg): 18.1%

• P/E: 24x

Elmos Semiconductor designs semiconductor products primarily for automotive applications.

It operates in specialized areas where reliability, engineering relationships, qualification cycles, and product integration can matter as much as scale.

Added to the Universe.

7. LeMaitre Vascular

• Market Cap: $1.82B

• ROIC (5Yr Avg): 15.8%

• P/E: 27.9x

LeMaitre Vascular develops medical devices used primarily in the treatment of peripheral vascular disease.

It is a relatively focused medical-device business operating in specialized niches, with a model that deserves closer study around margins, reinvestment opportunities, competitive positioning, and long-term runway.

Added to the Universe.

8. Yelp

• Market Cap: $1.14B

• ROIC (5Yr Avg): 16%

• P/E: 10x

Yelp remains a meaningful local-services discovery platform connecting consumers with businesses.

The interesting part of the thesis is not headline growth. It is the combination of an established platform, cash generation, capital returns, and the gap that may exist between the perception of the business and its underlying economics.

Added to both the Universe and The Mispriced 25.

9. Franklin Electric

• Market Cap: $4.38B

• ROIC (5Yr Avg): 14.7%

• P/E: 28.6x

Franklin Electric manufactures systems and components used to move water and fuel.

It operates in less visible parts of the industrial economy, with exposure to groundwater, water infrastructure, pumping systems, and related applications.

Not an obvious narrative stock. Potentially an interesting business.

Added to the Universe.

10. Ross Stores

• Market Cap: $73.25B

• ROIC (5Yr Avg): 23.6%

• P/E: 27.8x

Ross Stores is one of the largest off-price apparel and home-fashion retailers in the U.S.

The model is built around purchasing inventory opportunistically and offering branded products at discounted prices. Scale, buying capabilities, inventory discipline, and store economics make the business worth following.

Added to the Universe.

Universe vs. Index

These two decisions should not be confused.

Universe inclusion means: this is an asset we are willing to research and potentially own inside the framework.

Index inclusion means: given the current combination of quality, price, odds, and portfolio construction, the idea currently deserves scarce space inside The Mispriced 25.

The hurdle for the second decision is much higher.

With these changes, the research funnel gets a little broader.

But the goal stays the same:

• Find good assets.

• Understand what the market is pricing in.

• Think in probabilities.

• And only size the ideas where the asymmetry is strong enough.

Nobody wants it at $38.

That’s usually when things start to get interesting.

A falling stock is not automatically a cheap stock.

And a great brand is not automatically a great investment.

But at this price, Nike is worth studying.

The question is no longer whether Nike is a famous company.

It is whether the market has become too pessimistic about what the business can earn from here.

Why it’s worth watching

Nike is one of the strongest consumer brands in the world.

But revenue, margins, China, and its direct-to-consumer business have all moved in the wrong direction.

That combination — high business quality, weak recent fundamentals, and a much lower valuation — creates an interesting setup.

Not necessarily an opportunity yet.

But something worth watching closely.

What it does

Nike designs and sells athletic footwear, apparel, and equipment.

Footwear accounts for roughly two-thirds of revenue.

The business reaches consumers through both wholesale partners and Nike’s own stores and digital channels. In FY2026, NIKE Brand wholesale revenue was about $27.5 billion, while Direct generated roughly $17.7 billion.

The economic engine is simple:

• Nike invests in product innovation, athletes, storytelling, distribution, and brand.

• Consumers pay a premium because of what the Swoosh represents.

That brand has been built over decades.

The harder question is whether Nike can convert that brand equity back into consistent growth and higher free cash flow per share.

Why it may be mispriced

Nike’s stock has collapsed because the problems are real.

FY2026 revenue was about $46.4 billion, down from $51.4 billion in FY2024.

Revenue was roughly flat on a reported basis and down 2% currency-neutral. NIKE Direct fell 8% currency-neutral, while Greater China declined 13%.

Margins have also deteriorated.

Operating profitability is well below prior levels, while competitors such as On, Hoka, adidas, New Balance, and others have gained relevance in important categories.

The market is no longer pricing Nike like an untouchable compounder.

At roughly $38 against FY2026 diluted EPS of $2.10, the stock trades around 18x depressed earnings.

That is not obviously distressed.

But it is very different from the valuation investors were willing to pay when expectations were much higher.

The potential mispricing is therefore that the market may be pricing today’s operational problems too far into the future.

If Nike only needs to stabilize earnings and return to modest per-share growth, the current valuation becomes much more interesting.

What matters

Four things matter from here.

1. Revenue growth

The first signal is simple: can Nike return to positive currency-neutral growth?

A turnaround without organic growth is mostly financial engineering.

2. Direct and digital

NIKE Direct declined 8% currency-neutral in FY2026.

The company needs to prove that its owned channels can return to growth without relying heavily on promotions.

3. China

Greater China remains one of the clearest tests of the thesis.

Revenue there has fallen materially, and continued weakness would suggest the problem is more structural than cyclical.

4. Cash flow and returns on capital

FY2026 ROIC remained healthy at roughly 18.7%, but below prior levels, while estimated free cash flow was only around $2.2 billion.

The real confirmation would be a return toward $4–5 billion of free cash flow and ROIC above 20% without requiring aggressive discounting.

That would tell us the engine is actually recovering.

The risk

The biggest risk is not that Nike disappears.

It is that Nike remains a great brand but becomes a mediocre compounder.

Consumers have almost no switching costs.

Nike competes every day for attention against adidas, On, Hoka, New Balance, Anta, and others.

The failure case is therefore easy to imagine: the Swoosh remains globally recognized, but competitors keep taking share in the fastest-growing categories.

China never fully recovers.

Digital remains weak.

Nike regains volume through promotions rather than product desirability.

Margins stay structurally below historical levels.

In that world, today’s valuation may not be particularly cheap.

The brand survives.

The economics change.

That is the risk.

Where we stand

Status
Watchlist

Fair Value
$42–52 per share

Central Fair Value
~$47

Margin Of Safety
~15–20%

Buy Zone
≤ $32

Expected CAGR — Base Case
~14% over five years

Thesis
Nike does not need to return to its old growth rate for the current valuation to work.

The base case assumes roughly 7% EPS growth over five years, around $2.95 of FY2031 EPS, and a 22x terminal multiple. Including modeled dividends, that produces an estimated annualized return of roughly 14%. The key variable is whether Nike can restore normalized EPS and free cash flow per share while maintaining high returns on capital.

The bottom line

Nike is not interesting because it fell from $150 to $38.

Past prices tell us very little about intrinsic value.

Nike is interesting because the expectations embedded in the stock have changed dramatically while the underlying franchise still retains meaningful strengths.

Today, we have a powerful brand attached to a struggling operating engine.

That creates uncertainty, but it may also create mispricing.

For now, Nike belongs in The Universe because the gap between what the company has been and what the market now expects it to become is large enough to study.

Until next time,

Luca

Alfa Financial Software Holdings — ALFA

1. Why We Are Watching It

We are watching Alfa Financial Software Holdings because:

• Its software sits deep inside the operations of asset finance companies, creating meaningful switching costs.

• The business is becoming more recurring: H1 2026 ARR grew 17%, subscription revenue grew 14%, and subscription TCV grew 22%.

• The economics are strong: high margins, a capital-light model, strong historical cash generation, and no bank debt.

The simple thesis:

Alfa may be evolving from a high-quality software-and-services business into a more valuable recurring software compounder, without needing aggressive capital intensity to get there.

The key question:

Can subscription revenue become a larger part of the business while Alfa keeps its margins, cash conversion, and competitive position intact?

2. What It Does

Alfa Financial Software develops, implements, and supports software used by banks, OEM finance companies, and other asset finance businesses to manage loans, leases, and financing contracts.

Its core product, Alfa Systems, handles areas such as originations, servicing, and collections.

The platform is live across 37 countries.

It makes money through:

• Subscription revenue from SaaS and recurring services.

• Delivery revenue from implementation, upgrades, migrations, and project work.

• Software Engineering revenue from development work and some remaining perpetual license recognition.

H1 2026 revenue was £65.1 million:

• Subscription: £24.1m

• Delivery: £32.4m

• Software Engineering: £8.6m

3. What We Like

The attractive parts:

• Mission-critical product: Alfa Systems runs important operational infrastructure for large asset finance businesses. Replacing software like this can be expensive, disruptive, and risky. That creates real switching costs. The company also highlights long customer relationships and complex enterprise implementations.

• Recurring revenue is moving in the right direction: ARR reached £48.5m, up 17%. NRR was 110%. Subscription revenue increased 14%, while Subscription TCV grew 22%. This is the part of the story we want to see becoming more important over time.

• Very good economics: the earlier investment memo showed 2025 operating margin of 31.6%, FCF conversion of 97%, ROIC around 36.7%, and FCF margin around 29%. The business requires relatively little physical capital.

  High ROIC matters because a company that earns attractive returns on capital has a much better chance of compounding value when it can reinvest intelligently. Free cash flow matters because reported earnings alone do not tell us how much cash is actually available to owners.

• Strong balance sheet: Alfa ended H1 2026 with £22.2m of cash and no bank debt. That reduces financial fragility and gives management flexibility.

• Possible growth runway beyond the current core: management continues investing in US Auto Originations, Fleet, Commercial Finance, cloud adoption, and AI-enabled implementation tools. Alfa invested £19.6m in software development in H1 alone. AI may also be unusually interesting here. Alfa believes it can reduce implementation effort and shorten time to value. One internal migration pilot reportedly reduced effort on a significant task by roughly 75%.

4. What Scares Us

The risks:

• It is still not pure SaaS: delivery remains the largest revenue stream. Revenue and margins can therefore move with project timing, implementation intensity, and employee utilization.

• Execution and people risk: Alfa needs skilled people to implement complex enterprise systems. H1 employee engagement declined to 64% from 78% after product engineering departures. Poor retention or weak execution could slow delivery and future customer wins.

• Sales timing risk: Large enterprise contracts take time. In H1, some projects were delayed and one late-stage customer paused a project after management and internal priorities changed. Repeated delays over several periods would be a signal.

• Cash conversion and margin quality: H1 operating free cash flow conversion fell to 76% from 88%, while gross margin declined to 60.2% from 64.2%. Some of this was caused by severance, FX, working capital, and lower engineering chargeability, but it still deserves attention.

• Cyber risk: mission-critical software creates a moat when it works. It creates a serious liability when it does not. A meaningful security or availability failure could damage retention, reputation, and future sales.

5. Quality Score

Business Quality — 4/5

• Strong vertical specialization

• High switching costs

• Mission-critical product

• Long customer relationships

• Strong position in a niche market

Financial Quality — 4.5/5

• Strong historical free cash flow

• High ROIC

• High operating margins

• Net cash balance sheet

• Capital-light economics

Management / Network Quality — 4/5

• Historically disciplined capital allocation

• Product reinvestment continues

• No bank debt

• Management incentives include EPS and TSR

• Excess capital has historically been returned to shareholders

Total Score

12.5/15

6. Fair Value Range

Our rough fair value range:

Read more

A network can grow without its token becoming more valuable.

TVL can rise. Users can return. Stablecoins can expand. Transaction counts can remain high.

None of those facts automatically answer the question that matters to an investor:

How much of that economic activity actually accrues to the asset we own?

Our latest review of the crypto universe inside The Mispriced Research makes that distinction increasingly important.

Across Bitcoin, Ethereum, Solana and the DeFi protocols we follow, the underlying picture is not broadly bearish. In several cases, network fundamentals have improved.

But neither is it uniformly attractive.

In much of the universe, prices have recovered while fee generation has weakened. In others, TVL has expanded without an equivalent improvement in monetization. And in some cases the protocol itself looks substantially stronger than the economics of the token representing it.

That creates an unusual setup.

Crypto infrastructure looks healthier than token economics.

The opportunity, therefore, is no longer simply to own the protocols with the best activity metrics. It is to distinguish between network growth, economic capture and price.

That is the framework we use below.

Subscribe now

1. The market: stronger networks, weaker monetization

The most consistent signal across the current dataset is a divergence between capital/activity metrics and fee economics.

Ethereum is a clear example. TVL increased from roughly $41.2 billion to $47.8 billion, active addresses increased from 557,000 to 627,000, the staking ratio rose above 35%, and exchange reserves declined slightly. Yet 365-day fees fell from $244 million to $218.2 million. The ecosystem became healthier while direct economic capture remained weak.

Solana shows a related pattern. TVL, RWA exposure and monthly REV improved, but 365-day fees fell to $251.7 million while its market capitalization rose materially. Its market-cap-to-fees ratio expanded to roughly 227x. The network is proving adoption faster than it is proving sustainable monetization.

The same phenomenon appears deeper in DeFi.

Aave’s TVL increased 21% to $30 billion while fees fell 10.1%. Lido’s TVL increased 25.8% while fees fell 10.5%. Jupiter’s TVL increased while annual fees fell nearly 35%. Jito’s TVL reached roughly $1 billion while annual fees declined from $192.2 million to $82.2 million. Raydium’s TVL rose sharply while fees, active addresses and transactions weakened.

This does not imply that these protocols are failing.

It implies that TVL is not revenue, usage is not value accrual, and protocol quality is not automatically token quality.

That distinction matters especially after prices recover.

A high-quality asset bought at an unreasonable price can still produce a poor investment outcome. The same principle is useful here. Crypto requires an additional layer: even a high-quality protocol may have a weak token if the mechanism connecting economic activity to tokenholder value is incomplete.

2. Bitcoin

Thesis

Bitcoin’s cycle thesis remains intact.

Momentum has improved, miner economics have strengthened and the price has recovered above the 200-day moving average.

The issue here is price: Bitcoin moved from a favorable asymmetric setup toward a much thinner margin of safety.

Mispriced Analysis

The market may still be underestimating Bitcoin’s upside in a constructive cycle.

But the mispricing is smaller than it was at lower prices.

Our updated model places the floor around $52,900, the base value around $95,400 and the bull-cycle value around $163,600.

The asset may still have upside but the entry is no longer obviously asymmetric.

Our View

• Thesis status: INTACT

• Attractive entry zone: $53K–$68K

• Base scenario: $95,400

• QPOS Score: 52.6/100

3. Ethereum

Thesis

Ethereum’s network thesis remains intact.

Its token thesis is less convincing.

TVL, active addresses, staking and supply absorption improved. But fees continued to decline and direct value accrual remains weak.

This is the central Ethereum tension.

The ecosystem is working better than the token is monetizing it.

Mispriced Analysis

Our bullish interpretation is that the market may underestimate Ethereum’s role as settlement infrastructure, collateral and institutional-grade onchain plumbing.

The bearish interpretation is that investors may be overpaying for ecosystem activity that migrates economically toward L2s rather than ETH itself.

Both can be true simultaneously.

Our View

• Thesis status: INTACT

• Attractive entry zone: $1,750–$2,200

• Base scenario: $2,700–$3,700

• QPOS Score: 52.2/100

4. Solana

Thesis

Solana remains one of the strongest execution-layer assets in the crypto universe.

Usage is substantial. TVL improved. RWA activity improved. Monthly REV increased.

But annual fee generation does not yet justify the degree of valuation expansion.

Mispriced Analysis

The market appears to be pricing Solana on future monetization rather than current monetization.

That may ultimately be justified if its growing activity base produces materially higher recurring fees.

For now, the asymmetry is less compelling.

The network is better than it was in the previous review, but the price has risen enough that the improvement in fundamentals has not translated into a better expected-return profile.

Our View

• Thesis status: INTACT

• Attractive entry zone: $60–$85

• Base scenario: $115–$150

• QPOS Score: 56.7/100

5. Aave

Thesis

Aave remains one of the highest-quality DeFi protocol assets in the universe.

TVL increased to roughly $30B, active addresses improved, dilution remains limited, and Aave retains a strong position in onchain credit.

The weakness is valuation.

Price increased by more than 30% while annual fees declined by roughly 10%.

Aave has become stronger as a liquidity network and less attractive as an entry.

Mispriced Analysis

The market may still underestimate the long-term value of Aave as core DeFi credit infrastructure.

Its liquidity moat, scale and relatively clean dilution profile justify a quality premium.

But that premium is already being recognized. The mispricing today is therefore smaller than the protocol quality alone might suggest.

Our View

• Thesis status: INTACT

• Attractive entry zone: $70–$95

• Base scenario: $130–$200

• QPOS Score: 57.9/100

6. Lido

Thesis

Lido is one of the more interesting changes in the current review.

The protocol thesis strengthened.

The token thesis remains incomplete.

TVL increased substantially without a comparable increase in market capitalization.

That improved relative valuation.

Mispriced Analysis

The potential mispricing comes from Lido’s scale as staking middleware relative to its relatively small token valuation.

LDO still lacks sufficiently direct and recurring token value accrual.

The opportunity exists precisely because the market is applying a discount to that uncertainty.

If credible recurring accrual develops, the discount could narrow. If it does not, protocol success may continue to exceed tokenholder economics.

Our View

• Thesis status: INTACT

• Attractive entry zone: $0.28–$0.40

• Base scenario: $0.62

• QPOS Score: 60/100

7. Jito

Thesis

Jito remains intact, but monetization weakened materially.

This is another case where price moved enough to partially compensate for weaker economics.

Jito still matters inside the Solana ecosystem. The open question is how much of that relevance ultimately accrues to JTO.

Mispriced Analysis

Jito’s valuation improved as FDV declined and TVL increased.

The market may therefore be discounting the protocol too aggressively if Jito can stabilize fees and eventually improve JTO accrual.

Our View

• Thesis status: INTACT

• Attractive entry zone: $0.32–$0.48

• Base scenario:$0.60–$0.95

• QPOS Score: 58.8/100

8. Jupiter

Thesis

Jupiter remains a strong protocol with weaker monetization.

TVL increased to roughly $3.3B, active addresses remain resilient and transaction activity remains high.

But annual fees fell by approximately 35%.

The network thesis therefore remains intact.

The token thesis has become less convex.

Mispriced Analysis

The market may still undervalue Jupiter’s position within the Solana trading stack.

But monetization has become the limiting variable.

Usage and TVL remain strong enough to protect the protocol thesis.

They do not yet protect the token thesis.

Jupiter remains reasonably valued relative to TVL.

The problem appears when valuation is compared with monetization.

The updated base case was reduced because fee generation has weakened materially. The market is still assigning substantial value to Jupiter’s usage and strategic role in the Solana ecosystem, but current activity is not yet translating into equally strong token economics.

Our View

• Thesis status: INTACT

• Attractive entry zone: $0.15–$0.18

• Base scenario: $0.18–$0.32

• QPOS Score: 41.5/100

9. Raydium

Thesis

Raydium’s thesis weakened, but did not break.

The protocol gained TVL while losing meaningful momentum in fees and user activity.

Raydium is now less a fee-driven thesis and more a bet on the durability of its liquidity position and a future recovery in activity.

Mispriced Analysis

Raydium appears cheap relative to TVL.

But some of that cheapness may be justified if declining fees and activity indicate structural rather than cyclical weakness.

The token may be cheap relative to the capital sitting inside the protocol, but that capital is becoming less economically productive.

The mispricing only becomes compelling if fee and user activity stabilize.

Our View

• Thesis status: WEAKENED

• Attractive entry zone: $0.60–$0.72

• Base scenario: $0.85–$1.20

• QPOS Score: 39.6/100

10. Pendle

Thesis

Pendle’s thesis remains intact, but the current entry is less attractive.

TVL remains around $1B, active addresses improved and transactions increased modestly.

The negative signal is monetization.

Annual fees declined from $23.1M to $21.2M while the token price increased from $1.55 to $1.77.

The protocol has not deteriorated.

The price-to-fundamentals relationship has.

Mispriced Analysis

Pendle still has meaningful optionality if onchain yield markets continue to deepen.

It also has something many crypto protocols lack: real fee generation.

But the current valuation already reflects more of that potential.

Market cap/fees increased from 11.5x to 14.4x and FDV/fees from 18.9x to 23.5x.

The expected-value model remains positive, but the upside declined materially from the previous review.

This is still an interesting protocol.

It is no longer an obvious valuation dislocation.

Our View

• Thesis status: INTACT

• Attractive entry zone: $1.15–$1.55

• Base scenario: $1.95–$3.20

• QPOS Score: 60.6/100

11. Uniswap

Thesis

Uniswap’s protocol thesis remains intact and has slightly strengthened.

TVL increased from $4.6B to $5.8B, transaction count rose from 222.6M to 267.3M, and active addresses improved to 119.1K. The protocol continues to show strong usage, deep liquidity, and a durable competitive position.

The investment thesis is weaker at the current price. The market is paying substantially more for a protocol whose usage improved, but whose monetization did not.

Mispriced Analysis

The core mispricing is no longer that Uniswap looks obviously cheap.

At $5.88, UNI is already trading close to its base fair-value range.

Market cap/TVL increased from 0.52x to 0.64x, while market cap/fees rose from 3.33x to 5.33x. FDV/fees increased from 4.74x to 7.44x. UNI is therefore no longer cheap on gross fees, and the current price increasingly depends on future fee capture, burn, or another credible value-accrual mechanism.

The most important variable is token capture.

At a 10% fee-capture rate, UNI would trade at roughly 53x market cap to captured value and 74x FDV to captured value. Even at 15%, those multiples remain demanding. This means the current valuation requires either stronger growth or materially better token economics.

Our View

• Thesis status: WEAKENED

• Attractive entry zone: $3.25–$4.75

• Base scenario: $5.00–$6.75

• QPOS Score: 41.5/100

Conclusion

Network fundamentals are often improving faster than token economics.

The market appears to be rewarding network relevance before proving economic capture.

Takeaways

• The crypto ecosystem is healthier than aggregate token monetization suggests. TVL and usage remain resilient across much of the universe.

• Token value accrual is the central research question. Protocol success and token success cannot be treated as synonymous.

• BTC remains structurally strong, but its margin of safety has compressed materially after the rebound.

• ETH and SOL have improving network fundamentals but weak current monetization relative to valuation.

• AAVE remains one of the highest-quality DeFi assets, but price appreciation has reduced the asymmetry.

• LDO currently shows one of the more interesting valuation/fundamental divergences: TVL increased substantially while market cap barely moved.

• JTO also offers improved valuation, but its fee collapse prevents a high-conviction position.

• JUP remains strong operationally but weak on incremental monetization.

• RAY deserves the most caution among the reviewed DeFi names because fees, users and transactions weakened together despite higher TVL.

• PENDLE remains intact, but the price has moved from clearly attractive toward a more balanced risk/reward.

• UNI remains one of the stronger protocols operationally, but the investment setup has become much less asymmetric.

Until next time,

Luca

The goal is not to explain every market move but to keep the portfolio process visible, consistent, and accountable.

I run a concentrated portfolio built from the same research process I publish here.

I invest my own capital alongside the research, and I usually keep the portfolio focused — typically around 15 positions.

Research Goal

Building a concentrated portfolio of liquid opportunities that can compound capital at attractive rates over the long term.

I look for a small number of positions where quality, valuation, expected return, and downside line up better than average.

The long-term target to build a process that can aim for 15%+ annualized returns over time.

Portfolio Snapshot

Since Nov 2025

• True Time-Weighted Rate of Return: +36.4%

• Return on Invested Capital: +15.3%

• Realized Gains: +10.6%

• IRR: +96.2%

• Dividend Contribution: +1.05%

August Moves

BUY

• NXP Semiconductors — $221.94

• Deckers Outdoor — $85.60

• Wavestone — €33.60

• Cirrus Logic — $108.66

ADD

• The Trade Desk — $13.50

• NAPCO Security — $35.28

• Nextpower — $89.59

• BTCS — $1.45

TRIM

• ResMed — $234.18

EXIT

• Sanlorenzo — €38.44 — +29.0%

• Anima Holding — €7.16 — +17.9%

• Halozyme — $104.87 — +50.8%

Current Portfolio

15 holdings

Read more

The Best Buys come from The Mispriced Research Index — our internal ranking system used to compare investment opportunities across the same framework.

Subscribe now

Best Buys — August 2026

Every month, we narrow the output of our research system down to the opportunities that stand out most across three different lenses:

• 🏆 High-Ranked — the strongest overall opportunities in the Research Index

• 🏷️ Undervalued — the widest gaps between market price and fair value

• 🌱 Compounders — the most attractive expected value

🏆 High-Ranked

These are the names sitting near the top of our internal ranking system:

Read more

Regal Rexnord Corporation — RRX -4.27%↓

1. Why We Are Watching It

We are watching Regal Rexnord because:

• The business has been transformed into a higher-margin industrial platform.

• It has exposure to attractive markets including automation, data centers, aerospace, defense, and industrial efficiency.

• The stock has moved into a valuation range where the downside/upside equation is becoming more interesting.

The original transformation is meaningful. From roughly 2018 to 2025, gross margin increased from about 27% to 38%, adjusted EBITDA margin moved from about 15% to 22%, and free cash flow grew from roughly $285 million to $520 million.

More recently, demand has improved. Q2 daily orders increased 8.8%, organic sales grew 3.3%, and the Automation & Motion Control segment grew organically by 15.6%, supported by data centers, discrete automation, and aerospace & defense

The simple thesis:

Regal Rexnord could become a high-quality industrial compounder if stronger end-market exposure, margin improvement, and integration benefits translate into durable free cash flow while the balance sheet continues to improve.

The key question:

Can Regal Rexnord turn its transformation into sustainable free cash flow growth without leverage, cyclicality, and execution risk consuming too much of the upside?

2. What It Does

Regal Rexnord is an industrial technology company focused on moving, controlling, and transmitting power.

Its products include motors, bearings, gearing, couplings, brakes, drives, actuators, air-moving systems, and other motion-control components.

It sells to OEMs, distributors, and industrial customers across markets including:

• Automation

• HVAC

• Data centers

• Aerospace and defense

• Medical

• Energy

• Mining

• General industrial

Customers are not simply buying components.

They are often paying for reliability, energy efficiency, uptime, customization, and lower operating risk.

That matters because many of these components sit inside mission-critical applications.

3. What We Like

The attractive parts:

• A real, if imperfect, industrial moat: switching costs can be meaningful in integrated OEM applications, the products are often mission-critical, and Regal Rexnord benefits from technical know-how, customer relationships, scale, service, and supply-chain capabilities.

• Improving business quality: the company has materially improved its margin profile through its transformation: 38% gross margin, 22% adjusted EBITDA margin, and around 9% adjusted free cash flow margin.

• Exposure to better structural growth markets: roughly half of sales are exposed to secular markets, with automation, aerospace and defense, medical, and data centers among the important growth areas. Recent order growth gives us more evidence that these exposures are becoming meaningful.

• A clearer path to cash generation: management currently guides to around $600 million of 2026 free cash flow, or roughly $9 per share. The more important question is whether normalized FCF per share can eventually reach the estimated $10.80–$13.00 range.

At $169.49 the margin of safety has improved materially.

What we’re really looking for:

A durable asset with attractive economics, clear risks, and a price that gives us room to be wrong.

4. What Scares Us

The risks:

• Leverage: net debt to adjusted EBITDA, including synergies, is around 3.06x. It is moving in the right direction, but debt still limits financial flexibility. Until leverage is clearly below 3x, aggressive buybacks or additional large acquisitions would make the story less attractive.

• Free cash flow quality: 2026 FCF guidance is around $600 million, but working capital has become a bigger drag. The business needs to prove that cash conversion improves rather than remaining an annual adjustment story.

• Cyclicality: this is still an industrial company. Residential HVAC, pool-related demand, general industrial activity, and other short-cycle markets can weaken. A better portfolio does not remove the cycle.

• Execution risk: the thesis still depends on productivity improvements, acquisition synergies, favorable mix, integration, and continued growth in higher-quality verticals.

5. Quality Score

Business Quality — 4/5

• Good market position

• Mission-critical products

• Meaningful OEM relationships and switching costs

• Better exposure to secular growth markets

• But no network effects or strong recurring-revenue model

Financial Quality — 3/5

• Improving margins

• Good underlying cash-generation potential

• Attractive reported ROIC, though acquisition accounting makes normalization important

• FCF conversion still needs proof

• Leverage remains meaningful

Management / Execution Quality — 3/5

• Significant business transformation has been executed

• Margin improvement has been real

• Deleveraging is currently the right capital-allocation priority

• Integration still needs to prove itself

• Insider alignment is not particularly strong

Total Score

10/15

6. Fair Value Range

Our rough fair value range:

Read more

This series is where we study one asset at a time, define the risk, and decide what price would make the bet worth taking.

Subscribe now

This is not financial advice.

This week’s Research Note is about:

Cirrus Logic — CRUS -1.65%↓

1. Why We Are Watching It

We are watching Cirrus Logic because:

• The business generates strong margins and free cash flow.

• Returns on capital are close to 20%.

• A broader product pipeline could reduce its dependence on smartphones over time.

The simple thesis:

Cirrus Logic is a high-quality semiconductor business priced below our estimated value, but its dependence on one customer prevents us from treating it like a fully durable compounder.

The key question:

Can Cirrus build meaningful revenue streams outside its largest customer before that concentration becomes a problem?

2. What It Does

Cirrus Logic designs low-power mixed-signal chips used primarily in smartphones, laptops, and other electronic devices.

Its products help customers improve:

• Audio performance

• Power efficiency

• Camera control

• Haptics and user interaction

• Hardware and software integration

It makes money by selling proprietary components to original equipment manufacturers and contract manufacturers.

Its main product areas include:

• Smartphone audio chips

• High-performance mixed-signal products

• Camera controllers

• Power-management chips

• PC and laptop audio solutions

Cirrus is a fabless semiconductor company. It designs the chips but relies on outside manufacturers to produce them. Its value comes from engineering knowledge, intellectual property, integration, and its ability to help customers bring better products to market.

The company has thousands of patents and technical switching costs once its products are designed into a device. But it does not have network effects or contractually recurring revenue.

3. What We Like

The attractive parts:

• Strong cash generation: Cirrus converts a large share of its earnings into cash. Its free cash flow margin has historically been unusually strong for a semiconductor business.

• High returns on capital: five-year average ROIC has been around 19%–20%. That suggests the company has been able to generate attractive profits from the capital required to operate.

• Clean balance sheet: Cirrus holds net cash rather than net debt. That reduces financial risk and gives management flexibility during weaker industry periods.

• Asset-light economics: the fabless model requires limited capital expenditure. Recent analysis estimates capex at roughly 1% of sales, making normalized earnings a reasonable proxy for owner earnings.

• A more credible reinvestment pipeline: potential growth areas now include camera controllers, smart power chips, AI PC audio products, analog front ends for smart metering, and longer-term opportunities in data-center power, grid infrastructure, and electric-vehicle charging.

• Disciplined repurchases: management has regularly used free cash flow to reduce the share count. The board added a new $500 million repurchase authorization in 2025 after an earlier $500 million authorization.

The company’s operating margin is around 23%, with gross margins above 50%. Its economics look like those of a quality compounder. The unresolved question is whether those economics can survive meaningful customer diversification.

What we’re really looking for:

A durable business with attractive economics, clear risks, and a price that gives us room to be wrong.

4. What Scares Us

The risks:

• Extreme customer concentration: Approximately 90% of revenue comes from one customer. A change in product strategy, sourcing, pricing, or internal chip development could materially affect Cirrus.

• Limited bargaining power: a dominant customer can negotiate aggressively, add a second supplier, reduce Cirrus’s content per device, or move parts of the design in-house.

• Diversification remains unproven: PC, automotive, industrial, and general-market products are promising, but they remain too small to offset a major loss in smartphone revenue.

• Fabless supply-chain risk: Cirrus depends on third parties for manufacturing capacity. Supply constraints, production delays, or unfavorable wafer commitments could affect revenue and margins.

5. Quality Score

Business Quality — 3/5

• Strong position in specialized mixed-signal chips

• Valuable engineering knowledge and intellectual property

• Moderate technical switching costs after a design win

• No network effects

• Durability weakened by customer concentration

Financial Quality — 4/5

• Strong free cash flow

• ROIC near 20%

• Net-cash balance sheet

• Attractive margin structure

• Limited capital intensity

Management / Execution Quality — 2/5

• Disciplined share repurchases

• Conservative balance sheet

• Good execution in core products

• Broader product pipeline is developing

• Diversification has not yet produced meaningful revenue independence

Total Score

9/15

6. Fair Value Range

Our rough fair value range:

Read more

This series is where we study one asset at a time, define the risk, and decide what price would make the bet worth taking.

Subscribe now

This week’s Research Note is about:

ExlService Holdings — EXLS 2.15%↑

1. Why We Are Watching It

We are watching ExlService Holdings because:

• It combines data, analytics, AI, and operational services inside complex enterprise workflows.

• More than 75% of revenue is recurring, with net revenue retention above 111%.

• The business has improved its ROIC, margins, and cash generation while maintaining a relatively light capital base.

• Enterprise AI could create a long reinvestment runway, especially in regulated industries where data quality, context, and trust matter.

The simple thesis:

EXL could become a durable compounder by helping large companies move AI from experimentation into mission-critical production.

The key question:

Can EXL turn strong revenue growth into durable free cash flow per share without allowing stock-based compensation, acquisitions, or AI commoditization to weaken the economics?

2. What It Does

ExlService Holdings helps large companies use data and AI to make decisions, automate complex processes, and modernize important workflows.

Its main customers operate in areas such as insurance, healthcare, and banking.

EXL makes money through:

• Data, analytics, and AI services

• Ongoing management of complex business processes

• Proprietary platforms, tools, and intellectual property

• Expansion and cross-selling within existing customer relationships

More than three quarters of revenue is recurring.

3. What We Like

The attractive parts:

• Mission-critical customer relationships: EXL is integrated into workflows that clients cannot change casually. Its industry knowledge and position inside regulated processes create meaningful switching costs.

• Strong recurring economics: more than 75% of revenue is recurring. Net revenue retention is above 111%, suggesting that existing customers tend to expand their spending over time. The company’s average net promoter score over the past five years has also been above 80.

• High returns on capital: reported ROIC increased from 8.9% in 2020 to 20.2% in 2025. That matters because growth creates the most value when a company can reinvest at attractive returns.

• Strong cash conversion: EXL generated approximately $298 million of free cash flow in 2025, equal to 119% of net income. Capital expenditures were only $52.6 million, which supports the view that this is a relatively asset-light business.

• A credible enterprise AI position: EXL works at the intersection of industry data, domain knowledge, operational execution, and AI. That combination may be more durable than selling generic AI services alone.

What we’re really looking for:

A durable business with attractive economics, clear risks, and a price that gives us room to be wrong.

4. What Scares Us

The risks:

• AI could weaken the service model: AI is both the opportunity and the threat. More accessible models could automate labor-intensive work faster than EXL can replace it with higher-value platforms and intellectual property.

• Stock-based compensation is rising: second-quarter stock-based compensation increased from $16.4 million to $24.6 million, or roughly 50%. Adjusted earnings exclude this cost, but shareholders still bear its economic effect.

• Revenue growth is not fully reaching GAAP earnings: second-quarter revenue grew 15.6%, but GAAP EPS increased only from $0.40 to $0.42. GAAP operating margin declined from 15.8% to 14.7%. The top-line result was strong. The per-share economic result was less convincing.

• The iMerit acquisition adds execution risk: the acquisition could cost up to $310 million, including upfront and future consideration. It may strengthen EXL’s AI capabilities, but the company still has to prove the integration, talent retention, and returns on invested capital.

• The moat is meaningful, but not absolute: EXL has high switching costs, recurring revenue, and strong vertical expertise. It does not have a monopoly, a dominant network effect, or an irreplaceable technology platform. Large consultants, IT service providers, and AI specialists remain credible competitors.

5. Quality Score

Business Quality — 4.5/5

• High recurring revenue

• Strong customer retention

• Meaningful switching costs

• Mission-critical workflows

• Attractive positions in regulated industries

Financial Quality — 4.5/5

• ROIC above 20% in 2025

• Strong historical free cash flow conversion

• Low capital intensity

• Moderate leverage

• Good reinvestment potential

Management / Execution Quality — 4/5

• Strong historical operating execution

• Organic investment increased substantially

• Capital allocation has supported higher ROIC

• The balance sheet has remained flexible

• Share count declined in the latest quarter

Total Score

13/15

6. Fair Value Range

Our rough fair value range:

Read more

It is liquid.
It is scarce.
It has a growing institutional market structure.
It offers a return stream that is different from traditional businesses and bonds.

But an interesting asset is not automatically an index candidate.

The question is not whether Bitcoin could rise.

The question is whether the current price offers a sufficiently attractive combination of quality, price, odds, and sizing.

Today, the answer is not clear enough.

That is why Bitcoin is not in The Mispriced 25 right now.

The thesis is not broken

Bitcoin’s long-term case remains credible.

Its monetary policy is transparent. The network continues to operate without a central issuer. Mining difficulty increased over the past year, while supply held by long-term holders reached an all-time high of roughly 14.85 million BTC during the second quarter.

That matters.

Long-term holders added approximately 313,000 BTC during a quarter in which the price declined by around 14%.

Stronger holders appear to be absorbing supply from shorter-term participants.

This is constructive.

It is not enough on its own.

The current setup is mixed

Bitcoin closed the second quarter near $58,544.

At that level, it was below three important reference points:

• Short-term-holder realized price: approximately $70,327

• 200-day moving average: approximately $75,371

• On-chain mean: approximately $76,660

These levels had shifted from support to resistance. Historically, that configuration has been associated with weaker market environments.

At the same time, Bitcoin had not fully returned to the lower cost bases that have often marked more complete resets.

The ArkInvest Report identified realized price near $53,000 and investor price near $49,000. Bitcoin was trading above both.

This leaves an uncomfortable middle ground.

The price has fallen enough to damage momentum.

But it may not have fallen far enough to create a clear margin of safety.

QPOS: The current Bitcoin assessment

Quality

Bitcoin’s network quality remains relatively strong.

The supply schedule is predictable. Liquidity is substantial. Institutional access has improved. Long-term-holder behavior is constructive, and realized volatility has remained near multi-year lows despite the decline.

But quality is not uniformly improving.

Active ownership and transaction volume weakened. Miner revenue was under pressure. The broader market still depends heavily on confidence, liquidity, custody infrastructure, and regulation.

Bitcoin may be a high-quality monetary network.

It is not a stable cash-generating business.

That distinction matters.

Price

Bitcoin is difficult to value because it does not produce cash flow.

There is no earnings yield, free-cash-flow model, or book value to anchor the analysis.

Price must be assessed through adoption, scarcity, network activity, positioning, liquidity, cost bases, and the expectations already reflected in the market.

At roughly $58,500, the price was below major trend levels but still above the $49,000–$53,000 area associated with deeper historical resets.

That does not make Bitcoin expensive.

It makes the margin of safety unclear.

Odds

The evidence supports several competing outcomes.

Bear case

ETF outflows continue. Treasury companies face higher financing costs. Bitcoin fails to reclaim its major means and moves toward, or below, the $49,000–$53,000 cost-basis range.

Base case

Seller exhaustion limits the downside, but weak marginal demand keeps Bitcoin in a broad consolidation. Price remains volatile without establishing a durable trend.

Bull case

Long-term holders continue absorbing supply. ETF flows recover. Bitcoin reclaims its short-term-holder cost basis and major trend levels, confirming that the correction was a reset rather than the beginning of a longer contraction.

The current expected value may be positive over a long horizon.

The near-term distribution remains too wide for a high-conviction index position.

Sizing

Bitcoin’s volatility is lower than in previous cycles, but it remains high relative to most assets considered for The Mispriced 25.

It also carries risks that are difficult to model:

• Regulatory change

• Custody and operational failures

• Market-structure stress

• Leverage embedded in treasury companies

• Dependence on marginal liquidity

• Technological and security risks

A Bitcoin position would therefore need to be smaller than a typical index position unless the entry price and thesis strength improved materially.

Sizing cannot repair a weak setup.

It can only limit the damage if the thesis is wrong.

What would put Bitcoin back into consideration?

There are two credible paths.

1. A better price

Bitcoin trades into or below the $49,000–$53,000 cost-basis region, selling pressure becomes exhausted, and the network remains healthy.

This would not create an automatic index entry.

It could create a more attractive asymmetry: lower expectations, clearer downside reference points, and a larger potential margin of safety.

2. Better confirmation

Bitcoin stabilizes and reclaims key levels between roughly $70,000 and $77,000.

At the same time:

• ETF flows turn consistently positive

• Treasury-company financing stress declines

• Network activity improves

• Long-term-holder supply remains firm

• The recovery is not driven by excessive leverage

This would mean paying a higher price.

But it could improve the odds that the trend and demand structure have genuinely repaired.

The cheapest entry is not always the best entry.

The goal is not to capture the exact bottom.

The goal is to enter when the relationship between downside, upside, and evidence becomes attractive.

The decision

Status: Watchlist.

Bitcoin is not excluded because the long-term thesis has failed.

It is excluded because the present setup does not yet offer enough clarity across price, odds, and sizing.

That could change quickly.

Until then, doing nothing is also a decision.

The Mispriced 25 is not designed to own every important asset.

It is designed to hold 25 assets where the evidence, expectations, and downside create a sufficiently attractive research case.

Bitcoin may qualify again.

It does not qualify today.

This series is where we study one asset at a time, define the risk, and decide what price would make the bet worth taking.

Subscribe now

This is not financial advice.

This week’s Research Note is about:

Adyen — $ADYEN

1. Why We Are Watching It

We are watching Adyen because:

• It is one of the highest-quality payment businesses in Europe.

• It has strong cash generation, high margins, and no material financial debt.

• It still has a credible runway across enterprise payments, Unified Commerce, Platforms, and financial products.

The simple thesis:

Adyen is a high-quality payments compounder that can keep growing owner earnings per share at attractive rates if it protects its take rate, margins, and customer relevance.

The key question:

Can Adyen keep compounding owner earnings per share at mid-teens rates without giving up pricing power or destroying capital through acquisitions?

2. What It Does

Adyen is a global payments platform for large merchants and platforms.

It helps companies accept payments online, in stores, and across countries through one infrastructure layer.

It makes money through:

• Processing and settlement fees

• Acquiring services

• Risk management

• Issuing and financial products

• Payment terminals and related services

The customer pays Adyen because payments are mission critical.

3. What We Like

The attractive parts:

• Mission-critical infrastructure: Adyen sits inside the payment flow of large merchants. If the system fails, transactions fail. That creates real switching costs.

• Strong economics: the business produces high EBITDA margins, strong free cash flow, and needs limited capital to grow.

• Global single-platform model: Adyen’s value is simplicity. Large merchants can use one integration across countries, channels, and payment methods.

• Long growth runway: Unified Commerce, Platforms, geographic expansion, and financial products can keep expanding the revenue base.

What we’re really looking for:

A durable asset with attractive economics, clear risks, and a price that gives us room to be wrong.

4. What Scares Us

The risks:

• Growth disappointment: the stock still prices in strong compounding. If growth slows faster than expected, the valuation can compress.

• Take rate pressure: payments is a competitive market. If Adyen has to give up economics to keep volume, owner earnings growth could disappoint.

• Margin pressure: Adyen is investing in people, products, and expansion. That is fine if it creates future earnings. It is a problem if returns fade.

• Capital allocation risk: the Talon.One acquisition is a meaningful deal. The strategic logic may be sound, but the return on capital is not yet proven.

5. Quality Score

Business Quality — 5/5

• Durable business

• Strong enterprise positioning

• High switching costs

• Mission-critical product

• Global scale

Financial Quality — 5/5

• High margins

• Strong cash generation

• Low capital intensity

• No material financial debt

• Attractive reinvestment potential

Management / Execution Quality — 3/5

• Strong execution history

• Founder-led culture

• Long-term orientation

• But Talon.One increases capital allocation risk

Total Score

13/15

6. Fair Value Range

Our rough fair value range:

Read more

They fail because they lack a decision system.

Markets are noisy. Prices move. Narratives change. Fear and greed take turns pretending to be insight. In that environment, the investor’s first job is not to predict every outcome. It is to protect the quality of the process.

Benjamin Graham’s core lesson was not simply “buy cheap stocks.” It was deeper than that. Successful investing requires a sound theoretical framework for making decisions, and the emotional discipline to keep that framework from being corroded by market behavior.

The more irrational the market becomes, the more opportunity it may offer to the prepared investor.

That is the philosophy behind The Mispriced.

Research indexes built to act as a decision machine.

Why a decision machine matters

Every investment decision is a bet made with incomplete information.

We do not know the future. We do not control the outcome. We can make a good decision and get a bad result. We can make a poor decision and get lucky.

Annie Duke calls this the problem of separating decision quality from outcome quality. In uncertain environments, the result is only partial feedback.

That is exactly why process matters.

A research index forces the question before emotion takes over:

What did we believe?

Why did we believe it?

What price were we paying?

What could go wrong?

How much risk did the idea deserve?

What would make us change our mind?

The goal is not to remove human judgment. That is impossible. The goal is to make judgment more consistent, more accountable, and less vulnerable to mood.

QPOS: the operating system

At the core of our index methodology is QPOS:

• Quality.

• Price.

• Odds.

• Sizing.

Every opportunity is evaluated through four questions.

Quality: Is this asset, business, or investment structure worth owning?

We are looking for durability, balance sheet strength, reinvestment ability, cash generation, capital discipline, and downside resilience. Quality does not mean comfort. It means the asset has something real behind it.

Price: Are we being paid enough for the risk?

Price is where expectations live. A great asset can be a poor investment at the wrong price. A hated asset can become attractive when the market’s expectations fall too far. Graham’s margin of safety still matters because the only risk no investor can eliminate is the risk of being wrong.

Odds: Is the expected value asymmetric?

We are not looking for certainty. We are looking for setups where the range of possible outcomes appears skewed in our favor. Bear case. Base case. Bull case. Probability. Payoff. Invalidation.

Sizing: How much capital does this idea deserve?

Conviction without sizing discipline is dangerous. Liquidity, volatility, correlation, thesis strength, downside risk, and time horizon all matter. The question is not only “Is this interesting?” The question is “How much risk has this earned?”

QPOS does not replace human judgment.

It strengthens it.

Passive is not always passive

Traditional market-cap weighted indexes are often described as neutral.

That is only partly true.

They are rules-based. And rules create behavior.

When a stock rises enough, it becomes a larger part of the index. Sometimes it enters the index after a large move. When a stock falls, it becomes a smaller part of the index. Sometimes it exits after a large decline.

The result is simple: market-cap weighted indexes can end up buying recent winners after they have gone up and selling recent losers after they have gone down. Research Affiliates has argued that “passive” indexes are not as neutral as they appear and can embed hidden costs inside their selection and weighting rules.

This does not mean passive investing is bad.

It means indexes are not magic.

They are rules.

And rules matter.

We do not want to own something simply because it has become large. We want to own it because it has earned a place through quality, price, odds, and sizing.

Systematic does not mean blind

The word “quantitative” often sounds cold, opaque, or over-engineered.

We do not see it that way.

A systematic process is simply a way to reduce avoidable mistakes. Alpha Architect describes quantitative value investing as a tool that helps value investors acknowledge their own fallibility, protect against behavioral errors, and exploit the behavioral errors of others.

Their stated objective is direct: buy the cheapest, highest quality stocks in the market.

That is close to how we think.

We are not trying to outsource judgment to a machine.

We are trying to build a repeatable process that makes our judgment harder to fool.

The greatest threat to long-term investment performance is often not the market.

It is the investor.

Buying near the top because everyone feels safe. Selling near the bottom because everyone feels scared. Overreacting to recent results. Confusing volatility with risk. Confusing comfort with safety. Confusing a good outcome with a good decision.

A decision machine exists to slow those instincts down.

The best of both worlds

The Research Index is not designed to be passive.

But it is also not designed to be discretionary in the loose, emotional sense.

We want the best of both worlds:

• The discipline of an index.

• The selectivity of active research.

• The transparency of rules.

• The flexibility of judgment.

• The patience of long-term ownership.

• The willingness to diverge from the benchmark.

We are not going to obsess over tracking error.

Tracking error measures how closely a portfolio tracks a benchmark. That is useful if the goal is to behave like the benchmark. It is not useful if the goal is to find what may be mispriced.

We expect to look different.

We want to look different.

Not for the sake of being contrarian, but because mispricing usually does not live in the center of consensus.

If an idea cannot survive the QPOS framework, it does not belong in the index. If it can, we are willing to let it look unusual.

Concentration is a feature, not a flaw

The Research Index is designed to hold roughly 20 to 30 assets.

That is intentional.

We want diversification, but not dilution. We want enough positions to reduce single-asset risk, but not so many that the index becomes a closet benchmark.

Strict criteria create concentration.

That is not a problem. It is the point.

A research index should express research.

It should not become a warehouse for every decent idea.

We aim to be long-term, buy-and-hold investors. We seek to own assets that can compound in value over time. But “long-term” does not mean passive neglect. Every position needs a thesis, a risk note, an invalidation rule, and a sizing logic.

Hold through volatility.

Do not hold through thesis decay.

That distinction matters.

Asymmetry is rarely comfortable

Some of the most interesting opportunities are not clean, popular, or easy to hold.

Asymmetric assets often come with volatility, controversy, uncertainty, or career risk. That is why they can become mispriced in the first place.

Bitcoin is one example.

For many investors, the perceived risk is volatility. But the deeper risk may be structural underexposure to an asset with historically convex return behavior.

WisdomTree noted that Bitcoin’s historical returns have been highly concentrated in a small number of trading days, and that missing the 30 best trading days from January 1, 2014 to March 23, 2026 would have reduced cumulative returns from more than 9,000% to 26%.

That does not make Bitcoin risk-free.

Far from it.

It means timing perfection may be the wrong standard. Some assets punish investors who try to be too precise.

The QPOS question is not “Is this volatile?”

The better question is:

Are we being paid for the volatility?

The research index as a learning loop

A good research index is not just a list of holdings.

It is a learning system.

Each decision creates a record. Each record creates feedback. Each feedback loop improves the next decision.

Why did we add this asset?

What did we expect?

What were the odds?

What did we miss?

Was the outcome driven by skill, luck, or changing facts?

What should change in the process?

This is where investing becomes less about prediction and more about calibration.

We will still make mistakes.

That is guaranteed.

The goal is not to eliminate errors. The goal is to make them smaller, less emotional, and more useful.

A bad result is not always a bad decision.

A good result is not always a good decision.

The research index helps us remember that when the market tries to make us forget.

Subscribe now