李录(Li Lu) 视角 — HOOD

Verdict

Outside My Circle — and, separately, the price would already rule it out. This is not a knowable business to me at a wide margin of safety; it is a market-sentiment-geared retail trading platform priced at 45 times earnings. On both counts I stop, and I say so plainly rather than manufacture a verdict on something I cannot know.

Circle of Competence — the Boundary

Outside. Let me draw the exact edge, because the boundary — not the size of the circle — is the whole discipline. "An ability without a boundary is not an ability… The most important part of the circle of competence is the boundary."

The test I apply is simple and unforgiving: can I write, in one or two sentences, how this business makes money and why it will keep making it for the next decade — and can I predict that future with high certainty? Try it on Robinhood and the qualifiers multiply immediately. The revenue is a stack of things that each move with the crowd's mood, not against it: transaction-based revenue from options, equities, and crypto (crypto revenue down 47% year-over-year this quarter alone, per DATA.md ④); net interest income that swings with the rate curve; a young event-contracts / prediction-market line growing 320% off a tiny base; and a flurry of just-closed acquisitions (Bitstamp, WonderFi, TradePMR, and a pending MIAX stake) that have not yet been digested. That is not a sentence. That is a paragraph of moving parts, and the paragraph is the answer.

Where does my knowledge run out? Three places, and each one cuts straight through the thesis rather than around it:

1. Payment for order flow. A large share of transaction revenue depends on a practice that the SEC and the EU have repeatedly threatened to restrict or ban — the company itself lists it as a material risk (DATA.md ⑫). I cannot price the probability or timing of a regulatory regime change. That is exactly "the regulatory regime you don't follow" at the boundary.
2. Crypto and retail-trading psychology. The earnings engine is geared to bull-market retail enthusiasm and crypto volumes. I forecast businesses; I do not forecast the crowd's appetite for speculation, and a business whose earnings are the crowd's appetite is one I cannot project.
3. The price of a fast-moving fintech narrative. This is the "story-tech" my method explicitly marks down — not from contempt, but because most of it is outside my circle at the prices people pay for it.

When the boundary cuts through the thesis, the verdict is Outside My Circle — a full stop, not a small starter position. I will not fake conviction on the unknowable. "If you stray outside of your circle of competence… there will be some moment when the market takes you to the cleaners." For me, this business sits past the edge.

Margin of Safety (run first)

I run the valuation gate first, by design, because the deep research is too expensive to spend on a business the price already rules out. Here it rules it out.

• Conservative intrinsic value range: I will not pretend to a precise DCF on earnings this cyclical — a long projection on a business this sentiment-geared would be exactly the "false precision dressed as analysis" I warn against. But the cross-checks are clear enough:
  • Comparables / earnings: P/E 45.26x trailing, 46.58x forward; P/S 18.19x; PEG 2.69x (DATA.md ③). These are paid-up multiples for a business with negative net income as recently as FY2021–FY2023 (−$3.69B, −$1.03B, −$0.54B; DATA.md ④).
  • Asset value: P/B is 9.01x book of ~$9.69B. Buying a brokerage at nine times tangible book is the opposite of my asset discipline (count fixed assets plus working capital, do not count on goodwill). There is no asset margin of safety here.
  • Free cash flow: P/FCF 27.86x, an FCF yield of just 3.59% (DATA.md ③) — and that FCF margin of 65% is flattered by an asset-light model, while FCF itself was negative in FY2024 (−$170M) and FY2022/FY2021.
• Current price vs. value → margin of safety: Negative. At every conservative lens — earnings, sales, book, free cash flow — the price sits above, not far below, any value I could defend. There is no protective gap; there is a premium.
• Gate: "The price at which you buy should always be far, far below the company's intrinsic value." A margin of safety here is not merely under my ~25% floor — it is absent. The gate fails on the spot. The fact that the stock is down ~35% year-to-date and ~39% off its 52-week high (DATA.md ②) does not change this: a fall from an expensive price can still leave an expensive price. Cheapness is the entry ticket, and this name has not bought one.
• Data sourced? user-provided (DATA.md, sourced to SEC filings and stockanalysis.com). No numbers invented.

Is It a Good Business?

A genuinely good business and an unknowable, expensive one are different questions, and I keep them apart. On the merits there is real quality here — but quality I cannot buy at a margin of safety is not an opportunity for me.

• Returns on capital: ROE 21.46%, ROIC 8.38%, ROA 5.19% (DATA.md ⑦). The ROE clears my ~15% preference, but the ROIC at 8.4% is modest, and DATA.md correctly flags that the equity base was distorted by years of heavy stock-based compensation (FY2023 SBC of $871M was 47% of revenue), so I do not trust these ratios as a stable read on the engine.
• Moat — wider in 10 years? Honestly, unclear, and partly narrowing. The strongest widening mechanism is the Gold subscription (4.3M users, +36% YoY) building recurring revenue and stickiness, plus accumulating regulatory licenses (DATA.md ⑩). Against that: zero-commission is now industry-standard with no pricing differentiation, the customer base is young and low-balance (ARPU $157), and the core earnings still ride on transaction volumes the incumbents (Schwab, Fidelity, Interactive Brokers) can match. A moat I cannot confidently say will be wider in a decade is, for my purposes, not yet a moat I can underwrite.
• Financial conservatism (second line of defense): Net cash ~$5.66B is a genuine strength. But the operating reality is leverage-to-sentiment, not balance-sheet leverage: current ratio 1.11x, debt/equity 1.40x (DATA.md ⑤). For a long hold, the financial fortress matters less than the earnings fortress, and the earnings are anything but conservative — they were deeply negative three years running.

The summary: better than the bears claim, but not a business whose future I can predict with high certainty, and not at this price.

Who Is Running It?

I judge character first, like an investigative journalist, and from DATA.md I can only do the screen-level pass — not the shoe-leather work my method actually requires.

What the record shows (DATA.md ⑨, ⑫): founder-CEO Vlad Tenev controls the company through Class B super-voting shares alongside co-founder Baiju Bhatt, so public Class A holders have no effective governance voice — the standard founder-control structure. There is a real regulatory track record I cannot wave away: a $45M SEC settlement (Jan 2025, compliance violations) and a $30M FINRA settlement (Mar 2025, AML deficiencies). And insiders have been net sellers over the trailing three months.

None of this is a knockout integrity veto on its own — these are common features of a fast-growing fintech founder's company. But two things matter to my method. First, the AML and compliance settlements are precisely the kind of conduct an investigative journalist's read of the court and regulatory documents exists to surface, and they are not reassuring. Second, the super-voting structure means that fairness to minority holders is structurally at management's discretion, which raises, not lowers, the bar I would need to clear on character before trusting the reported cash flows. I have not done that on-the-ground work, and I would not pretend to conviction on the people from a press release and a settlement docket.

Deep-Research Status

The work is not finished — and for me it should not be started, because the boundary and the price already end the analysis honestly. Most mistakes come from acting on incomplete or inaccurate information, so let me state the completeness plainly: I have a data snapshot, not encyclopedic knowledge. I have not READ EVERYTHING — not the full 10-K risk factors line by line, not the lawsuits and regulatory orders behind the $75M of settlements, not the four acquisition agreements (two of which, TradePMR and MIAX, have undisclosed prices in DATA.md ⑧), not the management on the ground. To go further would be to spend my scarcest resource — research time — on a name the valuation gate already failed. That is the gate working as designed. I am not betting on a probability here because I like the idea; I am declining to bet because the work would be wasted.

What Did I Miss?

The honesty inversion. Let me build the strongest bull case I can, because the danger is always the thing I didn't look at because I'd already reached my answer.

The strongest bear-to-my-pass / bull-for-the-stock case: Robinhood in 2026 is no longer the meme-stock toy of 2021. FY2025 revenue grew ~51% to $4.47B with $1.88B net income and, for the first time, buybacks ($653M) exceeding stock-based comp ($305M) — a real capital-allocation inflection (DATA.md ⑥, ⑧). Net deposits hit a record $68B, platform assets are $307B (+39% YoY), Gold subscriptions and recurring revenue are compounding, and the company is methodically buying its way into crypto infrastructure (Bitstamp), wealth management (TradePMR), prediction markets, and derivatives-exchange ownership (MIAX). If even half of that $600B TAM ambition is real, today's 45x earnings could look cheap against the earnings of 2030. That is a coherent growth thesis, and I want to state it at full strength.

Where this thesis is "greatly affected by psychology": right here. It rests on (a) continued bull-market retail engagement, (b) crypto volumes recovering rather than the −47% trend continuing, (c) PFOF surviving regulators, and (d) four near-simultaneous acquisitions integrating cleanly. Each of those is a forecast of the crowd, the regulator, or management's execution — not a forecast of a durable business I can project. The thesis asks me to pay a premium today for an outcome that depends on sentiment staying favorable. That is the speculator's forecast, not the investor's. Investors forecast companies' future performance, while speculators forecast other market participants' short-term behaviour.

Could I be wrong? Yes — the stock could compound handsomely from here if the bull case lands. But "I might miss an up-move" is not the same as "I have a knowable business at a wide margin of safety." I am wrong, in my framework, only if I have called something unknowable that is in fact knowable to me at a protective price. I have re-checked, and I have not. The honest conclusion is that this is a fine business to watch from outside and a poor one to underwrite. If anything, the inversion confirms the pass.

Macro / China Thesis

Not applicable as a source of edge. This is a US-listed, US-retail-driven business; there is no China/Asia mispricing here that my cross-cultural research moat could exploit, so my civilizational thesis gives no signal. The only macro that touches the case is ordinary US cycle exposure — interest rates (the rate curve at ~3.8% short / ~4.5% long supports net interest income today, DATA.md ⑭) and retail-trading sentiment — and a case that leans on US rates and crypto mood staying friendly is macro speculation, not the kind of decades-long civilizational tailwind I underwrite. My framework returns nothing here except: this is not the market the West misprices that I am built to know.

Concentration & Patience

• One of the 5–10 I'd actually own? No. I expect only five to ten truly knowable opportunities in an entire career, and I size each one to matter — roughly a fifth of capital behind a finished thesis. This is not a finished thesis; it is an unknowable business at a premium price. It does not earn a 20% slot or a 2% one. Concentration without finishing the boundary and research work is not my method — it is gambling with size.
• Hold/sell test: Not held, so nothing to sell. For the record: the burden of proof on a sell is on the bear, but there is nothing to defend here in the first place.
• Default holding period / act or wait? My honest move is non-action. Most of the time the right thing to do is nothing, and holding cash while I wait for a knowable opportunity at a wide margin of safety is not idleness — it is the honest response to an empty opportunity set. This name is empty for me.

In My Words

I will say this carefully, because the temptation with a fast-growing, freshly-cheapened-looking name is to talk yourself into it. I cannot. Robinhood is a real and improving business — the capital-allocation turn in 2025 is genuine, the subscription flywheel is genuine, and I do not dismiss the people running it as charlatans. But my whole method begins before any of that, with two questions: can I predict this business's future with high certainty, and is the price far, far below a conservative estimate of its value? For me the honest answer to both is no. The earnings ride on the crowd's appetite for trading and crypto, on a payment-for-order-flow regime I cannot price, and on four acquisitions not yet digested — that is a paragraph of moving parts where I need a sentence of certainty. And the price, even down a third this year, is a premium to every conservative lens I can apply, not a discount. So I draw the boundary where it actually falls — with this business on the outside — and I stop. You should never fool yourself, because you are the most easily fooled. The intellectually honest verdict is not a clever pass-with-a-watchlist; it is simply: this one is unknowable to me, and even if it weren't, it is not cheap. I do nothing, and I am comfortable doing nothing.

基于 2026-06-13 共享数据;本分析为单一大师框架的演绎,非投资建议。