Robinhood Markets, Inc. HOOD

Key data

The business

Robinhood is a phone app where mostly younger Americans invest in stocks, options and crypto, park cash, and borrow against what they hold. It makes money three ways. It takes a cut of each order routed to market makers, biggest in crypto and options, the piece that swings the most. It earns net interest on customer cash and margin loans, a steadier stream that rises and falls with the Fed. And it sells a paid Gold tier that bundles a higher cash yield, bigger instant deposits and research into a recurring fee. The moat is thin in the usual sense, moving your account is easy, but a free, fast product, a brand lodged with a whole generation, and the Gold bundle keep people from leaving. The thing the customer actually holds is the app and, for Gold members, a cash-back card.

The numbers

Five quarters show the machine and its one soft spot.

The story is in the middle two rows. Revenue fell 17% from Q4 2025 to Q1 2026 as crypto volume dried up, then set a record in Q2, up 32% on the year and 23% off the Q1 low. Q2 rebuilt to a record without crypto doing the heavy lifting, which means the newer lines, event contracts especially, kept carrying. Q2 EPS of $0.62 beat the $0.44 the Street looked for by 41%.

Revenue has gone 3.3x from the 2022 trough to 2025, about 49% a year, and the company swung from losing $1.0B in 2022 to earning $1.9B in 2025. First-half 2026 EPS of $1.00 is up 27% on the same half last year. This is real compounding, not an accounting story.

What the 54x pays for is the line below, and it is a lot.

The multiple assumes revenue roughly doubles again by 2028 and earnings compound with it. What I believe that the tape does not: the transaction line is still more crypto-and-rate-levered than a record quarter suggests, and the Q1 drop is the honest signal, not the noise. The single print that settles it is Q3-Q4 transaction revenue against net interest as the Fed cuts.

Management

Insiders sold about $69M across 65 sales over the past year and bought nothing. Founder-CEO Vlad Tenev accounted for about $28.6M of that in three sales on July 6, 2026, and the Form 4 plan status is not disclosed in this run, so whether those were pre-scheduled or discretionary, which changes what they mean, cannot be told from here. Set against that, the buyback has been well-timed: $653M repurchased in 2025 and $250M in early 2026, at prices below today's, a fair use of cash rather than empty support. Management beat EPS in three of the last four quarters and missed one by a rounding error. The record is good. The unread signal is the sale plan status.

How it fails or surprises you

Rate cuts squeeze the steady leg. Net interest on customer cash and margin loans is the part that does not need a hot product, and it falls as the Fed cuts. If short rates drop through late 2026 while balances stay flat, this stream shrinks just as the transaction line needs it most. Watch net interest revenue sequentially in the Q3 print.

The rotation stalls, not rotates (right tail). Every hot product so far has been replaced by a hotter one, and event contracts are ramping fastest. If prediction markets scale into a durable, high-margin line the way options did, the transaction business stops looking cyclical and starts looking like a platform. The market pays for none of this today. Event-contract revenue disclosure in Q3-Q4 reveals it first.

Q1 2026 is the tell (proves the read wrong). Revenue fell 17% in one quarter when crypto went quiet, yet the stock kept climbing. If that repeats and the record Q2 was the peak, transaction revenue rolls and 54x has no floor. A second down quarter in transaction revenue would break the "stepped-up permanently" case cold.

Closing thoughts

At 54 times earnings this is mostly world one: the market already prices the rotation machine and the steady interest income, so the edge is small and sits on which way the next quarter's revenue mix breaks. The people paying 54x believe transaction revenue has stepped up for good and that Gold's recurring fee de-risks the swings. My read leans the other way, because Q1's 17% sequential drop when crypto cooled says the biggest revenue leg is still cyclical, and at this multiple a simultaneous crypto lull and rate-cut cycle leaves no room. A print narrows it: Q3-Q4 transaction mix and net interest as the Fed eases. The fatter tail is the left one, because at 54x a double-hit compounds fast, but the right tail, event contracts scaling and Gold turning recurring, is live and unpaid-for.

The bet is still that Robinhood keeps pulling young investors in and getting them to do more, invest, borrow, subscribe. What breaks it is crypto volume rolling over while the Fed cuts, draining the transaction take and net interest at the same time. The one pair that tells you first: transaction revenue and net interest revenue, sequentially, in the Q3 report. If both fall together, the record quarter was the top.