Record Revenue at Lower Prices, and the Courts Turn on Sports Contracts: One Trigger Fires Each Way, Hold Maintained

Results vs. Consensus

Q3 2026 Scorecard

The revenue consensus most widely cited into the print was $1.44B. One major provider carried $1.42B, against which the result is in line. EPS consensus sat at $0.56, with two compilations at $0.55. Those estimates predate the new Adjusted EPS measure, so GAAP diluted EPS is the like-for-like comparison. Outlets that compared the $0.60 adjusted figure reported a beat, and outlets that compared the $0.51 GAAP figure reported a miss.

Year-over-Year Comparison (Q3 2026 vs. Q3 2025)

Sequential Comparison (Q3 2026 vs. Q2 2026)

Revenue. A 9% sequential increase to $1,424M is well above the flat quarter we modeled in July from management's "similar area" volume guidance. Our caution on the trading book was wrong. The composition is the issue. Options revenue rose 17% on 17% more contracts at an unchanged 44 cents per contract. Equities revenue fell 6% while notional volume rose 4%, so revenue per dollar traded dropped about 10%. Event contracts rose 11% on 21% more contracts. The crypto increase is mostly a new revenue source inside the line. Net interest grew 7% sequentially but fell 9% year-over-year, which comes down to one line: securities lending swung to a loss. The $16M revenue shortfall against consensus is a take-rate shortfall, and management said October spreads are tighter again.

Margins. The Adjusted EBITDA margin held at 57%, against 58% a year ago, and revenue less total operating expenses reached 48% of revenue against 44% in Q2. Marketing rose only 2% year-over-year to $104M and technology rose 7%. The pressure sits in the lines that scale with new businesses and with credit. Operations costs rose 139%, mainly Trump Accounts customer support and card growth. Provisions rose 173%. General and administrative costs fell 12%. Q3 is the best margin quarter of the year on our estimates, because the fourth quarter carries the deferred expense.

EPS. GAAP diluted EPS of $0.51 fell 16% year-over-year. Three items explain the gap to the year-ago $0.61: the $0.07 investment mark, the higher tax rate, and provisions $45M higher. The new Adjusted EPS of $0.60 removes the mark, a small CFO-transition charge and $9M of acquired-intangible amortization. It fell 3% year-over-year against a restated $0.61. The company amended the measure the day after the release to correct the tax effect of the adjustments. The Q3 figure did not change; the Q2 and nine-month comparatives did. We use GAAP EPS as our primary basis.

Revenue Line Performance

Unit economics and the crypto figure excluding Robinhood Chain are our calculations from reported revenue, reported volumes and the CFO's stated chain revenue.

Options and Equities: Record Volume, Lower Price

Options revenue of $399M and 905 million contracts traded are both records. Revenue per contract held at about 44 cents, unchanged from Q2 and down from about 50 cents a year ago. The 10-Q gives the breadth behind it: contracts traded per options trader rose 66% year-over-year while the number of customers placing options trades fell 11%. The business is growing through more intense use by fewer people.

Equities show the same pattern with a price cut on top. Notional volume reached a record $992B, up 4% sequentially, yet revenue fell 6% to $121M. Revenue per dollar traded was about 1.22 basis points, against 1.35 in Q2 and 1.33 a year ago. The 10-Q attributes the year-over-year pattern to higher notional per trader, partly offset by lower rebate rates on the mix of symbols traded.

"So on equities, it's usually a mix shift between single names and ETFs, and it also depends on what the spreads are. Options is also similar. It depends on what the vol is and also the spreads." — Shiv Verma, CFO

Assessment: Bull 1, the active-trader franchise, stays ON TRACK on volume and market share. The equities revenue miss of about 8% against consensus is the clearest sign that the Street over-modeled unit economics. A shrinking count of active option traders deserves more weight than it got on the call.

Event Contracts: Faster Volume, Lower Price, Better Mix

Event-contract revenue reached a record $173M, up 11% sequentially and more than six times the year-ago $27M. Contracts traded rose 21% to 16.4 billion, so revenue per contract fell about 8%, to roughly 1.05 cents from 1.15 cents. Management attributed the change to its new dynamic pricing model and Gold discounts and said the pricing change did not lower its take rate. The mix improved: non-sports contracts rose to about 30% of the quarter from about 15% in Q2. Rothera passed $100M of annualized revenue in its first full quarter.

"I think we're less excited about doing what everyone else is doing, which is just all sports all the time because I think that's rapidly going to get commoditized and is just, frankly, a little bit less differentiated from my standpoint." — Vlad Tenev, Chairman and CEO

Assessment: Bull 3 stays ON TRACK commercially. The non-sports mix is the most valuable data point, because it is the part of the business the state court cases do not reach. The legal exposure of the sports majority is covered in Key Topic 2 and is now a separate bear point.

Cryptocurrencies: The Chain Is Inside the Line

Crypto revenue rose 20% sequentially to $120M, the first increase in four quarters, on total notional volume of $49.8B (+23%). Robinhood App volume rose 13% to $20.6B and Bitstamp volume rose 32% to $29.2B. Monthly app volume strengthened through the quarter, from $4.3B in July to $8.9B in September. The 10-Q, however, explains the year-over-year decline as fewer customers placing crypto trades (-37%), lower notional per trader (-13%) and lower rebate rates, "partially offset by revenues earned from Robinhood Chain." The chain fees therefore sit inside this line.

"Additionally, the Robinhood chain in its first quarter of adoption saw really strong usage, as Vlad mentioned before. So that was another driver of transaction revenue in the quarter." — Shiv Verma, CFO

On the call the CFO put the chain's net revenue at about $25M. If all of it is in the crypto line, crypto trading revenue was about $95M, down about 5% from $100M in Q2. The blended rate per dollar traded also slipped, which management attributed to the stablecoin-based Robinhood Earn product, now over $500M of assets.

Assessment: On the line as reported, our July upgrade trigger of crypto "flat or better sequentially" fired, and we grade it as met. On the trading book alone it was not met, but the decline narrowed to about 5% from 25%, and September volumes were the best of the quarter. Bear 1 stays CONTAINED: crypto is 8.4% of revenue, and a bottom in the trading book is closer than it was in July.

Net Interest: Securities Lending Turned Negative, Everything Else Grew 18%

Net interest revenue of $415M fell 9% year-over-year, or $41M. One line explains more than all of it. Securities lending, net swung from $87M a year ago to a loss of $19M, as interest paid on stock-loan balances exceeded interest earned. The 10-Q attributes this to "a lower average rate on higher stock loan balances as the hard-to-borrow market weakened." Without securities lending, net interest revenue grew about 18% year-over-year, from $369M to $434M, and about 15% sequentially.

The other moving parts follow the strategy. Margin interest rose 59% to $244M on a record $22.1B Margin Book, although the book grew only 2% from June. Credit card interest, net nearly tripled to $50M. Interest on segregated cash fell 47% to $55M because Robinhood now pays interest on customer free credit balances under the February cash-program change. Cash Sweep revenue fell 30% to $45M on lower rates. Net interest revenue was 2.26% of average interest-earning assets, annualized, down from 2.34% in Q2 and 3.00% a year ago, although the gross yield on those assets rose to 2.41% from 2.33%.

The rate backdrop has turned. The 10-Q notes that the Federal Reserve raised its policy rate in September 2026, and management gave a new sensitivity figure in answer to a question about hikes.

"The actual math is for every 25 basis point hike or decline, it's about a 1.5% change to revenue. So all else being equal, if the Fed hikes another 25 basis points, that should be a lift. The really simple math is we have about $80 billion of interest-earning assets. About half of that $40 billion is interest-sensitive assets because the sweep program and a few other things aren't interest sensitive." — Shiv Verma, CFO

That is about $100M of annual revenue per 25 basis points, on a revenue base management puts at $5.5B to $6B.

Assessment: The headline decline is mostly one volatile line, and the core balance-sheet business grew at a high-teens rate. That is better than the 9% headline. Two cautions keep Bull 5 AT RISK. Securities lending was a material profit pool in 2025 and is now a small loss, with no sign of a quick return. And the Margin Book, which drove net interest growth all year, grew only 2% in the quarter.

Other Revenues: Gold and Trump Accounts

Other revenues rose 61% year-over-year to $142M. Gold subscription revenue rose 32% to $62M on 5.18 million subscribers. The "other" component, which holds the Trump Accounts service revenue, more than doubled to $73M from $34M. Proxy revenue was $7M, unchanged. The line was flat sequentially because Q2 holds most of the year's proxy revenue: $57M for nine months, of which Q3 contributed $7M.

Assessment: The most predictable revenue in the model is now 10% of the total. The Trump Accounts amount is not disclosed, but the $39M year-over-year increase in the component that holds it gives a rough upper bound for the program's quarterly revenue. On management's description the program earns service fees on a cost-plus basis, so its margin is thin.

Key Operating Metrics

Key Topics & Management Commentary

Overall Management Tone: Management was as confident as in July and more promotional, giving a live product demonstration and the September launches more airtime than the income statement. The forward data were selective: October deposits, event-contract volumes and slightly lighter trading volumes were volunteered, while tighter spreads and the shift of deposit growth toward existing customers came out only under questioning. Nothing was said about the Ohio sports-contract agreement signed the day before the call.

1. Grading the July Commitments

In July management made seven forward statements that we recorded as the checklist for this quarter. Three were delivered or exceeded, two are on track (one without a number), one was contradicted in Robinhood's favor, and one was immaterial.

Assessment: Management's forward statements continue to be conservative on activity. In July we used the "similar area" volume guidance to model a flat Q3 of $1.29B to $1.34B. Revenue came in at $1,424M, so our caution on the trading book was wrong. The commitment that matters most for the rating, deposit growth, delivered in the full-year sense management defined and not in the quarterly sense we used for our triggers.

2. Sports Contracts: The Courts Turned

The legal position of sports-related event contracts worsened in the quarter. The release mentioned it only in generic risk language, and the call not at all. The 10-Q, filed the day after the call, sets out the sequence. In August the Ninth Circuit affirmed the denial of Robinhood's request for an injunction against Nevada, where Robinhood stopped new sports contracts in December 2025. In September the Sixth Circuit, in a parallel case, affirmed the denial of an injunction in Ohio and reversed one granted in Tennessee. Robinhood then agreed with Michigan to stop offering sports contracts and close existing ones. On October 2 the Ohio Casino Control Commission sent Robinhood a cease-and-desist letter, and on October 26 Robinhood agreed to stop new sports contracts in Ohio by November 2. Each agreement trades the exit for the state's non-enforcement while appeals run. Robinhood has conditionally petitioned the Supreme Court, and the Third Circuit decision in New Jersey still favors the industry.

Management's only strategic comment on the subject was about direction, not exposure.

"And at least from our standpoint, we have been investing a lot in the non-sports contracts. You saw us announce KPI prediction markets. We've been seeing a lot of growth in the crypto contracts and commodities." — Vlad Tenev, Chairman and CEO

The exposure is meaningful but bounded. Event contracts are 12% of revenue, and management put the non-sports share of the quarter's mix at about 30%. If revenue follows that mix, sports contracts are roughly 8% of total revenue. The exits so far cover three states. The risk is that two appellate courts now support state authority, which invites more states to act before the Supreme Court decides.

Assessment: This is the adverse regulatory development we named in July as a downgrade trigger, and we grade that trigger as fired. We add a sixth bear point for the legal status of sports contracts, at MATERIALIZING. The offset is the mix: if the non-sports share keeps rising, a narrow ruling would cut growth without breaking the line.

3. Rothera, the Crypto.com Stake and the Venue Strategy

Robinhood restructured how it sources prediction markets. Rothera, the exchange and clearinghouse it owns with Susquehanna, passed $100M of annualized revenue and over 6.4 billion contracts since launch. In September Robinhood also signed agreements with Crypto.com and an affiliate under which it may receive up to a 10% equity interest over three years as consideration for routing a selection of event contracts. The 10-Q describes the consideration as "entirely variable". Management said the two venues will carry the "vast majority" of flow by early 2027, and that most other venues will be dropped.

"But additionally, with no cash outlay, we received an initial 1% equity ownership stake in crypto.com and og.com. And we have the ability to earn up to 10% ownership depending on how much we partner together in the years to come." — Shiv Verma, CFO

The CEO described Rothera as a neutral venue for other brokers, since it runs no consumer business of its own, and said Robinhood does not operate it day to day.

Assessment: The structure earns exchange economics on one venue and equity upside on the other, at no cash cost. It does not answer the question we have asked for three quarters: what Robinhood earns per contract on Rothera against a third-party venue. Our July upgrade trigger on Rothera economics did not fire.

4. Net Deposits, MAUs and the Funnel

Net deposits were $15.7B, a 17% annualized rate, against $21.7B (28%) in Q2 and $20.4B (29%) a year ago. By month they were $5.6B in July, $4.0B in August and $6.1B in September. From July the metric includes Trump Accounts contributions and Treasury seed funding, which the July call's "$4 billion area" guidance had excluded. None of the release, the 10-Q or the monthly metrics supplement says how much of the quarter came from that source. Trailing twelve-month deposits were $71.0B, a 21% rate.

The engagement numbers were weaker. Monthly active users fell 8% sequentially to 13.0 million, the same level as December 2025, and slipped to 45% of funded customers from 50%. Funded customers grew by only about 0.3 million: 0.6 million new, 0.1 million returning and 0.4 million churned. Trump Accounts are excluded from the funded-customer count.

"I would say it's more of the latter. We don't over-index in any 1 month or quarter. And the best way to see it is October. As we mentioned, October was already a record of over $13 billion of net deposits." — Shiv Verma, CFO

October is the counterweight. Excluding the $6B Dell contribution, the month's roughly $7B annualizes to about 21% of the September asset base, above the 20% target, although it may also include other Trump Accounts flows.

Assessment: Neither of our July deposit triggers fired: 17% is above the 15% downgrade line and below the 20% upgrade line. Bear 5 stays EMERGING. The composition worries us more than the rate: deposits now lean on existing customers and on a federal program, while the active-user count fell and new-customer growth slowed back to the Q1 pace.

5. Credit: Card Losses Are Now Visible

Provision for credit losses rose to $71M, up 173% year-over-year and 27% sequentially, and reached 5.0% of total net revenue against 4.3% in Q2 and 2.0% a year ago. Of the $71M, $61M relates to the credit card. The Gold Card reached 1.3 million customers with over $19B of annualized payment volume, and card receivables reached $1,841M, up from $1,457M at June and $742M a year ago.

The 10-Q gives the loss detail that the release does not. In Q3, write-offs on purchased card receivables were $31M and payments to the partner bank on off-balance-sheet receivables were $9M, against $5M of recoveries. That is about $35M of net losses on average card receivables of $1,692M, or roughly 8% annualized on our calculation. Past-due receivables rose to 6% of the book from 5% at December, and the share of balances with FICO scores below 640 rose to 5% from 3%. The allowance on purchased receivables is $134M, or about 8% of the on-balance-sheet book.

"So credit card, we're rolling out more delivery because you have to get the unit economics right, you have to get the credit risk right. Banking is just like any of the other products that we launch." — Shiv Verma, CFO

The comparison that matters is with card revenue. Credit card net interest was $50M in the quarter, below the $61M card provision, on an annualized yield of 11.8%. Interchange income adds to that, and provisions on a fast-growing book are front-loaded under current accounting. But a loss rate near 8% leaves a thin spread over a yield below 12%, and that is before card-related operations costs, which rose $18M year-over-year.

Assessment: Bear 4 moves from EMERGING to MATERIALIZING. Our downgrade trigger is provisions above 6% of revenue, and at 5.0% it has not fired. Another 27% sequential increase in Q4 on flat revenue would take provisions above 6% and fire it. The card is now large enough that its loss rate deserves a place in the release, and Robinhood does not give one.

6. Expenses: A Lean Third Quarter Moves Cost Into the Fourth

Adjusted Operating Expenses and SBC were $664M, up 8% year-over-year and 4% sequentially, against revenue growth of 12% and 9%. GAAP total operating expenses of $739M rose 16% year-over-year. The release attributes the increase to provisions, Trump Accounts and Rothera costs, and growth investments. Operations costs alone rose 139%. The full-year range came down by $25M at the top to $2.675B–$2.75B.

"So as we do that, we're going to lower the top end of our prior outlook range by $25 million, and we now expect full year adjusted OpEx in SBC to be between $2.675 billion and $2.75 billion, and this is $13 million lower at the midpoint versus our prior outlook." — Shiv Verma, CFO

The arithmetic matters more than the cut. The first three quarters total $1,912M. The new range therefore implies fourth-quarter Adjusted OpEx and SBC of $763M to $838M, up 15% to 26% on Q3. In July the range implied an average of about $739M a quarter for the second half. Q3 came in $75M below that average, and the guide moved only $13M at the midpoint, so most of the undershoot was timing. Management named three Q4 items: more Trump Accounts spend inside an unchanged ~$100M full-year budget, the credit card scale-up, and investments pulled forward from 2027.

Assessment: The expense discipline is real, and the 2026 range has now come down twice. But the third quarter's 57% Adjusted EBITDA margin should not be annualized. On flat revenue, the guided range takes the Q4 Adjusted EBITDA margin down by roughly seven to twelve points. Bear 2 stays CONTAINED, with Q4 as its test.

7. A New Adjusted EPS, Amended the Next Day

Robinhood introduced Adjusted Net Income and Adjusted EPS this quarter. The measures remove gains and losses on investments (mainly the Robinhood Ventures funds), restructuring charges and related SBC, CFO-transition SBC, and amortization of acquired intangibles, net of tax. Normal SBC stays in.

"So we're introducing adjusted EPS, and this is to just better reflect our core operating profitability before any gains or losses on investments or some of these other nonoperating items. But at the same time, we're going to remain accountable for all of our regular ongoing costs. So we're still including normal course SBC in our adjusted EPS." — Shiv Verma, CFO

On October 28 the company filed an amended release solely to revise the tax effect of the non-GAAP adjustments. The amendment did not change GAAP results or the Q3 Adjusted EPS of $0.60. It did change the comparatives. Q2 Adjusted EPS moved to $0.55 from $0.50, and nine-month Adjusted EPS to $1.54 from $1.51. On the corrected figures, Adjusted EPS fell about 3% year-over-year in Q3 ($0.60 against $0.61) and exceeds GAAP EPS by $0.03 over nine months.

Assessment: The measure is defensible. It removed Q2's $129M of gains as well as Q3's $67M of losses, and it keeps normal SBC, which many companies exclude. Two points count against it. It also removes amortization of acquired intangibles, a recurring cost of an acquisitive strategy. And a tax-presentation error in a brand-new measure, corrected within a day, is not what we expect from a company of Robinhood's size. We use GAAP EPS as the primary basis and show the adjusted figure alongside it.

8. Trump Accounts: Seventy Million Accounts and a $6B Deposit

The program changed scale in Q3. The Treasury opened accounts automatically for nearly 70 million eligible children, and the Dell family's $6.25B pledge put $250 into the accounts of 25 million of them. Robinhood is the broker and sole initial trustee, working with BNY. Of the $13B of October net deposits, $6B was the Dell contribution. Other revenues rose 61% year-over-year, and the release attributes the growth mainly to Trump Accounts service revenue and Gold subscriptions.

"I think it can become the default vehicle for philanthropic giving in this country, right? And we're already seeing the beginnings of that with the Dell's, but also the other donors -- and we don't really have a great default option. I think Trump accounts can become that." — Vlad Tenev, Chairman and CEO

The next stage is plumbing for employer contributions and for corporate and philanthropic donors. Management expects to keep investing, and the program's roughly $100M 2026 cost is part of the Q4 expense step-up.

Assessment: Bull 4 is the clearest positive surprise of the year. A federal account for nearly every American child, with Robinhood as trustee, is distribution no competitor can buy. Two qualifiers belong in the model. The Dell money and Treasury flows are deposits of a different quality: they are long-dated, small per account and locked until adulthood, so they add to platform assets and to the deposit metric without adding trading revenue. And the economics remain service fees on a cost-plus program.

9. Robinhood Chain: A Revenue Line Sooner Than Expected

In its first quarter on mainnet, Robinhood Chain became a top-five chain by revenue and volume, with over $80B of decentralized-exchange volume and over $50M of chain fees. About half of the fees go to the network it is built on, which leaves about $25M of net revenue inside the crypto line. A questioner noted that late-August and September congestion lifted fees. Management said a normal environment means "a couple of basis points" on notional and that it will not optimize for fees.

"So I don't really want to be in the business of like charging high fees for on-chain usage. What I'd rather do is gain the market share, make the developers choose us as the best chain, give them more products." — Vlad Tenev, Chairman and CEO

The CEO also signaled more spending, saying the company will ramp investment in the developer community because the chain's growth "was higher than our expectations."

Assessment: In July we called the chain a strategic option, not a revenue line. At about $25M net in its first quarter it is a revenue line, though Q3 was likely a high-water mark for fees. It also complicates the crypto line, which now mixes a cyclical trading book with an infrastructure fee stream.

10. Wallet Share: Banking Becomes a Destination

The wallet-share products kept scaling. Gold subscribers reached 5.2 million, up 34%, at an 18% adoption rate and about 40% of new customers. Banking reached nearly $5B of deposits from over 350 thousand funded customers, with about 40% signed up for direct deposit. Strategies reached 350 thousand funded customers and over $2B of assets. The CEO demonstrated a peer-to-peer payments feature that lets a recipient accept funds through Apple Pay or Google Pay without a Robinhood account.

"And this quarter, for the first time, Robinhood Banking became the #1 withdrawal destination for Robinhood brokerage customers. This is actually even before it's generally available." — Vlad Tenev, Chairman and CEO

Banking still requires a brokerage account. Management tied independent onboarding, and with it performance marketing, to general availability, without a date.

Assessment: Our April test for Banking was direct-deposit attach above 35% as deposits passed $5B. Attach is about 40% and deposits are "nearly $5 billion", so that test is all but met. Bull 2 stays ON TRACK. ARPU growth slowed to 5% from 24%, but that is mostly the crypto comparison.

11. The Roadmap: Perpetuals, 24/7 Trading, Agents and New Markets

The September event announced 17 products and features, including Robinhood Agents, 24/7 trading of select US equities, perpetual futures for eligible US customers built on Bitstamp, and earnings contracts in the prediction markets hub. Agentic Trading has about 200 thousand accounts and $1.6B of notional volume to date. Internationally, Robinhood launched crypto trading in the UK and received authorization in Dubai, with a product launch planned for 2027.

"So I see no reason why our offering can't get to a similar, if not greater scale than the DeFi offering as time goes on." — Vlad Tenev, Chairman and CEO

Assessment: The roadmap is the bull case's best argument, and the execution record supports it: US perpetuals, which in July had no timeline, were announced within two months. None of these items is a 2026 revenue driver. Perpetuals and 24/7 equities are 2027 inputs, and the CFO said part of the Q4 expense step-up is spending pulled forward for products launching in 2027.

Guidance & Outlook

Robinhood does not guide revenue or EPS, so the expense range and the October data are the whole outlook. Management framed Q4 as a strong start following a record quarter.

"So before we move to Q&A, I just wanted to give a little bit about how Q4 looks to start following a record Q3. October net deposits are over $13 billion in the month, great to see, and that includes $6 billion of the Dell family gift to Trump accounts, as Vlad mentioned earlier." — Shiv Verma, CFO

The two halves of the October data point in different directions. Deposits are at a record, and even without the Dell money the month's $7B is above the Q3 monthly pace of about $5.2B. Trading activity and take rates are flat to slightly down, and the revenue lines that move most with the trading book are therefore unlikely to grow in Q4. Event contracts are the exception at about +10%.

Implied Q4 setup: Equity, options and crypto volumes about 5% lower, slightly tighter take rates, event contracts about 10% higher, and continued growth in balances and Gold. On those inputs we model Q4 revenue of $1.40B to $1.48B, with $1.44B central, roughly flat on Q3. Against that, the expense guide implies an increase of $99M to $174M in Adjusted OpEx and SBC. With provisions near $80M and a tax rate near 21%, that puts Q4 GAAP EPS at about $0.41 to $0.55, with $0.49 our central case, before any investment marks. Q4 2025 EPS was $0.66.

Street at: Before the print, Q4 consensus was $0.63 of EPS on $1.54B to $1.55B of revenue. It has since moved to about $0.58 to $0.59. Our $0.49 central case is about 15% below that, mostly because we hold revenue roughly flat and take the expense guide at face value. For the full year, compilations sat at $2.17 to $2.22 before the print, and at least one has since cut to $2.07, against our $2.00.

Guidance style: Conservative on expenses and absent on revenue, as before. The full-year range has come down twice since April. The Q4 activity color was more cautious than July's was, and July's turned out too cautious.

Analyst Q&A Highlights

How much of the 20% deposit target comes from new customers versus existing ones

An early sell-side question went straight to the metric management asked to be judged on. It noted that, as the denominator grows, the reported rate has been running a little under 20%, and asked how the target splits between new accounts and wallet-share gains. The answer moved first to the October figure and the year-to-date record. It then made a concession the prepared remarks did not: recent growth has come more from existing customers than from new ones, and funded-customer adds in Q3 were "similar to Q1", not Q2.

Q: "I wanted to just ask on the net deposit growth, which I know seasonally 3Q tends to be a touch lighter. I'm really evaluating this KPI on a core basis. But as the denominator has grown larger, you've been getting -- running a little bit below that 20% target. And I just wanted to get a better sense as to -- as we unpack some of the building blocks, how much of that 20% is expected to come from new account growth or new store growth versus same-store growth?"
— Steven Chubak, Wolfe Research

A: "In terms of new store sales versus existing customers, it is both. More recently, it's been a little bit more of the existing customers. We got a lot of customers very, very fast in the prior years. And now as we ship new products, we're gaining share of wallet and they're starting to trust us more. As we've said before, though, we are starting to turn the dial to try to accelerate net funded accounts again. And so Q3 was similar to Q1."
— Shiv Verma, CFO

Assessment: This was the most useful answer on the call for the deposit debate. Growth from existing customers depends on market levels and on the wallet-share products. Growth from new customers is the part that compounds the base. Management is now saying the second engine slowed after Q2. Most of the fixes it named, standalone banking onboarding and new products it says will launch in 2027, will not help this year. That supports a deposit rate near the 20% line, not a return to the 28% of Q2.

What the expense guide implies for the fourth quarter and for 2027

The question observed that third-quarter expenses came in below expectations and asked about the moving parts, including Trump Accounts timing and the 2027 run-rate. The answer confirmed the timing point: about $100M of Trump Accounts spend for the year is unchanged, so some of it moves into Q4. It added two more Q4 items, the credit card scale-up and investments pulled forward from the 2027 plan. Management gave no 2027 figure and pointed to the January-February plan.

Q: "So the number for the quarter came in, I think, a bit lower than a lot were expecting. Your full year guide was lower just a little bit. Curious if you could talk about the moving parts. I know there's some timing of Trump expenses that are expected to come in, in the back half. And curious what you can share on the run rate into '27."
— Ben Budish, Barclays

A: "In Q4 specifically, there's a couple of things that we're rolling out that we're investing in. First, we're doing even more for the Trump account. So you should expect some more expenses there. Second is the credit card. We're continuing to scale the credit card, both Gold and Platinum. So there'll be some investments there. And then lastly, we're in the middle of annual planning right now, but there's a couple of items we're really excited for. And so we pulled those forward and we started to make a couple of investments there."
— Shiv Verma, CFO

Assessment: The answer confirms what the arithmetic shows. The third-quarter expense line was partly timing. Three named Q4 items, one of them spending pulled forward from 2027, explain why the full-year range implies a fourth quarter roughly 15% to 26% above the third. Consensus that extrapolated Q3's $664M will need to come up.

Why take rates came in light, asset class by asset class

The question stated plainly that Q3 take rates were lighter than the Street expected and that the decline seemed to continue into October. The answer went through each asset class. Equities and options depend on product mix and spreads. In crypto, the stablecoin-based Robinhood Earn product is diluting the blended rate. In event contracts, the new dynamic pricing and Gold discounts change the result. The October read was "similar" and "maybe a touch lighter".

Q: "Can you flesh out a little bit more about what you're seeing in terms of the drivers on take rates as 3Q came in lighter than what a lot of us were expecting. It sounded like that declines continued to start the fourth quarter."
— David Smith, Truist Securities

A: "On crypto, the main driver was our Robinhood Earn, our stablecoin product. So that product is doing great and now has over $500 million of assets there. But because it's stablecoin, we monetize it less than the normal platform. And so that's what drove it down. If you take out the stablecoin, it was basically similar to the quarter before in the same ZIP code."
— Shiv Verma, CFO

Assessment: Each explanation holds on its own. Together they describe a company that is choosing volume and share over unit economics in three of its four trading lines. That is a reasonable choice for a platform and a negative surprise for a revenue model. Equity revenue per dollar traded fell about 10% from Q2, and event-contract revenue per contract fell about 8%. Management told us to expect October to be no better.

Equity stakes in a prediction-market rival while running a captive exchange

A pointed question on venue strategy asked how Robinhood reconciles routing between Rothera and the venue in which it now owns equity, and why it would hold stakes in a competitor to its own exchange. Management named Rothera the primary exchange and the new partner the main second exchange. It said most other venues will be dropped by early 2027, and that the equity stake monetizes flow that previously earned nothing extra.

Q: "You now own equity in two prediction market venues while running your own exchange with Susquehanna. How do you reconcile the routing decisions between the different venues? And what is the strategic logic of holding stakes and competitors to Rothera?"
— Patrick Moley, Piper Sandler

A: "What you should expect is -- rather will be our primary exchange and crypto.com/og.com will now be our main second exchange, and we'll be getting off most of our competitors. By the end of the year into early next year, you should expect the vast majority of our flow goes through both of those. From the financial standpoint, what we're doing now is, in addition to sending them flow, we're also becoming owners."
— Shiv Verma, CFO

Assessment: The structure is clever. Robinhood earns exchange economics through Rothera and equity upside through the second venue, and the cash cost is zero. It also removes the last argument for routing to independent venues. The open question is unchanged from July: Robinhood still has not given its economics per contract on Rothera versus a third-party venue, so the value of the move rests on management's description.

Rothera's pricing, its sports mix and third-party brokers

The question asked how Rothera plans to win share, how it prices, and whether it will onboard other futures commission merchants. The CEO separated Robinhood from Rothera's operations and positioned the exchange as a neutral venue for brokers that do not want to route to a competitor's first-party platform. The CFO gave the quarter's two most specific prediction-market numbers: non-sports contracts reached about 30% of the mix, and the pricing change kept the overall take rate the same.

Q: "I was hoping you might be able to just update us on your Robinhood strategy -- your Rothera strategy at this point. How do you plan to generate market share gains? And maybe you could speak to the pricing strategy at this point. And then maybe finally, maybe just an update on how you're thinking about adding other FCMs to the platform beyond just the hood brokerage."
— James Yaro, Goldman Sachs

A: "On the non-sports side, if you look in the quarter, it was actually about 30% was non-sports, and that was up from 15% the quarter before. ... We made it dynamic. We listened to customers, but we didn't impact our take rate because we were able to share some with them and take some of that back. And then on top of that, we made it even better with gold."
— Shiv Verma, CFO

Assessment: The mix shift toward non-sports contracts is the most important prediction-market data point on the call. It answers the bear case that the line is a sports-betting proxy with state-by-state legal exposure. The take-rate claim is harder to check against the reported numbers. Revenue per contract fell about 8% sequentially, so whatever was held on the pricing grid did not hold in the reported revenue.

How Robinhood Chain makes money, and how much

The question acknowledged that network congestion in late August and September raised fees, then asked about the financial opportunity and whether monetization can improve without making the chain less competitive. The CEO put developer activity and user experience ahead of fees. The CFO gave the quarter's numbers: roughly $50M of gross fees, about half shared with the underlying network, and "a couple of basis points" on notional in a normal environment.

Q: "Obviously, great to see early traction there. Also appreciate it benefited from some congestion in late August and September, which helped gas fees. So maybe seeing some normalization in fees there. ... And then just bigger picture, too, can monetization improve over time without making the chain less competitive?"
— Devin Ryan, Citizens

A: "On monetization, only piece I'll add is it was about $50 million in gross revenue in the quarter. We shared the schedule before. We shared some of it with Arbitrum. It was about 50%. So even on a net basis, about $25 million or $100 million annualized. We didn't call it a cylinder just yet. We want to give it more time to bake. You should expect in a normalized environment, it's probably a couple of basis points on notional values what we monetize."
— Shiv Verma, CFO

Assessment: About $25M of net revenue in its first quarter is far ahead of what we modeled for the chain in July, when we called it a strategic asset rather than a revenue line. Two qualifiers apply. The questioner said congestion helped, and management declined to call the figure a $100M business line. We treat Q3 as a high-water mark for fees, not a run-rate.

Whether ARPU can reach $300

The last question on the call asked whether ARPU, roughly doubled in two years, could reach $300 and what would have to go right. Both executives declined to set an ARPU goal. The CEO argued that faster international and younger-customer growth would pull the average down even as total revenue rose. The CFO called ARPU an output metric and named net deposits as the best predictor of revenue.

Q: "Keeping it sort of high picture, ARPU doubled over the last 2 years ... how do you see ARPU evolving from here? And if I pick an ARPU figure like $300, do you see this as feasible? And given all that you've got going on, what does Robinhood need to get right to get to a figure like that?"
— Ken Worthington, J.P. Morgan

A: "ARPU is an output metric. So we actually don't go to teams on that. We grow them on market share and then ROI. And what we've said before is we believe Robinhood can 10x from here. ... The highest predictor in correlation to revenue is just net deposits."
— Shiv Verma, CFO

Assessment: The answer is reasonable, and its timing matters. ARPU growth slowed to 5% this quarter from 24% in Q2, mostly because the crypto comparison from a year ago is hard. In July we called ARPU the cleanest proof of the wallet-share thesis. Management now calls it an output metric and points to net deposits instead, which grew at 17% annualized this quarter, below the 20% target. We will keep tracking both.

What They're NOT Saying

1. The Ohio exit, agreed the day before the call: On October 26 Robinhood agreed to stop new sports contracts in Ohio by November 2, after a cease-and-desist letter on October 2. The release does not mention it, the prepared remarks do not mention it, and no one asked. It appeared in the 10-Q filed the next day, together with the September Michigan exit and the adverse Sixth and Ninth Circuit rulings.
2. The net Robinhood Chain revenue in the filings: The release gives over $50M of chain fees, which the CFO called gross revenue, and he put the net figure at about $25M on the call. The 10-Q says only that crypto revenue was "partially offset by revenues earned from Robinhood Chain." Without a filed net figure, investors cannot see the crypto trading book on its own.
3. How much of net deposits is Trump Accounts money: Since July the deposit metric includes Trump Accounts contributions and Treasury seed funding. We checked the release, the 10-Q and the monthly metrics supplement, and none of them splits the amount out. That matters because net deposits are the metric management asks investors to judge it on.
4. A credit card loss rate: The 10-Q gives Q3 write-offs of $31M on purchased receivables, $9M of payments to the partner bank on off-balance-sheet receivables and $5M of recoveries, but no net charge-off rate. Our calculation is about 8% annualized. The release gives neither.
5. Securities lending's swing to a loss: The release cites "lower securities lending activity." The 10-Q shows the line went from an $87M contribution a year ago to a $19M loss. It was not discussed on the call, although it explains more than the entire year-over-year decline in net interest.
6. Why buybacks nearly stopped: Repurchases were $25M in Q3, against about $124M of program repurchases in Q2 (plus $290M bought alongside the convertible offering) and $250M in Q1, while the stock traded between about $87 and $125. No one on the call addressed it.
7. Rothera economics, for a third quarter: Rothera passed $100M of annualized revenue, but neither the release nor the 10-Q compares Robinhood's revenue per contract on Rothera with routing to a third party.

Market Reaction

• Pre-print setup: HOOD closed at $94.56 on October 27, down 16.4% year to date against a +12.8% S&P 500, down 20.8% over the trailing 30 days and down 35.2% over twelve months. The close was 24% below the September 3 closing high of $124.72. The 52-week closing range was $65.16 to $147.08. The options market implied a move of about 10% either way.
• After hours: Shares fell 3% to 5% in the first prints after the release, and the decline widened to roughly 7% to 8% by the end of the evening.
• Reaction session (October 28): The stock opened at $89.23, down 5.6%, traded between $89.15 and $92.60, and closed at $91.43, down 3.3% or $3.13. The S&P 500 rose 0.3%. Among peers, Coinbase rose 0.8%, Interactive Brokers 1.6% and Schwab fell 0.6%.
• Volume: 39.8 million shares against a 30-day average of 20.1 million, about 2.0 times normal.
• Since our July note: From the $86.60 close on July 30, HOOD is up 5.6% against 4.1% for the S&P 500, consistent with a Hold.

The session recovered more than half of the overnight loss. Buyers stepped in near the open, and the close was 2.5% above it. That shape fits a market that sold the headline misses overnight and then weighed them against a record quarter and a record October for deposits.

Coverage blamed the crypto decline, the GAAP EPS and revenue misses, and the softer October color. On the reaction day it added a miss on monthly active users against expectations, and the Ohio agreement once the 10-Q was filed. We weight the take-rate commentary most heavily. It is the only one of those items that carries into Q4 estimates, and it explains why the revenue consensus missed for a quarter in which equity, options and event-contract volumes all set records.

The sell-side response repeated July's. Thirteen published actions cut price targets, several others reiterated theirs, and none changed a rating. Targets are being reset toward the price, not abandoned, and most remain well above it. With the stock already 24% off its September high going into the print, the setup was not demanding, and the 3.3% decline is small for a quarter with a GAAP miss, a revenue miss and an adverse legal disclosure.

Street Perspective

Debate: Is the Drawdown an Entry Point?

Bull view: The stock is 24% below its September high on a quarter of record revenue and a record October for deposits. On the Street's 2027 estimates it trades near 31x, and the average price target sits far above the price. Every operating pillar is intact, and the one-off investment mark explains the EPS miss.

Bear view: Street 2027 estimates near $3.00 assume revenue growth of about 30% next year, after 12% this quarter and with take rates still falling. The fourth-quarter consensus has not absorbed the expense guide. On realistic estimates, 2026 EPS will be roughly flat on 2025's $2.05, and the multiple is in the mid-40s.

Our take: The bears have the better of the estimate argument and the bulls have the better of the franchise argument. Our 2027 estimate of about $2.45 is about 18% below the Street, and on it the stock trades at 37x. That is fair for a business with this growth and these optionalities, not cheap. The entry point is below our range, not at the middle of it.

Debate: Are Falling Take Rates a Choice or a Problem?

Bull view: Each decline has a deliberate cause. The stablecoin product dilutes crypto, dynamic pricing and Gold discounts lower event-contract prices, and lower-rebate ETFs dilute equities. Each choice buys volume, share and loyalty, and equity, options and event-contract volumes all set records.

Bear view: Three of four trading lines earned less per unit in Q3, management guided to tighter spreads in October, and the revenue miss came entirely from monetization. When a platform's growth comes from more intense use by fewer active traders, falling unit prices hurt more.

Our take: The bulls are right about the reasons and the bears are right about the arithmetic. We model per-unit revenue flat to slightly lower through 2027, which is one reason our estimates sit below the Street.

Debate: Will the Courts Shrink Prediction Markets?

Bull view: The federal regulator supports preemption and has sued several states, the Third Circuit sided with the industry, and the Supreme Court is likely to resolve the split. Non-sports contracts doubled to about 30% of the mix and are outside the state gaming fight, and Rothera's neutral-venue model gives Robinhood an advantage in either outcome.

Bear view: Two circuits now hold that state gaming law applies, dozens of states back that position, and Robinhood has exited sports contracts in three states. Class actions seeking to recover customer losses are multiplying. The business that just passed crypto in revenue could lose most of its sports volume state by state.

Our take: This is a binary legal outcome that the stock cannot price precisely, and the base case is more state exits before any Supreme Court ruling. We treat it as a capped risk. In a bad outcome, event-contract revenue would be cut, not eliminated, given the non-sports mix, and a favorable ruling would be an upgrade catalyst.

Debate: Is Engagement Peaking?

Bull view: Deposits are a better measure of engagement than active users, and October set a record. Banking became the top withdrawal destination before general availability, and nearly 70 million Trump Accounts give Robinhood a pipeline of future customers no rival has.

Bear view: Monthly active users fell 8% to the December 2025 level, new customers slowed, and the number of people placing options and crypto trades fell year-over-year. The growth is coming from heavier use by a stable core, and that depends on market conditions.

Our take: The bear data are more current, and management confirmed that deposit growth now leans on existing customers. We want to see Q4 active users and new customers recover before treating Q3 as a seasonal dip.

Model Update & Valuation Framework

Valuation framework: At the $91.43 reaction-day close and 911 million diluted shares, the market capitalization is about $83B. That is 45.7x our 2026 EPS estimate of about $2.00, and 41x to 42x the $2.17 to $2.22 Street range before the print. In July the stock traded at 44.4x the then-consensus of $1.95 for the same year. The multiple on 2026 earnings has not changed. On 2027, the stock trades at 37x our $2.45 and 31x the Street's roughly $3.00. The difference is mostly revenue: we expect about 20% growth in 2027 and the Street about 30%.

Fair value: July's $80 to $95 range equalled 41x to 49x the then-consensus of $1.95 for 2026. We apply that band to our own, higher $2.00 estimate, which gives $82 to $97. As a cross-check, 34x to 40x our 2027 estimate gives $83 to $98. The bull case is the Street's 2027 estimate at 38x, about $114, which needs the 30% revenue growth and a favorable Supreme Court outcome. The bear case is $2.20 at 30x, about $66, which assumes further take-rate pressure and the loss of most sports volume. At $91.43 the stock sits near the middle of our range, with about 25% upside to the bull case against about 28% downside to the bear case. That symmetry is the case for Hold.

Thesis Scorecard Post-Earnings

Grading the July Triggers

In July we wrote: "Upgrade to Outperform on any of: net deposits returning to a 20%-plus annualized pace in Q3; crypto printing flat or better sequentially; Rothera economics disclosed at a take-rate premium to third-party routing; or the stock trading below $75. Downgrade to Underperform on: net deposits sustaining a sub-15% annualized pace through Q3; an adverse regulatory or legislative development on event contracts; or credit-loss provisioning passing 6% of revenue."

One trigger fired in each direction, and we net them. We did not state in July how simultaneous triggers would net, so we say plainly how we applied them. Each trigger was written as sufficient on its own, and two sufficient conditions pointing in opposite directions cancel. The crypto trigger fired on the line as reported. On the trading book alone it narrowly missed, and we do not use that to discount it. The regulatory trigger fired without qualification. Our valuation work, which puts the stock near the middle of a fair-value range we derived independently, gives the same answer.

Overall: The thesis is unchanged in balance and more two-sided. The franchise is larger and more diversified than in July: record volumes, a chain that earns money, Trump Accounts at national scale and Banking that holds customers' cash. The quarter also produced three new negatives: falling unit prices, visible card losses and a turn in the courts on the largest part of the prediction-market business. The binding constraint is now unit economics, and net deposits are second.

Action: Hold, conviction unchanged at 6, fair value $82 to $97. We list every change to the triggers. The crypto, regulatory and two Q3 deposit triggers have been used and are retired. The Rothera, $75-price and 6% provision triggers stay as written. We extend the two deposit triggers to Q4 and remove the one-time Dell contribution from the test; our own Q4 estimate on that basis is 17% to 19%. We add three new triggers: a favorable Supreme Court ruling, an adverse one, and a sequential decline in event-contract revenue. If triggers fire in both directions again, they offset. Upgrade to Outperform on any of: net deposits at a 20%-plus annualized pace in Q4 excluding the Dell contribution; Rothera economics disclosed at a take-rate premium to third-party routing; the stock trading below $75; or a Supreme Court ruling that federal law preempts state gaming law for event contracts. Downgrade to Underperform on any of: net deposits below a 15% annualized pace in Q4 excluding the Dell contribution; credit-loss provisioning passing 6% of revenue; event-contract revenue falling sequentially in Q4, a quarter with midterm elections; or an adverse Supreme Court ruling on event contracts.