---
ticker: "HOOD"
company: "Robinhood Markets, Inc."
date: "2026-10-20"
view: "Avoid"
fair_value: 64.14
price: 98.54
price_source: "market data quote (intraday; previous close $96.52 on 2026-10-19)"
price_time: "2026-10-20T17:58:43Z"
price_vs_fair_value: "+53.6%"
bear: 14.18
bull: 143.21
exit_price: 79.53
p_price_justified: "under 20%"
restated: "2026-10-26"
method: "case-spread-2026-10-24"
original_fair_value: 64.14
original_view: "Avoid"
page: "https://alpha.dming.si/research/2026-10-20-hood"
---

# HOOD · Robinhood Markets, Inc. · Avoid

Research report as of 2026-10-20, partly unblinded. Page: https://alpha.dming.si/research/2026-10-20-hood

We value HOOD at $64.14 a share. The report's $98.54 price (Oct 20, 2026) was 54% above fair value: Avoid.

_Restated Oct 26, 2026 (method case-spread-2026-10-24); the report's original value was $64.14, Avoid._

| Lens | Range | Marks |
| --- | --- | --- |
| Our report | $14.18 to $143 | fair $64.14; original $64.14 |
| Our simulation | $33.08 to $116 | median $59.97 |
| Street targets (35) | $110 to $170 | average $139, median $138 |
| Price | | $98.54 |

Bands run from our bear to bull (the hollow mark is the original $64.14), our simulation's 10th to 90th percentile and the Street's low to high. Street targets forecast future prices, not value today. Market data, captured Oct 20, 2026.

| Figure | Value |
| --- | --- |
| Fair value | $64.14 |
| Price | $98.54 (market data, Oct 20, 10:58 AM PT) |
| Five-year return | -0.7% (a year, at the report's price) |
| Bear / bull | $14.18 / $143 |
| Exit price | $79.53 (on Dec 31, 2031) |
| Confidence | 35% |
| Price against our scenarios | The price is justified in under 20% of our scenarios (5,000 simulated outcomes, each drawing the uncertain drivers from ranges we chose: our judgement, not a measured chance) |

_This text argues the original call. The current figures are at the top._

As of 2026-10-20 · partly unblinded, revised after review (unblinded) · blind commit `6b3639c7c` (made as `91e6cd240`, rebased unchanged) · market pass after it

**Original call, Oct 20, 2026 (before the Oct 26, 2026 restatement)**

_As of 2026-10-20 · partly unblinded · revised after review_

- **Fair value today:** $64.14 per diluted share (cost of equity 8.88%, beta 0.854); the blind value, set before we saw the price, was $46.65 (record.json blind_fair_value)
- **Price:** $98.54, market data quote, 2026-10-20T17:58:43Z (intraday; previous close $96.52) → fair value 34.9% below the price
- **Five-year exit price:** $79.53 on 2031-12-31: the average of the long fade ($83.96) and the outside view ($75.10, at the peers' 14.5x EV/EBIT); fading over 10 or 20 years instead gives $76 or $83. A cross-check formula (the value-driver formula) gives 83.4x against the fade's 19.2x year-six NOPAT (net operating profit after tax); see section 3
- **Five-year return:** -0.7% a year from today's price on our central case, payouts and net cash included (bear -28.1%, bull +14.0%); below the 12% a year we require in the central and bear cases, above it only in the bull
- **Purchase price:** $56.40: the price at which the central case earns our required 12% a year
- **Bear / bull today:** $15.95 / $122.89, each a single consistent business outcome (weights 25/50/25 give a case-weighted $66.78); the bull is above the price
- **Confidence:** 35%: our chance that, after Q3 2026 results, central fair value is still within 15% of $64.14 (subjective)
- **Moat:** Narrow: the cheapest, fastest-shipping retail app with 28.6M funded customers, but no lock-in beyond habit, and nearly half of revenue comes from market-maker rebates (payment for order flow) and event contracts (contracts that pay out on a real-world outcome) that a regulator can cut
- **Market expects:** revenue to keep pace with platform assets: about 1.57% of them every year to 2031 ($14.6B of revenue in 2031 against our $10.4B), or our 2031 EBIT valued at 25.6x instead of the peers' 14.5x
- **Price needs:** no single headline driver reaches the price within its range (best: Fed funds 6% by 2031 gives $83; cost pass-through 10% gives $82; pass-through is the share of each added revenue dollar that becomes cost)
- **Odds:** P(price justified): under 1%, the share of our simulated scenarios in which fair value reaches the price (simulated 10th-90th percentile $42-$73, median $59; the trading yield on assets is not a simulated driver, see section 7)
- **Key driver:** the Fed funds rate (the US central bank's policy rate) path: 2.0% by 2031 is -$8.12 a share, 4.2% is +$6.75 (central 3.2%, the FOMC's longer-run median); then trading revenue over platform assets (0.65% in 2031 central)
- **Street:** average target $138.60 (median $138, range $110-$170; 14 targets in the last month averaging $140.43, up from $129.22 over the year); 2027 revenue $6.9B and EPS $3.03 (21 analysts); an expectations reference, not validation

**We rate Robinhood an Avoid at $98.54.** Our fair value is $64.14 a share,
34.9% below the price, and the central five-year return is -0.7% a year
against the 12% a year we require. The price is right only if Robinhood's
revenue stays near 1.57% of the money customers hold on its platform ("platform
assets") every year to 2031. We expect that ratio to fall to 1.12%, because it
has already fallen from 2.26% in 2023 to 1.74% in 2025 and to about 1.4% in the
first half of 2026. In our simulated scenarios the odds that fair value reaches
the price are under 1%, but our bull case is above the price.

**How to read this report.** Each report values a company three ways. The
central case is our best estimate and decides the call. The bear case is a
severe but plausible bad outcome; the bull case is the matching good outcome.
Cost of equity (8.88%) is the yearly return investors require for holding the
stock, and we discount future cash to today at that rate. Beta (0.854) measures
how strongly the stock moves with the market and sets the cost of equity. The
five-year exit price is our estimate of the stock's value on 2031-12-31, the
average of two estimates: the "fade", which lets the company's return on new
capital drift toward its peers' over time, and the "outside view", which values
the company at what similar companies trade at (here 14.5 times EV/EBIT:
enterprise value, the value of the business before its debt, divided by
operating profit before interest and tax).

**What "blind" means here.** We first value the company without seeing its
market price. That blind value ($46.65) is committed in the repository at
`6b3639c7c`. The filings we read were masked: passages that could reveal the
market price were hidden. The report is partly unblinded because the masked
10-Q still shows the $2.2B convertible notes (bonds that can convert into shares)' 12.6M shares (a conversion price
near $174.6, set at a 65% premium to the 2026-06-22 close), because the script
that sets the peers' exit multiple read market data key-metrics (market data is the market-data
vendor), and because we chose the beta having seen both candidates.
assumptions.md lists each; none entered the blind value of
$46.65. The market pass (market.py, market.txt,
odds.json) changed no input. The revision below did: it was made
after we saw the price and two outside reviewers' findings, so every changed
number in this report is unblinded (see "Revision after review").

## Revision after review: three corrections lift fair value from $46.65 to $64.14

The blind model (commit `6b3639c7c`, central $46.65, bear $14.13, bull
$73.95) went, with the market data, to two outside reviewers whose job was to
find its flaws. Our chief investment officer (CIO) ruled on each finding. Three
were accepted and applied; the rest are sensitivity lines or were rejected. All
values below come from model.py and output.txt, which
reproduces the blind values before applying any change.

**Disclosure.** All three accepted changes were made after the price was seen,
and all three raise value, which is the direction of the price. The beta and
the convertible corrections fix departures from our written method; the third,
trading revenue over platform assets, is a judgement on the shape of trading
revenue, taken from filing evidence (10-K 2022, 10-K 2025, 10-Q Q2 2026), not
from the price, and it is the largest of the three: $10.24 of the $17.49. Read
the $64.14 as the result of a correction made with the price in view, and the
$46.65 as the blind value.

### Dollar bridge, central case, in the order applied

| Step | Change | Per share | Value |
|---|---|---|---|
| 0 | Blind central (the rerun reproduces it) | | $46.65 |
| 1 | Beta: the Blume-adjusted peer median 0.854 (method default) in place of Damodaran's Brokerage & Investment Banking levered 1.17, which is at 136% debt to equity; HOOD's debt weight is 2%. Cost of equity 10.20% to 8.88% | +$6.92 | $53.57 |
| 2 | Convertible: the 0% note due 2029-10-01 at present value at 5.29% ($1.89B) in place of $2.2B principal, still counted once as debt | +$0.34 | $53.91 |
| 3 | Trading revenue over average platform assets decays from 0.81% (2027) to 0.65% (2031) in place of 0.81% to 0.51% | +$10.24 | $64.14 |
| | Total | +$17.49 | $64.14 |

Bear: $14.13, then $15.61 (beta, +$1.48), $15.95 (convertible, +$0.34),
$15.95 (trading revenue over platform assets, unchanged: see below). Bull:
$73.95, then $86.68 (beta, +$12.73), $87.02 (convertible, +$0.34), $122.89
(trading revenue over platform assets, +$35.87). Case-weighted (25/50/25):
$45.35 to $66.78.

**Beta (accepted).** The 1.17 we first used is Damodaran's beta for his
brokerage industry, which is heavily indebted ("levered"); HOOD is not. The
method's default is the Blume-adjusted median beta of the peer set fixed in
assumptions.md (the Blume adjustment pulls a measured beta toward 1): 0.67 x
0.782 + 0.33 = 0.854. Sensitivity lines on the new central: Damodaran levered
1.17 gives $55.79 (-$8.36); Damodaran unlevered (debt stripped out),
cash-corrected and relevered to HOOD, 0.69, gives $69.93 (+$5.79) (the 0.69 is
the outside reviewer's figure and we did not rederive it); HOOD's own market data beta
2.354, Blume-adjusted 1.907, gives $43.69 (-$20.46).

**Convertible (accepted).** The note pays no coupon, so its worth to the
equity is the discounted principal. At 5.29% (the method's risk-free rate, the
floor for a cost of debt) over the 2.95 years from 2026-10-20 to 2029-10-01 that
is $1.89B. The review gave $1.92B; the difference is $0.03B, $0.03 a share, and
we used the figure the stated date gives. The share count is untouched: the note
converts only above $174.60 and no case's exit reaches that.

**Trading revenue over average platform assets (accepted as the central
path).** The blind model drove trading revenue from the number of customers
times activity per customer and let platform assets grow on their own, so
trading over assets fell from 0.81% (2027) to 0.51% (2031) while assets per
customer more than doubled.

We recomputed the history from the filings in `data/` (trading revenue over the
average of opening and closing platform assets; assets $98.0B, 62.2, 102.6,
192.9, 322.1 at year-ends 2021 to 2025, 10-K 2022 and 10-K 2025):

| Year | 2022 | 2023 | 2024 | 2025 | H1 2026 annualised |
|---|---|---|---|---|---|
| Transaction revenue ($B) | 0.814 | 0.785 | 1.647 | 2.628 | 1.399 (H1) |
| Over average platform assets | 1.02% | 0.95% | 1.12% | 1.02% | 0.81% (0.66% without event contracts) |

The outside reviewer's claim holds: about 1.0% in 2022 to 2025. Over 2023 to 2025
trading per average funded customer rose 2.98x ($33.8 to $100.7 per customer)
and platform assets per average customer 2.78x; the outside reviewer gave 2.9x and
2.7x, and the small gap is our use of year-end customer counts (23.0M, 23.4M,
25.2M, 27.0M; the 2022 count is net cumulative funded accounts, a different
definition). Two limits on the history: 2025 platform assets include $51.8B
acquired, of which $42.5B is RIA (registered investment adviser) assets (TradePMR) that do not trade through
Robinhood, so 2025's 1.02% is understated against 2022 to 2024; and the H1 2026
ratio is the only one with event contracts, which are a new revenue line. So
some decline from 0.81% is real (event contracts are a cyclical, regulated
line, and H1 2026 is 0.66% without them), but a fall to 0.51% needs trading per
customer to stand still while the filings show it moving with assets.

New central: 0.81%, 0.77%, 0.73%, 0.69%, 0.65% for 2027 to 2031. New bull:
at least 0.81% every year (its own per-customer path starts above, at 0.91% in
2027 and 0.85% in 2028, and is not cut). The factors are fixed numbers on the
case, so every per-driver sensitivity (activity, crypto, event contracts,
rebates) still moves trading in proportion. Costs rise with the added revenue
through the unchanged 35% pass-through, 0.5% credit cost and 3% inflation.

We kept the bear at its own trading path (0.68% in 2027, about 0.49% from 2028). Its
story is a 2022-style winter with activity at 0.55 to 0.75 of Q3 2026, rebates
cut 20% and event contracts at 35%; those cuts are structural and justify a
ratio below the central, but the real 2022 held 1.0% because trading and
assets fell together, and the bear's assets recover to $546B by 2031 on a
0.49% ratio, which is below H1 2026 without event contracts (0.66%). We judge
the bear conservative here and kept it because the CIO's ruling was to keep it
unless its story says otherwise and the story only partly does. With the
central yield path as a floor the bear would be $22.22 (+$6.26), and the
case-weighted value +$1.57.

### Other findings change nothing: the central case stays at $64.14

| Finding | Rule | Effect on central | Reason |
|---|---|---|---|
| Payout convention | Free cash kept at 2.8% after tax and paid at exit | $62.72 (-$1.43) | Full payout at the cash-flow date is the stated convention and a sensitivity of the method's own; the authorisation covers about a tenth of it. The outside reviewer's line is a small cost of the convention. |
| Tax rate | 18% in place of 22% | $66.29 (+$2.14) | 22% is a normalised rate (federal 21% plus state, less the stock-pay benefit; assumptions.md). Reported effective rates were 10.7% (2025) and 17.9% (H1 2026), lowered by stock-pay tax benefits and, in Q2, the release of the California valuation allowance; we have no evidence that either lasts. The release also means the net deferred tax asset in the bridge ($0.179B, 10-K 2025) may be stale; we did not rebuild it, and $0.1B is $0.11 a share. |
| Operating leverage (how much of each added revenue dollar becomes cost) | Cost on each added revenue dollar 0.30 in place of the model's | $67.76 (+$3.62) | The model's incremental cost is the 0.35 pass-through plus a 0.5% credit cost, not 0.40; we show 0.30 against it, and 0.45 gives $56.91 (-$7.23). The 0.35 is anchored to 2025 (12% of added revenue) and the 2026 cost guide; no dated evidence independent of the price moves it. |
| Inflation | 2% in place of 3% | $66.59 (+$2.44) | No dated evidence independent of the price; 3% is the cost base's rate in the blind model. |
| Five other findings | Passed or rejected by the CIO | none | Not itemised here. The exit multiple stays at the peers' 14.5x, which our method fixes in advance (we judged 1.00x of it). |

## 1. Robinhood earns from trading, interest on customer balances and subscriptions

Robinhood is paid three ways (2025, 10-K; H1 2026, the first half of the year, 10-Qs; US$ billions):

| Engine | 2025 | H1 2026 | What moves it |
|---|---|---|---|
| Trading (options, equities, crypto, event contracts, other) | 2.628 | 1.399 | customers x activity per customer x rebate per unit, set against about 1% of platform assets in 2022-25 |
| Net interest (margin, customer cash, sweep, card, stock lending) | 1.514 | 0.748 | balances x the Fed funds rate plus or minus a spread |
| Subscriptions and services (Gold, a paid subscription; proxy; other) | 0.331 | 0.228 | subscribers x fee |
| Total revenue | 4.473 | 2.375 | |

Trading pays mostly through payment for order flow and exchange-style fees,
so it rises and falls with retail risk appetite: transaction revenue fell 42%
in 2022 (crypto -52%, equities -59%, options -29%; 10-K 2022) and rose 60% in
2025. Net interest moved the other way in 2022 (+66%), because rates rose as
trading fell. That offset is the business's main defence, and it ends if the
next downturn comes with rate cuts.

Customer money is not Robinhood's. At 2026-06-30 it held $12.0B of segregated
customer cash, a $21.6B margin book funded by customer free credits and stock
loan, and $29.7B swept to partner banks off its balance sheet. Only $5.36B of
cash is corporate. The bridge counts that cash, less an operating need of
$2.0B (about four times the required net capital of Robinhood Securities (RHS), its broker-dealer subsidiary) and the $2.2B convertible
at its present value of $1.89B: $2.16B, $2.37 a share. Everything held for
customers enters only through the spread it earns.

## 2. Our Q3 2026 estimate and what would break it

Q2 2026 (8-K 2026-07-29): revenue $1.308B, transaction $0.776B (options 0.342,
event contracts 0.156, equities 0.129, crypto 0.100), net interest $0.389B,
other $0.143B; diluted EPS $0.62, helped by a $0.106B gain on deconsolidating
Robinhood Ventures Fund I (RVI), its listed closed-end private-markets fund.
Funded customers 28.4M, Gold 4.84M, platform assets $368.7B, net deposits $21.7B (28% annualised). The guide for adjusted operating
expenses and stock-based compensation (SBC) was cut to $2.675-2.775B.

Our Q3 2026 estimate, built from the July, August and September 1-24 monthly
volumes (company figures quoted in search results; the releases themselves could not be fetched):

| Line | Estimate | Confirms the case | Breaks it |
|---|---|---|---|
| Total revenue | $1.42B | $1.38B or more | under $1.30B |
| Options | $0.400B (908M contracts at $0.44) | rebate per contract $0.42 or more | under $0.40 a contract: rebates compressing |
| Equities | $0.133B ($997B at 1.33 bp) | 1.25 bp or more | |
| Crypto | $0.120B | | under $0.09B: a deeper crypto winter than the central 1.35 normal |
| Event contracts | $0.164B (14.3B at 1.15 cents) | $0.14B or more and no state ban in force | under $0.12B, or a ruling that closes sports contracts in several large states (events kept at 35%: -$7.08 a share) |
| Net interest | $0.430B | $0.41B or more with the margin book at $21B or more | under $0.39B at a higher Fed rate: the spread fit is wrong |
| Net deposits | about $14.6B | $11B or more (12% annualised) | under $9B: below the bear's 2027 rate |
| Gold subscribers | about 5.1M | 5.0M or more | under 4.95M |
| Adjusted opex and SBC | about $0.725B, guide held | guide at or under $2.775B | guide raised above $2.775B without revenue to match (pass-through 45%: -$7.23) |
| Diluted EPS | about $0.55 | | |

Any one "breaks" cell is a revision, not a verdict; two in the trading lines
together mean the activity normal (our assumption that trading per customer settles at 0.88 of Q3 2026) is too high. Transaction revenue
annualised over average platform assets under 0.75% in Q3 says the central
decay to 0.65% is too slow (H1 2026: 0.81%).

## 3. Our forecast and how it becomes $64.14

The model is model.py; its output is output.txt;
inputs and labels are in assumptions.md (written blind; the
revision changes beta, the convertible and the trading yield, recorded above).

Central (US$ billions):

| Year | Revenue | of which trading | Net interest income (ex corporate cash) | EBIT | EBIT margin | NOPAT | Reinvestment |
|---|---|---|---|---|---|---|---|
| 2026 | 5.15 | 3.13 | 1.54 | 2.13 | 41% | 1.66 | 0.31 |
| 2027 | 6.32 | 3.63 | 2.15 | 2.83 | 45% | 2.21 | 0.59 |
| 2029 | 8.35 | 4.86 | 2.82 | 3.81 | 46% | 2.97 | 0.69 |
| 2031 | 10.36 | 6.01 | 3.55 | 4.73 | 46% | 3.69 | 0.73 |

2025's EBIT margin, before interest on corporate cash, was about 44% (10-K
2025, inference); H1 2026's 40%. The central case holds margins there: costs
rise 3% a year plus 35% of each added revenue dollar. Trading per customer
normalises to 0.88 of Q3 2026, because Q2 and Q3 2026 are high-activity
quarters, and trading over platform assets decays from 0.81% to 0.65%. Platform
assets grow from $400B to $994B on net deposits fading from 18% to 10% a year
and a 6% market return, and revenue per dollar of assets falls from 1.74%
(2025) to 1.12%.

Exit price: we use our default method, enterprise value on EBIT (operating
profit before interest and tax). We considered the variant used for financial
companies and did not take it (recorded in `Forecast.changes`): Robinhood does
not lend deposits it owes, and the broker peers' EBIT also carries net
interest. The outside view is what similar companies trade at: 14.5x, the
median of the ten-year EV/EBIT medians of Charles Schwab (SCHW), Interactive
Brokers (IBKR), LPL Financial (LPLA) and Cboe (CBOE); HOOD has too few years of
its own. Our method fixes that multiple in advance, and we judged HOOD deserves
exactly it (1.00x). The fade starts at the central volume-only return on new
capital (56%) and fades to the peers' median return on tangible equity (23%,
which the model calls the mature return). The fade ($83.96) is 12% above the
outside view ($75.10), inside the 30% gap beyond which our method requires an
explanation. The
value-driver formula gives 83.4x against the fade's 19.2x year-six NOPAT
because growth after 2031 (8%) is close to the cost of capital (8.78%), where
the formula's denominator nearly vanishes; it is a check that does not fit a
business still growing at year five, and we do not move the fade toward it.
Consistency check: central NOPAT growth in 2031 is 10.1% against 8% after (7.6% in
year seven). The bull's is 22.2% against 11% after, a sharp step down, so the
bull's exit may be low rather than high; we did not change it.

Sensitivities, each alone (output.txt):

| Change | Value | Change per share |
|---|---|---|
| Fed funds 2.0% by 2031 / 4.2% | $56.02 / $70.89 | -$8.12 / +$6.75 |
| Trading left at its per-customer path (0.51% of assets in 2031) | $53.91 | -$10.24 |
| Event contracts kept at 35% | $57.06 | -$7.08 |
| Cost pass-through 45% / 25% | $56.91 / $71.39 | -$7.23 / +$7.24 |
| Order-flow rebates cut 20% from 2028 | $58.56 | -$5.58 |
| Growth after 2031 5% / 11% | $60.26 / $68.73 | -$3.88 / +$4.58 |
| Net deposits 12% / 24% in 2027 | $60.96 / $67.80 | -$3.18 / +$3.65 |
| Activity normal 0.78 / 0.98 | $60.98 / $67.31 | -$3.17 / +$3.16 |
| Crypto normal 1.0 / 1.7 | $61.93 / $66.36 | -$2.21 / +$2.21 |
| Mature return 14% / 30% | $62.57 / $64.77 | -$1.57 / +$0.62 |

The rate path matters more than the cycle because net interest is 30-39% of
revenue and almost all of it falls straight to profit, while a trading slump
also shrinks the cost base through the pass-through. Beta, payout, tax,
operating leverage and inflation lines are in the revision section.

## 4. The bear case is $15.95 and the bull case $122.89

Both were set from history before any price.

- **Bear, $15.95.** A 2022-style winter (the 2022 trading downturn) that this time comes with cuts:
  activity falls to 0.55 of Q3 2026 in 2028 (the 2021-2023 fall was -44%),
  crypto to 0.60, platform assets -25% in 2027, the margin book to 3.8% of
  assets, Fed funds to 2.0%, event contracts barred in most states (35% kept)
  and order-flow rebates cut 20%. Revenue falls from $5.15B to $3.19B in 2028
  and recovers to $4.60B by 2031; EBIT margin 17-25%; mature return 14%
  (IBKR's), growth after 2031 4%. The 2022 offset from rising rates is
  absent, which is what makes this worse than 2022. Its trading yield was not
  changed (see the revision).
- **Bull, $122.89.** A lasting upturn: activity 1.20 in 2028, crypto twice
  today's rate, net deposits 24% falling to 15%, platform assets $1.43T by
  2031, event contracts unrestricted and growing 40% then 10%, and trading
  holding at least 0.81% of platform assets. Revenue $16.95B in 2031 at a 52%
  EBIT margin, close to the price-implied path (section 7); mature return 30%
  (SCHW 2018-2025), growth after 2031 11%. The bull rose from $73.95 mainly
  on the trading yield (+$35.87) and the beta (+$12.73).

The simulation behind the odds (drivers.json) carries the same pair as two
alternative worlds at a quarter each, with activity, crypto, margin and
market return correlated.

## 5. Stock pay is counted as a cost; the model pays out all free cash

Stock pay (stock-based compensation) was $0.305B in 2025 and $0.197B in H1 2026 (10-K, 10-Q); it is
counted as a cost. The board authorised $1.5B of buybacks in March 2026 over
about three years; H1 2026 repurchases were $0.664B, of which $0.290B was tied
to the convertible offering. The model pays out all free cash flow (about
$12.3B over 2026 H2 to 2031, central), so the value does not depend on
whether it goes out as buybacks; keeping it at 2.8% after tax and paying it at
exit is $1.43 lower.

Deals verified in the filings: TradePMR (closed 2025-02-26, cash and about 2M
restricted shares; net assets $0.169B; consideration masked in our copy),
Bitstamp (2025-06-02, cash; net assets $0.224B; consideration masked), MIAXdx
through the 90%-owned Rothera venture with SIG (2026-01-20, $0.079B), WonderFi
(2026-06-01, $0.178B cash), and PT Buana Capital Sekuritas and PT Pedagang
Aset Kripto in Indonesia (agreed December 2025, pending). RVI was
deconsolidated on 2026-06-25; the retained stake ($0.437B) is in the bridge.
Reinvestment carries $0.20B a year of acquisitions (bear $0.10B, bull $0.30B).

## 6. Regulation: event contracts and order-flow payments are the exposed lines

- Event contracts: suits by NV, NJ, MA, MI, WA, WI, KY, CT and tribes; new
  sports contracts stopped in Nevada; the Third Circuit ruled for Kalshi in
  April 2026 (10-Q). Central keeps 90% of the unrestricted path; the bear
  35%.
- Payment for order flow: best-execution suits and a New York Attorney
  General inquiry (10-Q). Central keeps rebates; the bear cuts them 20%.
- Crypto: Bitstamp gives licences in the EU and elsewhere; no specific
  charge beyond the crypto cycle.
- A $0.05B a year regulatory and legal charge (bear $0.15B) stands for
  settlements like 2022-2025's.

## 7. The price needs revenue of 1.57% of platform assets every year; the Street is above us

At $98.54 the market pays about $90B for the diluted shares (912M). Holding
our central operating path and cost of equity, the price needs an exit price
of $133.04 on 2031-12-31, against our $79.53: $121B of enterprise value,
25.6x our 2031 EBIT of $4.73B, or 14.5x an EBIT of $8.4B, 1.8 times ours
(market.txt).

The single figure that closes the gap is revenue per dollar of platform
assets. Our central lets it fall to 1.12% by 2031. The price needs about 1.57%
held every year to 2031, with costs on our own rules: $14.6B of revenue and
$7.4B of EBIT in 2031. The history (revenue over average assets under custody
or platform assets; 10-K 2022, 10-K 2025, inference):

| Year | 2021 | 2022 | 2023 | 2024 | 2025 | 2026 (ours; H1 annualised about 1.38%) |
|---|---|---|---|---|---|---|
| Revenue over average assets | 2.25% | 1.70% | 2.26% | 2.00% | 1.74% | 1.43% |

| Floor on revenue over average assets, each year | Fair value | Against the price |
|---|---|---|
| none (central, falls to 1.12%) | $64.14 | -34.9% |
| 1.20% | $69.88 | -29.1% |
| 1.40% (about 2026's rate) | $84.90 | -13.8% |
| 1.74% (2025's rate) | $111.51 | +13.2% |

The Street (Street analyst estimates, read 2026-10-20) is above us on both
revenue and margin through 2029:

| Year | Street revenue | Ours | Street EBIT | Ours (ex corporate-cash interest) | Street EPS |
|---|---|---|---|---|---|
| 2026 | 5.33 | 5.15 | 2.57 | 2.13 | 2.20 |
| 2027 | 6.90 | 6.32 | 3.93 | 2.83 | 3.03 |
| 2028 | 8.13 | 7.35 | 4.86 | 3.34 | 3.66 |
| 2029 | 10.11 | 8.35 | 5.17 | 3.81 | 4.44 |

Run through our model: the Street's revenue with our cost rules gives $82.42
(+$18.28); with the Street's EBIT as well, $87.87 (+$5.44). Even then the
price is $10.7 higher, which is the exit multiple: the Street's average target
of $138.60 needs HOOD valued well above the 14.5x of SCHW, IBKR, LPLA and CBOE
in 2031. On HOOD's own market data beta, 2.354 (Blume-adjusted 1.907, cost of equity
13.30%), our value falls to $43.69, and the Street's path gives $59.78. The
Street's 2030 revenue (9 analysts) of $9.14B is below its 2029 and below our
$9.39B.

The odds (odds.json): no single headline driver reaches the price within its
range; the simulation (Monte Carlo) puts P(fair value at least the price) under 1%, with a
10th to 90th of $42 to $73 and a median of $59. That median sits 8.2% below
the central value because several simulated input ranges (triangular distributions) are one-sided by design
(order-flow rebates can only be cut, event contracts can only be restricted).
The trading yield on platform assets is the largest driver the simulation
does not draw: it is worth $10.24 between the per-customer path and the
central decay, and $35.87 in the bull, so the simulation understates the
upside from it and the odds are a floor. The narrow range our report check flagged (the last of the four questions
answered below) is also because the drivers that move value most for the market, the revenue
yield on assets, the exit multiple and the cost pass-through below 25%, are not
simulated as wide as the market's own view; the blind ranges came from history
and were not widened toward the price.

One of our checklist questions asks whether a risk is charged twice. We checked three places. The
central case charges regulation once on revenue (event contracts kept at 90%)
and once as a $0.05B yearly legal cost; these are different items, and
together they were worth $1.53 a share in the blind model (not recomputed).
The bridge deducts the convertible at present value and does not also dilute
for it below its $174.60 conversion price. The beta, where the blind model
charged a levered industry figure to a company with little debt, is now the
method's peer median, with the Damodaran levered, Damodaran unlevered and
HOOD's own betas shown.

## 8. The case for the market's price, argued as hard as we can

1. **Assets, not trades, are the business now.** Robinhood is turning into a
   wealth platform: retirement accounts, RIAs through TradePMR ($50B),
   managed Strategies, banking, the card, Gold at 17% of customers and about
   40% of new ones. Platform assets grew from $102.6B at end-2023 to $368.7B
   at mid-2026. If each new dollar of assets carries the services the old ones
   do, revenue scales with assets and our 1.12% is too low; at 2025's
   1.74% our own model gives $111.51, above the price.
2. **Costs are fixed and the platform is built.** The 2026 cost guide was cut
   at Q2 while revenue grew; the Street's EBIT margin of 57-60% in 2027-28 says
   incremental revenue keeps most of itself. Our 35% pass-through plus 3%
   inflation charges every new product as if it needed its own team. At 25%
   pass-through our value is $71.39, at 0.30 $67.76, and with the Street's
   revenue and EBIT it is $87.87.
3. **The new products are options on large markets.** Event contracts went
   from 8.5B to 13.6B contracts a quarter in six months; the Third Circuit
   ruled for Kalshi in April 2026. Prediction markets, tokenised stocks in the
   EU, Bitstamp's institutional business and Asia each could be a business the
   size of today's options line, and our model gives most of them nothing
   beyond the Q3 run rate and a growth fade.
4. **The multiple is a growth multiple, not a broker's.** SCHW and IBKR are
   mature; HOOD grew revenue 52% in 2025 and 24% in H1 2026 (market data) while
   compounding customers and assets. A company still growing 20% a year in
   2031 deserves more than 14.5x, and our growth after 2031 of 8% may be low.
5. **Rates may stay high.** The Fed raised to 3.75-4.00% in September 2026.
   If inflation keeps it near 4%, net interest is $6.75 a share more than our
   central.
6. **The cost of equity could be lower.** At Damodaran's unlevered 0.69, +$5.79.
7. **Trading may hold its yield.** If trading holds 0.81% of platform assets,
   as it held about 1.0% in 2022 to 2025, that is the bull's world at
   $122.89, above the price.

Put together: the Street's operating path ($87.87) is within $11 of the
price, and a 0.69 beta or an exit above the peers' multiple closes the rest.
Each of these is plausible alone.

**Why we still say Avoid.** The central and the case-weighted values sit 32-35%
under the price, and the price needs the bull's world, which we weight 25%
and the simulation, with its blind ranges, puts under 1%. The history runs the
other way on the one that matters most: revenue per dollar of assets has
already fallen from 2.26% (2023) to 1.74% (2025) and to about 1.4% in H1 2026,
so "holding 1.57%" asks for a reversal, not a continuation. The growth that
would justify a higher multiple is mostly trading in a strong tape (Q2 2026
options contracts +32% on Q1), the same revenue that fell 42% in 2022; the
Street's 2030 revenue (9 analysts) is already below its 2029. HOOD's own beta
of 2.354 says the market itself prices it as far riskier than a broker. The
view is weaker than the blind one: the revision moved the central 37% toward
the price and the bull is above it, so evidence that the trading yield holds
moves us out of Avoid. **What settles it:** revenue over average platform
assets in Q3 and Q4 2026, and the trading share of it. Two quarters of total
revenue above 1.55% annualised, or transaction revenue above 0.81% of average
assets, move us materially toward the price; at or below 1.40% total and 0.75%
trading the central case stands.

## Four questions our report check raised, answered

Our automated report check flags results that need an explanation. Four applied here.

- **How much did the revision move the value (+$17.49, +37.5%)?** The largest change is the
  trading yield on platform assets, +$10.24 of the $17.49; it rests on the
  10-K 2022 and 10-K 2025 ratios recomputed above, not on a management claim
  or a Street figure. Beta is +$6.92 and the convertible +$0.34.
- **What world makes the price right, given odds under 1%?** Trading
  holds 0.81% of platform assets and total revenue holds about 1.57% of them
  every year to 2031: $14.6B of revenue and $7.4B of EBIT in 2031 (market.txt),
  or an exit at 25.6x on our EBIT. The bull is close to that world ($16.95B,
  52% margin, $122.89), and the Street's own path ($87.87) is within $11 of
  the price. We weight the bull 25%; the simulation does not draw the trading
  yield, so its P is a floor.
- **Why is our fair value 0.50x the Street's average target discounted one year at our cost of equity ($127.30)?** Driver
  by driver, through our model (section 7): the Street's revenue with our cost
  rules +$18.28 to $82.42; the Street's EBIT as well +$5.44 to $87.87; the
  remaining $10.7 to the price and the further $39.4 to the target discounted
  one year ($127.30) is the exit multiple above the peers' 14.5x and the
  Street's longer growth, neither of which we can source. Revenue is the
  largest driver: the Street's 2029 revenue of $10.1B against our $8.35B, and
  its 2030 figure of $9.14B is below its own 2029.
- **Why is the price outside our 10th to 90th percentile range?** Stated in section 7: the
  ranges came from history; the trading yield, the exit multiple and pass-through
  below 25% are not drawn.

## 9. What we could not source

- The September 2026 monthly release (and clean July and August texts): the
  pages refused a fetch, so Q3 rests on search-quoted company figures.
- The TradePMR and Bitstamp purchase prices (masked in our copies).
- The capped call's cap price (masked); valued at nil.
- Damodaran's unlevered, cash-corrected beta of 0.69: the outside reviewer's figure,
  not rederived.
- The post-release California deferred tax asset: the bridge uses the 10-K
  2025 balance.

## 10. Sources, dated

- SEC EDGAR: 10-K 2025, 10-K 2022, 10-Q Q1 and Q2 2026, 8-Ks
  2022-11-02 to 2026-07-29 (`data/filings/`, fetched 2026-10-20, masked).
- Q2 2026 call transcript, Quartr via market data (masked).
- Market data statements, HOOD and peers (fetched 2026-10-20, `data/`).
- Federal Reserve, FOMC statement 2026-09-16 and September 2026 projections.
- Damodaran, US industry betas, January 2026 (a sensitivity line after the
  revision).
- Market pass: market data quote, profile and analyst estimates for HOOD, fetched
  2026-10-20 (`data/market/`).