Hype fades; structure remains.
Gate says its equities product cleared $90 billion in notional volume over four months. The disclosure contains one integer and zero methodology. No statistical basis. No window start or end date. No active user count. No revenue figure. No fee schedule. No legal entity. No license number. The number cannot be cross-checked, because the venues that validate brokerage flow — exchange tape, DTCC clearing records — have no visibility into a centralized exchange's internal ledger, and the vendors that do track crypto volume, Kaiko and CCData, do not audit equity books.
One operator-reported figure, unverifiable by construction. That is the whole dataset. Everything below is arithmetic run against that figure, and I will label confidence where it matters.
Centralized exchanges have been expanding product categories for years, and the strategic logic here is not mysterious. Crypto revenue is a beta function of the crypto cycle. A venue that can point to a second, semi-uncorrelated revenue stream tells a better story to market makers, to institutional counterparties, and — if a listing is ever contemplated — to underwriters. The announcement says nothing about GT, the platform token, not once. That silence is itself informative: this is a capital-markets narrative, not a product narrative.
The product is not new either. eToro has run crypto-plus-equities since the last decade. Robinhood moved the other direction, equities into crypto. Coinbase began U.S. equity trading in 2025 behind a domestic license stack. Kraken bought a retail brokerage to acquire permission rather than build it. Gate's only structural advantage is a pre-existing crypto user base and the conversion funnel attached to it. The moat is distribution. Not engineering.
Here is where the analysis has to go hard, because the most loaded phrase in the release — "equity volume" — is doing work that three different business models could each claim.
Path A: real brokerage. Orders routed to a partner broker-dealer such as Alpaca, DriveWealth, or Interactive Brokers, with the user holding actual shares. Technically this is order routing plus clearing integration. Compliance cost: maximum. Business model: genuine.
Path B: tokenized equity. On-chain receipts representing shares or ETFs, issued through a partner like Backed or Dinari. Technically this is issuance plus in-house market making. Legally it is a security, and in the EU it falls under MiFID II — not MiCA, which explicitly excludes securities. Regulatory exposure: maximum.
Path C: contracts for difference. The platform is the counterparty. The user holds nothing. Technically the simplest. Economically a leveraged position against the house. And critically: notional volume here inflates with leverage of 5x to 20x, meaning $90 billion of tape could correspond to $4.5 billion to $18 billion of actual client capital.
The disclosure does not distinguish between them. One phrase covers three models with completely different risk profiles, capital requirements, and legal exposure. That is not a detail gap. That is the entire question.
Now the arithmetic that matters more than the headline.
Run the fee comparison. Crypto spot taker fees on a centralized exchange sit near 5 to 10 basis points. Retail equity trading in mature markets has compressed to zero commission in the U.S. — funded by order flow — and to a fraction of a basis point or a flat monthly fee in Europe. Apply a blended monetization of 0.5 to 2 basis points to $90 billion and you get $4.5 million to $18 million of revenue across four months.
Apply 5 basis points to the same $90 billion in crypto spot and you get $450 million.
Same notional. Roughly 1% to 4% of the revenue. Efficiency is not empathy, and it is also not income. A venue can push enormous notional through a zero-fee equity book and book almost nothing for it.
Which means the value transmission chain to the platform token is, on current evidence, broken. Exchange buyback and burn mechanics are typically indexed to platform revenue, not platform volume. If the equity book contributes single-digit millions against a crypto business contributing hundreds, the burn math barely moves. Reading this announcement as bullish for GT is a category error — conflating a vanity metric with a value anchor.
Scale it against the real competitive set. Four months is roughly 122 days. That is $7.4 billion a day, $225 billion a month, about $270 billion annualized. Top-tier retail brokerages move hundreds of billions in equity notional per day. Gate's equity business therefore lands somewhere in the range of 1% to 3% of a leading retail broker. Ninety billion dollars is a large number in crypto. It is a small number in equities. The headline exploits the gap between two coordinate systems — a standard move in cross-sector comparison, and it works because almost nobody converts the units.
One more layer the release leaves silent: whether the volume was bought. Crypto venues routinely seed new product lines with zero fees, rebates, and stock giveaways. Under that model, a meaningful share of the $90 billion could be market-maker flow and incentive farming rather than retail demand. No independent user count, no per-user volume, no repeat rate — precisely the three metrics that would separate organic adoption from subsidized tape.
The narrative being sold is that crypto traders are rotating into equities. Consider what that sentence means if it is true.
It means capital is leaving crypto. The same fact, viewed from the crypto market's side, is an outflow signal — and the release packages it as a platform triumph. Both readings are consistent with the data. Neither is proven by it. Equally supported alternatives: U.S. equities were in a bull run pulling in every retail cohort; the venue subsidized flow; or the flow is professional market making. The report gives no overlap data, no net position changes, no fund-flow figures.
I learned this pattern in 2017, auditing 45 ICO whitepapers by hand and finding 38 with no technical differentiation. The market priced the story, not the artifact. The lesson was not that founders lie. It was that a number without a methodology is a rhetorical device wearing a lab coat.
Then there is the compliance silence, louder than all of the above.
A venue that holds a securities license advertises it in the first paragraph. This disclosure names no entity, no jurisdiction, no regulator, no broker partner. Offering retail equity exposure to users in the U.S. implicates broker-dealer registration and FINRA membership. In the EU, MiFID II. In Hong Kong, an SFC Type 1 license. Assembling that stack from zero takes two to four years. This product has run four months. The structure is therefore almost certainly an offshore entity fronting for a licensed partner — Gate supplies the interface, the partner supplies the permission, and revenue splits accordingly, which further compresses margin.
That structure has a legal property worth stating plainly: a public claim of $90 billion in equity volume is a public claim of operating a securities business at scale. If a regulator later alleges unregistered securities activity, this announcement becomes evidence. Code does not feel. Neither does a subpoena.
On user protection — equity balances held on a centralized exchange book are generally not covered by SIPC or any equivalent investor protection fund. If the venue fails, the holder is an unsecured creditor. That asymmetry appears nowhere in the text.
What to watch, concretely. Three data points would make this announcement analyzable rather than merely loud: independent user count, realized revenue contribution, and the identity of the licensed counterparty.
None are present. Their absence is the signal. A venue with a license to display would display it.
The question for the next cycle is not whether a crypto exchange can sell equities. It is whether a channel-integration business running at 1% to 4% of crypto's revenue density can ever be more than a retention feature dressed as a growth curve. Check the retention rate in six months, not the notional volume. The tape is cheap. The user is not.
