Tokenized stocks can't take a $10,000 order

On Robinhood's own blockchain there is a token that stands for one share of Netflix. Last month 3,387 people traded it. The largest order any of them got filled, in a month, was $7,385.

That is not a rule anyone wrote down. Nobody is stopped at the door. It is what the market could bear. Ask it for $1,000 of Netflix in one go and it costs about 13% to get in. Ask for more and the number stops meaning anything, because there is nothing on the other side to trade with.

The thing being traded has a name. It is a tokenized stock, and there are 424 of them on public blockchains. Bitquery indexes every trade on every chain they live on, so we pulled 30 days of the complete record, 27.1 million trade legs, and measured how much money the market will take in one order before the price runs away. A volume figure never shows that. The answer is small, and the headline hides it.

What a tokenized stock is

A tokenized stock is a crypto token that stands in for one share of a real company. An issuer buys the share, holds it, and sells you a token against it. The token then trades on a blockchain, which means it moves in seconds, settles without a broker, and never closes for the night.

Five issuers carry almost all of it. Binance sells bStocks on BNB Chain. Backed sells xStocks, mostly on Solana. Robinhood issues its own tokens on its own chain. Ondo and Backpack Securities do the same across several chains. Between them they cover Apple, Nvidia, Tesla, the S&P 500 and the Nasdaq 100, plus a few companies you cannot buy on a normal exchange at all, like SpaceX.

The pitch is access. Buy Nvidia at 3am, from anywhere, in whatever size you want, without a brokerage account. The first three of those hold up.

01 A market made of pocket change

The typical tokenized-stock trade last month was $45. Roughly two thirds of them were under $100, and orders big enough to matter to anyone were rarer than three in every ten thousand.

That $45 is a count of trades, not a picture of the money. A quarter of all trades are under $10, and weighing every trade equally lets that dust set the number. Weigh by dollars instead and the middle of the market sits near $762.

Both readings are true and they answer different questions. Most trades are tiny. Most of the money still moves in clips of a few hundred to a few thousand dollars, which is exactly the range the next section shows this market cannot absorb.

The money piles up in a few names, too. One token, Binance's tokenized Nasdaq 100 fund, is over a third of the entire market on its own. The 10 largest are more than four fifths of it. Everything else divides what is left.

Counting listings tells you nothing about any of this. Ondo has issued more tokens than anyone and carries less money than any other real issuer. Backpack Securities runs a dozen tokens and does 20 times Ondo's business.

02 The part that works

Most of this machine works. The cost of a small round trip, measured inside each pool between what buyers paid and what sellers received in the same minute, sits at about 10 basis points while US markets are open. Overnight it is the same. At weekends, with every stock exchange on earth shut, it is the same again. For a few hundred dollars this is as cheap as a retail broker, which is not what anyone expects from a market that never closes and has almost nothing standing behind it.

Fewer people show up when Wall Street is closed. Money moving per hour falls by more than half from the trading session to the weekend, and the typical trade shrinks with it. The ones who do show up are not charged extra for the hour.

Prices hold up as well. Against the last printed price of the real share, the middle of the range is close to zero in every session, so there is no standing mark-up on being on a blockchain. Gaps that open close again quickly. Half are gone inside 10 minutes and three quarters inside half an hour.

03 Where it stops

Now raise the amount. A $10,000 order is not an institution. It is one person moving a real slice of their savings into a company they want to own.

Three contracts out of 424 can take that for less than 1%.

Depth is the money it takes to shift a price by 1%, and every pool leaves a record of its own. Each trade nudges the price by an amount that depends on how much is sitting behind it. Read enough nudges and the pool tells you its size.

The gap between the top of that table and the rest is the whole story. Binance's Nasdaq token takes about a million dollars to move its price 1%. Robinhood's SpaceX token, two rows down, takes a twentieth of that. Backpack's SanDisk token needs less again, so a $10,000 order in it costs more than half of what you put in.

This is why the volume number misleads. Backpack's SanDisk token traded $35.7 million last month, and those dollars really did change hands. You could not have been one of them.

04 The ceiling

Across every tokenized stock, on every chain, for a month, 38 trades were larger than $100,000. The biggest of them all went through Binance's S&P 500 token.

That ceiling explains why you will not find a pile of victims here. Go looking for people ruined by bad execution and there is not much to find, because the trades that would have ruined them never happened. Whatever this market costs its users is paid by the ones who sized an order, read the warning, and closed the tab.

You cannot measure the trades that did not happen. They are most of the story.

None of this makes tokenized stocks a con. Settlement really does run at all hours, the prices really do track the shares, and small orders really are cheap. But "trade Apple 24/7" and "move real money in Apple 24/7" are different promises, and only one of them has been kept.

What we could not measure

Checking a price cannot prove a contract is genuine. A competent fake prices itself to track the real token, which is exactly what makes it convincing. Counting them on BNB Chain alone found 1,307 impersonator contracts. Contracts we could not verify against their issuer are marked in the table above, and they are 4% of the money we measured. One of them, a "QQQB" on Solana, had 3 traders in a month and a single wallet behind almost all of it. It is excluded from every headline figure here, along with 2 others that failed the same test.

Some tokens rest on very few people. Backpack's SK Hynix token traded $169 million with fewer than 2,000 traders behind it, and one wallet accounts for a third of that. On Solana as a whole, a single address is 17% of all tokenized-stock trading. Those totals are real, and they are not the same thing as a broad market.

Costs above roughly $10,000 are projections. They come from a straight-line model of how price responds to size, and real markets bend, so the true cost at size is likely worse than shown rather than better. We checked every estimate at four different sampling windows and dropped the ones that moved.

The record

The contracts behind the figures above.

Method

Window pinned to 26 July to 24 August 2026, 30 complete days. The source table keeps a rolling month and deletes as it goes, so we copied it to disk before touching it. Trade data covers BNB Chain, Solana, Robinhood Chain, Ethereum, Base and Polygon, through the same DEX trade APIs the docs publish. Reference share prices come from 5-minute bars including pre-market and after-hours.

Tokens were identified by comparing each candidate's on-chain price against the real listed price of the ticker it claims. Genuine ones cluster tightly on the real share. Impersonators miss by orders of magnitude and fall out. Two tickers passed the price test and still had to be thrown out by hand, both matching Canadian mutual funds rather than the companies they appeared to name.

Three corrections shaped the totals. Nearly a fifth of raw trade rows were exact duplicates. A further quarter of the remainder were single swaps routed through two pools, which the record writes down twice. And the sender on a blockchain trade is a routing contract, not a person, so trader counts here use the wallet that signed. Applied together these take the headline down by about a third from the raw figure.