NYSE’s owner is backing 24/7 stock tokens without Wall Street’s usual price rules

The company that owns the New York Stock Exchange wants people to be able to buy stocks 24/7. Given the trading volume crypto and other tokenized assets have seen in the past year or so, the decision to introduce stock tokens comes as no surprise.

People want to trade all the time, and now institutions want a piece of that. What surprises people is that the new 24/7 exchange won't use NYSE prices to determine what stocks cost once the tradfi market closes.

The price people pay for off-market trading will differ based on how much stock others have put up for sale in a DeFi-style trading pool, not a new bid on NYSE.

The experiment isn't coming from some tiny crypto startup trying to get Wall Street's attention. OKXICE is owned by Intercontinental Exchange (ICE) and an affiliate of OKX.

In its public notice, OKXICE said stock tokens will trade around the clock using technology behind Uniswap, with buyers dealing against pools of assets rather than the familiar lines of bids and offers.

Sixteen companies had already objected to third-party tokenized versions of their stocks trading on OKXICE as of Oct.19, including Microsoft, Alphabet, and Goldman Sachs.

Fifteen of them were on the exchange's original proposed roster of 62 companies, so roughly 25% of that list had objected. The sixteenth, Cerebras, wasn't on the original roster. The rest hadn't objected publicly by then, which doesn't amount to support.

Same shares, different price

A tokenization provider would acquire ordinary shares through a registered broker-dealer and issue one corresponding digital token for each share held, with the underlying stock backing the tokens one for one.

The tokens would also come with arrangements covering dividends, voting, and other shareholder rights.

Under the SEC exemption issued in September, eligible tokenized shares must provide rights equivalent to the conventional shares they represent. That's why a blockchain receipt that just imitates a stock's price doesn't qualify under this program.

Ordinary US exchanges use order books, which collect offers to sell and bids to buy. Those prices are part of a wider system that connects exchanges through market data and rules meant to prevent trades from going through at inferior prices when better-protected quotations are available elsewhere.

Stock prices can differ between exchanges, but rules govern how those differences are handled.

In OKXICE's new arrangement, traders would deposit stock tokens and stablecoins into an automated pool, and a buyer would trade against the pool's inventory.

Software adjusts the price as the token-to-dollar balance shifts, so a single purchase can make each additional fraction of a share more expensive than the last.

DeFi protocols have used versions of this model for years, but it's not how a brokerage account typically buys Coca-Cola stock. On OKXICE, the stablecoin side could include USDC, and the proposed pools would operate on XLayer with Uniswap v4 software.

Traders would still need identity checks and permission to trade, so this isn't an anything-goes corner of DeFi dressed up as a stock market.

OKXICE's notice says the exchange won't be subject to Regulation NMS, the framework that includes some of the ordinary price protections connecting US equity markets.

The SEC has granted qualifying tokenized stock exchanges temporary, conditional relief from registering as conventional exchanges for five years, along with requirements covering shareholder rights, permitted trading activity, issuer objections, and formal trading halts.

Antifraud rules still apply, and brokers handling covered customer orders can have their own best-execution obligations. Those legal boundaries show that the token can represent the same ownership rights while its exchange follows different pricing rules.

How $1,000 can buy only $909 worth of stock

Take a hypothetical pool containing 100 stock tokens and $10,000 in stablecoins, with an initial price of $100 per token. At the beginning, that matches the assumed price of the stock on a regular exchange, so the two markets appear to agree on what one share is worth.

Put a $1,000 buy order into that pool, though, and the math suddenly becomes much more interesting. In a normal purchase at a fixed $100 price, that money would buy ten shares, ignoring fees.

In this pool, the purchase adds dollars and removes tokens, so the software sets progressively higher prices as the buyer drains the available inventory. Under a basic constant-product formula, which is the pricing family described in OKXICE's notice, the customer ends up with about 9.09 tokens rather than ten.

That works out to an average cost of $110 per token. The stock's reference price hasn't moved from $100 in this example, and the holder has received tokens representing roughly $909 worth of shares at that reference price despite spending $1,000.

The last token in the transaction is more expensive than the first, and the pool's price for the next tiny purchase has climbed to around $121. The model charges more as the order consumes inventory.

Give that pool ten times as many shares and dollars, and the same $1,000 purchase would average roughly $101 per token. Give it a hundred times as much, and the average would fall to about $100.10.

Larger pools absorb orders with less price movement, assuming the other conditions stay the same, which is why the amount available to trade is as important as the starting quote.

Original hypothetical calculations using a full-range constant-product pool and a stablecoin worth $1, with fees, outside price movements, and arbitrage excluded.

OKXICE's pools let providers concentrate their money within selected price ranges, which can make trades near the current price much cheaper than the simple model implies. The trade-off is that liquidity concentrated inside one range can run out, leaving a bigger order to reach into another range with less inventory and a worse price.

Two pools with the same total dollar value can therefore offer very different deals to the person pressing Buy.

In the simple model, a $1,000 purchase requires $100,000 of starting stablecoins to keep its average price within 1% of the assumed $100 reference, while reducing that markup to 0.1% requires $1 million.

Those numbers show why a banner displaying the total money deposited into a venue tells a customer much less than knowing how much of that money is available close to the price they want.

The platform plans to show estimated execution terms, fees, and the gap between the reference stock value and the customer's expected purchase price.

Users will also be able to set a slippage tolerance, essentially telling the software how much worse the quoted deal can get before it should refuse to execute.

That's useful when the pool moves while a transaction is being submitted, but it isn't a promise of the best available stock price, and the exchange warned that a rejected blockchain transaction can still incur a network fee.

The Sunday problem Wall Street usually sleeps through

Company news doesn't care about the stock exchange's opening hours. A CEO can resign on Saturday, a court can deliver news over the weekend, or a geopolitical crisis can unfold before trading opens on Monday morning.

Anyone who wants to act on that information has limited options for trading the underlying US shares when the main market is closed, even though some brokers already offer overnight sessions and other ways to trade beyond ordinary exchange hours.

Being able to react immediately could be truly valuable, especially if waiting until Monday brings more price volatility.

However, it's hard to find someone willing to take the other side of that trade at a reasonable cost. Liquidity providers selling tokens on Sunday risk discovering that the underlying stock is worth far more than expected on Monday, while those buying tokens face the opposite risk.

With the main equity market closed, both could struggle to buy or sell the ordinary shares needed to offset the exposure.

And yet OKXICE's smart contracts won't use a stock-price feed from the NYSE or elsewhere to determine the pool's execution price. The company says it will display outside market prices for reference, but the exchange software will use only the assets in the pool.

Someone checking a stock's last conventional quote could see one number while the available token transaction produces another, and the gap can widen when not enough traders keep the two markets in line.

Professional traders can exploit differences like that, at least in theory. If a token costs more than the ordinary share, they can try to buy the cheaper stock, convert or hedge that exposure, and then sell the more expensive token.

Their selling adds tokens to the pool and tends to pull its price down toward the outside market. Do the reverse when the token gets too cheap, and the gap can narrow from the other direction.

A lot of things could complicate this trade and make it harder or even impossible to execute: the underlying shares could be hard or impossible to obtain immediately for the hedge to be successful, the tokenization provider might have a long transfer and redemption time, and the cost of hedging could be more than the profit that could be made from the arbitrage.

That's why the market could end up with a large price discrepancy, with no one actually able to make money from it.

While stock ownership may have moved onto a blockchain, the ordinary shares backing it are still connected to financial institutions that don't necessarily operate 24 hours a day.

Traders could also face formal trading halts, as OKXICE must be able to stop trading a token when the underlying stock starts having problems.

The platform doesn't plan to apply additional external reference-price bands or circuit breakers beyond its stated stoppage rules, and liquidity providers aren't required to keep their assets in the pool when conditions get uncomfortable.

The ability to execute a trade at 3 a.m. therefore says little by itself about how attractive its price will be at that hour.

Why a quarter of the proposed companies have already objected

The updated 19 issuer list includes Microsoft, Alphabet, Goldman Sachs, Broadcom, UnitedHealth, and Verizon, among others.

Fifteen of OKXICE's 62 originally proposed companies had objected, with Cerebras bringing the overall issuer count to 16 despite not being on that initial list. Microsoft, Alphabet, and the other familiar names are exactly the sort of stocks a new trading venue would want people to recognize.

The objections aren't a vote against blockchain in general, and they don't necessarily stop another exchange from offering eligible tokenized shares under a different arrangement.

Under the SEC's rules, when an unaffiliated third party wants to tokenize a stock for this kind of venue, the issuer can submit an objection that prevents that specific proposed offering.

Most objectors haven't explained their reasoning, so treating the list as proof of a coordinated fight against crypto would overreach the evidence.

Only Applied Optoelectronics provided a useful glimpse of the concern: a company representative reportedly told The Wall Street Journal that it wasn't sure how the trading might affect the stock's price, liquidity, or volatility, and didn't want investors to mistake the third-party tokens for a product the company had sponsored.

It's easy to see why a public company's investor-relations team would want control over whether its name appeared on something that looks like an official new market for its shares.

The NYSE isn't alone in pursuing tokenized stocks, but its competing visions are strikingly different. The SEC has approved Nasdaq's tokenized-securities rule for a DTC-linked pilot that brings eligible securities into an existing exchange structure instead of sending their trades through independent pools.

Nasdaq's approach keeps familiar stock-market mechanisms in place, while OKXICE proposes using an entirely different engine to make the trade happen. The approvals themselves don't establish that every contemplated feature has launched.

OKXICE also proposes a way to test its pricing: pool balances and completed trades will be visible on-chain, along with published transaction details and daily pool-size information.

Once trading begins, this could let you compare the true cost of buying and selling the same stock token on a Sunday, during a busy weekday, and as the ordinary market reopens.

Actual liquidity data would also show whether the model's theoretically expensive trades happen in practice or whether enough money shows up to keep prices competitive.

That means on Sunday, buyers would get the shareholder rights attached to a familiar stock at whatever price a separate pool offers at that hour.

Even with $100 displayed as the reference price, spending $1,000 wouldn't guarantee ten shares, and the cost of buying immediately would depend on the inventory other traders were willing to supply.