Analysis: Wintermute’s $10 Million Annualized Race for Trading Priority Has Limited Ties to Its Cumulative Profit

A few milliseconds can be valuable to a crypto market maker—but paying for a speed advantage does not explain every dollar it earns.

On November 15, blockchain analyst Psyche reported that Wintermute had spent approximately $824,000 over the preceding 30 days on Hyperliquid’s Gossip Priority auctions. At that pace, the expenditure would amount to roughly $10 million over a year. That is an extrapolation, not a forecast: the firm’s bidding activity has already changed.

The more important distinction is between the cost of priority access and the profit attributable to it. Wallets identified by third-party labels as belonging to Wintermute have accumulated approximately $229.7 million in account profits since 2024, according to the analysis. Most of that total predates the auction mechanism, and the remaining profit cannot be assigned to priority access alone.

What Hyperliquid’s Priority Auction Offers

Hyperliquid is a platform for onchain trading, including perpetual futures. Its official documentation describes the protocol and its trading infrastructure.

According to Psyche’s analysis, the Gossip Priority mechanism launched in April 2026 and holds an auction every three minutes for three priority slots. The highest-ranked slot can receive information about transactions that have not yet executed roughly 60 milliseconds before participants without priority access.

For a market maker, that interval can create an opportunity to cancel an existing quote or adjust prices before an incoming order is executed. It is a form of informational priority, not a guarantee that the recipient will profit or that its own trades will execute first.

The arrangement has similarities to last look in traditional financial markets: a liquidity provider gets an opportunity to reassess a quote when faced with an incoming trade. The comparison is useful for understanding the economic incentive, though the mechanisms and market structures are not necessarily identical.

Why the $229.7 Million Figure Needs Context

The headline spending figure can make priority access look like the explanation for Wintermute’s historical returns. The timeline suggests a narrower conclusion.

Figures are based on the analyst’s third-party wallet labels and are rounded. They should not be read as an audited breakdown of Wintermute’s business performance.

The pre-launch profit shows that most of the cumulative total arose before the mechanism existed. Nor is the post-launch amount a clean measure of what priority access earned. The labeled accounts also hold substantial perpetual futures positions used for hedging, so their reported profit and loss can reflect market making, directional exposure and inventory management.

Likewise, comparing the $25.6 million in post-launch profit with the $3.58 million in auction spending does not produce a reliable return on that spending. It does not establish what the accounts would have earned without priority, or isolate other trading costs and positions.

The defensible interpretation is that faster access may improve market-making economics at the margin. Its contribution cannot be determined from the aggregate account figures alone.

Bidding Is Concentrated—and Not Constant

The analysis suggests that access is dominated by a small number of sophisticated firms. Wintermute, Selini Capital and one unnamed market maker reportedly accounted for approximately 99% of auction spending.

Wintermute focused primarily on slot 0 and won about 74.7% of the relevant auction rounds in October. But its behavior subsequently changed. From November 11, its daily expenditure reportedly fell from around $30,000 to about $10,000, with bids concentrated in the hours after the US stock market opened.

That shift matters when interpreting the annualized figure. Multiplying a recent 30-day expense by a year is useful for expressing scale, but it does not mean Wintermute will spend $10 million over the next 12 months. Auction prices, bidding frequency and the perceived value of priority can all change.

The reason for the reduced bidding is not established by the figures alone. It could reflect a change in strategy or market conditions, but the observed spending pattern does not identify the cause.

What This Means for Traders

The auction raises a question for decentralized exchange users: who gets to see trading information first, and how does that affect execution?

For someone submitting a large market order, early visibility may give a participating market maker time to adjust or withdraw a quote. The eventual fill could therefore differ from what the trader expected when placing the order. That does not prove every large order receives a worse price because of the mechanism; execution also depends on liquidity, volatility, order size and other participants.

For smaller market makers, the concern is competitive access. Firms unwilling or unable to pay for priority may have less time to react to incoming activity, potentially placing them at a disadvantage when quoting prices.

Traders concerned about execution can review available liquidity, consider limit orders where appropriate and avoid assuming that a displayed quote will remain unchanged for a large trade. Limit orders offer price control, but they also carry the risk of partial execution or no execution at all.

Finally, the reported wallet attribution relies on third-party labels rather than Wintermute’s official disclosure. The spending and profit figures are therefore best treated as an analytical estimate—not a verified account of the firm’s complete trading activity.