Cross-venueKilled

Arbitraging tokenized US stocks against Hyperliquid's stock perps

0.4 to 10 bps

the price gap between the two venues, against the 50 bps it would need to be worth trading

bps = basis points = a hundredth of a percent (100 bps = 1%).

In plain English

We tried to profit from the price gap between tokenized US stocks and Hyperliquid's version of the same stocks. The gap is tiny and professional traders already keep it closed, so there was nothing left to capture.

What was claimed

The same US stocks trade in two crypto venues at once, so buy the cheap one, sell the rich one, and pocket the gap.

The bar it had to clear

a price gap wider than the cost to trade it (over 50 bps)

What the data said

0.4 to 10 bps, the pros already keep it pinned tight

Why it's dead

spreads.fi routes orders to xStocks, tokenized US equities on Solana. The idea was a basis arb between those tokens and Hyperliquid's stock perps. A live snapshot during US market hours showed the gap between the perp's mark and its oracle (the price it trades at versus the reference price it tracks) running 0.4 bps on AAPL to 10 bps on MU, 5 to 100x tighter than the 50 bps it would need. Professional desks already keep that basis pinned.

The detail

Timing check passed (both programmes live), but a magnitude check (is the gap even big enough to be worth trading?) killed it before any capital. Live snapshot at the NYSE open: mark-versus-oracle gaps of 0.4 to 10 bps across 9 large-caps, a 25x spread within the universe, all far below the 50 bps bar. A side catch: the venue's growth mode had already cut the taker fee about 90% (to roughly 0.45 to 0.9 bps), so an earlier 4.5 bps cost assumption was 5 to 10x too high, but even near-free trading cannot manufacture a gap that is not there.

Kill date: 2026-04-30
Sample: live snapshot, 9 symbols
Method: Documented kill
Verdict: already arbitraged