102 cross-exchange spreads crossed the scanner today, led by a 41.67% USTC gap between Binance Futures and Hyperliquid and a 36.50% MAVIA gap on the same venue pair. Most of the field was ordinary CEX-to-CEX noise on BANK, LAB, AKE, DEXE, TAG, NIGHT and LA — tighter, more executable, and far less exciting.
102 events on the board today. That's a normal day's worth of cross-exchange noise, and 100 of them behaved like it — single-digit to low-double-digit spreads scattered across the usual CEX pairs. The top two did not behave like it. USTC printed a 41.67% gap between Binance Futures and Hyperliquid, and MAVIA followed with 36.50% on the exact same venue pair. When two of your biggest prints of the day share a venue pair, that's not coincidence, that's a pattern, and it's covered below.
One flag before the numbers: every volume field in today's feed — pump volume, dump volume, buy pressure, sell pressure — came back at $0.0M. That means this scan captured price dislocation, not confirmed fill size. Nothing below should be read as "this much size was available at this price." Pull a live order book before wiring capital against any of it. That's not a disclaimer, it's the single most important sentence in this report.
USTC — 41.67% spread. Buy Binance Futures at $0.005368, sell Hyperliquid at $0.007605. This is the day's headline number and it's also the one most likely to disappoint a trader who reads it as a spot flip. Binance Futures and Hyperliquid are both perp venues — there's no token transfer that captures this gap, only a basis trade: long USTC perp on Binance Futures, short USTC perp on Hyperliquid, collect the convergence and whatever funding accrues along the way. USTC lost its peg mechanism years ago, so there's no fundamental anchor forcing these two marks back together — they converge because arbers push them together, not because they have to. Liquidity risk is real at a $0.005 handle: Hyperliquid's USTC book is thin enough that a mid-five-figure order would move it alone. No fill-size data came through the feed, so assume the book is smaller than the spread makes it look. Verdict: executable as a small, actively-managed basis trade — sub-$10k per leg — not as a fire-and-forget arb.
MAVIA — 36.50% spread. Buy Binance Futures at $0.027260, sell Hyperliquid at $0.037210. Same venue pair as USTC, same mechanics, same caveat: this is a perp-to-perp basis trade, not a token transfer. Heroes of Mavia is a gaming-narrative token that swings hard on sentiment, so part of this gap may simply be one venue's mark catching up to a move the other already priced in, rather than a durable dislocation. Hyperliquid's MAVIA book is a fraction of what a name at this cap needs to support real size — probe orders only. Window duration wasn't logged in this feed, but structurally this class of spread (perp-DEX vs CEX futures, on a mid-cap with thin dual-sided books) tends to persist longer than spot spreads precisely because so few traders are set up to run the hedge — that's exactly why it hasn't already been arbed to nothing. Verdict: tradeable in small size for someone already running a Binance Futures / Hyperliquid basis book; not a candidate for a first-time arb attempt.
BANK — 16.41% spread. Buy Bitget at $0.146510, sell Bitunix at $0.158387. This one is closer to a clean spot flip, assuming BANK trades as spot on both venues with pre-positioned inventory on each side. The catch is Bitunix: it's a tier-2 exchange with materially thinner books and slower withdrawal processing than Bitget, and a 16% gap against a tier-2 venue is as often a symptom of thin liquidity as it is free money — the ask you'd actually fill against on Bitunix is probably worse than the quoted $0.158387 once you eat through a few levels of book. Verdict: executable in modest size (low four figures) with pre-funded balances on both sides; not executable at scale without moving the Bitunix book yourself.
LAB — 14.17% spread. Buy OKX at $0.118100, sell Binance Futures at $0.126000. This is the most "legitimate-looking" spread in the top five: two genuine tier-1 venues, both with deep books at the exchange level, disagreeing on a single token's price by an unusually wide margin. Capture path is a cash-and-carry style trade — buy spot LAB on OKX, short LAB perp on Binance Futures — collecting the basis as it closes plus or minus funding on the short leg, so check the Binance Futures funding rate before entering; a negative funding print eats into this fast. Because both venues are deep at the exchange level, the actual slippage constraint here is LAB's own liquidity, not either exchange's overall depth. Given two tier-1 venues rarely disagree this much for long, expect this one to have compressed quickly — if you didn't already have the hedge on, it's likely gone by now. Verdict: solid basis-trade candidate for anyone who can enter both legs within seconds of each other.
BANK — 12.37% spread. Buy Binance Futures at $0.141355, sell KuCoin at $0.147380. BANK's second appearance in today's top five, this time across a different venue pair and a tighter gap than the Bitget/Bitunix print (16.41%). Four venues — Bitget, Bitunix, Binance Futures, KuCoin — disagreeing on BANK simultaneously is the real story here, not any single spread: that's fragmented price discovery, not a one-off blip. Mechanically this is the same futures-vs-spot caveat as MAVIA and LAB — short the Binance Futures leg, hold spot KuCoin inventory, mind the funding. Verdict: don't treat this as a single trade, treat BANK itself as a watchlist add — a token disagreeing with itself across four venues in one session usually keeps doing it for a few more.
Binance Futures vs Hyperliquid was the standout pair today — it produced both of the top two spreads (USTC 41.67%, MAVIA 36.50%) and nothing else came close to that magnitude. That's a structural read, not a coincidence: Hyperliquid's user base skews toward directional, degen-leaning longs, which tends to bid perp marks persistently above the more heavily-arbed Binance Futures price on lower-liquidity names. Any ticker that shows up paired against Hyperliquid deserves a second look — and the same liquidity-ceiling caveat every time.
BANK's split across four venues (Bitget, Bitunix, Binance Futures, KuCoin) is the opposite pattern — not one dominant pair repeating, but genuinely fragmented price discovery across the whole token. OKX vs Binance Futures (LAB) is the cleanest tier-1-vs-tier-1 basis spread of the day, and Binance vs Coinbase (LA, 9.20%) is the familiar geography trade — a Coinbase premium or discount driven by that exchange's more US-retail-heavy order flow and heavier fee schedule pricing things slightly differently than Binance's global book. Overall shape of the day: two of the five largest spreads share one venue pair (a real pattern worth tracking), the rest are scattered CEX/CEX prints with no second repeat pair.
Tier-1-vs-tier-1 spreads like LAB's 14.17% (OKX/Binance Futures) close fast — both venues have competent market makers and bots watching the gap, so a spread that wide between two deep books is unusual precisely because it doesn't last. If you saw it, assume it was gone within minutes, not hours.
Spreads involving Hyperliquid or Bitunix behave differently and can persist for hours, because capturing them requires either bridging capital across venues or running a hedged basis position — both of which carry enough friction to keep casual competition out. That same friction caps your size: the illiquidity that lets the spread survive is the same illiquidity that will eat your fill the moment you push real size through it. On a $0.005-handle token like USTC, a headline 41.67% spread can shrink to single digits the instant your own order starts walking the book.
Practical sizing rule: on tier-1/tier-1 gaps (LAB), move fast and stay small — you're racing bots with lower latency than you. On tier-2/perp-DEX gaps (USTC, MAVIA on Hyperliquid; BANK on Bitunix), size to the book, not to your bankroll — start with a probe order in the low hundreds of dollars, and only scale up if your actual fill price stays within a tight band of the quoted price. If your probe order moves the touch price by more than 20-30% of the spread you're chasing, stop — you're now paying for your own liquidity, not capturing someone else's inefficiency.
Walkthrough using LA — the cleanest pure spot-to-spot pair today (Binance $0.059000 buy, Coinbase $0.063800 sell, 9.20% gross spread), assuming $10,000 notional per leg with inventory already positioned on both exchanges:
That 1.76-point drag is the good case — a liquid pair, both legs pure spot, no funding to account for. The futures-vs-perp-DEX spreads (USTC, MAVIA, LAB, the second BANK print) carry an extra variable on top of trading fees: funding rate on whichever leg is a perp position, which can add to or subtract from the basis depending on sign and how long you hold the hedge open. Minimum spread worth chasing, as a rule of thumb: don't touch anything under 2% net-of-fee on major, liquid tickers — the roundtrip fee drag alone on a Binance/Coinbase-style pair can run 0.2% to 1.2%, and that's before slippage. Demand 3%+ headroom on micro-caps like BANK, LAB, AKE, TAG and NIGHT to leave a slippage buffer, since thin order books eat into fills disproportionately as position size grows. And treat every Hyperliquid or cross-futures print as a basis trade requiring a funding-rate check first — a fee-only calculation on those will give you the wrong number.
BANK is the name to keep on the monitor — two simultaneous cross-venue spreads across four exchanges in one session usually means the fragmentation runs another session or two before consolidating. Watch Bitget/Bitunix and Binance-Futures/KuCoin specifically. USTC and MAVIA against Hyperliquid are worth a re-check around the next Asia-session open, when Hyperliquid's directional-degen positioning tends to peak and its perp marks drift furthest from Binance Futures. LAB's OKX/Binance-Futures gap has likely already closed given both are tier-1 books — don't expect a repeat there without a fresh catalyst moving one venue faster than the other.
Best windows to watch: early UTC hours during the Asia-session ramp, when perp-DEX-vs-CEX gaps (Hyperliquid pairs) tend to widen, and the few hours immediately following any listing or delisting announcement, when one venue's spot book re-prices faster than the other.
Numbers don't get excited, and neither do I. 102 spreads today, two of them — USTC and MAVIA, both against Hyperliquid — genuinely worth a basis trade if you can stomach a thin book and manage funding. The rest is the usual scattered fee-math exercise: BANK's four-way fragmentation, LAB's clean tier-1 basis, LA's Coinbase premium. Size to the book, not to the headline percentage, and remember today's volume fields were all blank — go verify depth yourself before anything else. Arbitrage Hunter — July 21, 2026.