Sixty-nine arbitrage windows opened across CEX venues today, headlined by a 22.49% Binance-to-Coinbase gap on CHZ. The tape shows a repeatable pattern: Binance and OKX pricing cheap, Coinbase, KuCoin, Bitget and Bitunix pricing at a premium, concentrated in thin small-cap names where the real edge shrinks fast once fees and slippage are applied.
The scanner logged 69 arbitrage events over the last cycle, the busiest print of the week for cross-exchange dislocations. The headline trade was CHZ, which opened a 22.49% gap between Binance spot at $0.014940 and Coinbase at $0.018300 — a spread wide enough to clear round-trip fees, a network withdrawal, and still bank low-20s percent net, assuming size could actually get filled on both legs. Behind it, HANA printed 17.34% between Binance Futures and Bitunix, and a cluster of ESPORTS spreads (13.42%, 11.82%, 11.43%, 10.65%) turned that single ticker into the most-arbitraged name of the session across four different venue pairs.
Notably, the pump/dump and buy/sell pressure totals for the cycle all read $0.0M — this was not a momentum session. No coordinated volume surge drove these gaps; they read as structural mispricing between thin small-cap books (ESPORTS, AVAAI, ARIA, HANA) and the deeper majors basis that shows up in DOT's 10.67% OKX-to-Coinbase spread. That distinction matters for execution: momentum-driven spreads close fast as the crowd arbitrages them away, while thin-book mispricing can sit open for longer simply because too few desks are watching a ticker like ESPORTS across four venues at once — but it also means size, not speed, is the binding constraint.
CHZ was the standout print: buy-side liquidity on Binance spot at $0.014940 against a Coinbase ask of $0.018300, a $0.003360 absolute gap. This is a large enough spread to survive the full cost stack — Binance's 0.10% taker fee, Coinbase's heavier 0.60% taker fee, and a flat network withdrawal — and still clear roughly 21.6% net on a $10,000 clip (see Profit Calculations below). The catch is depth: CHZ trades with respectable but not deep books on Coinbase, so size beyond the first few levels of the ask stack would walk the fill price up and compress the realized spread. Withdrawal timing is the other variable — moving CHZ off Binance typically clears in minutes on a normal network, but any AML/risk hold on a large outbound transfer would burn the entire edge before funds land on Coinbase. Verdict: executable in the $5,000-$15,000 range for a desk with pre-funded balances on both venues; risky as a cold-start trade requiring a live withdrawal.
HANA's spread ran between Binance Futures at $0.037055 and Bitunix spot at $0.040250. This is a futures-to-spot cross rather than a clean spot-to-spot arb, meaning the trade carries basis risk on top of execution risk — closing the futures leg means either settlement mechanics or an offsetting spot purchase on Binance to flatten exposure before moving inventory to Bitunix. Bitunix itself is the bigger flag here: a smaller-tier venue with materially thinner order books and a track record of slower, more manual withdrawal processing than Tier-1 exchanges. A spread this wide against a Tier-2 destination venue is as much a liquidity-risk premium as it is free money. Verdict: only cleanly executable for a desk that already holds pre-positioned HANA inventory on Bitunix — chasing it via a fresh transfer is a coin flip on whether the window survives the withdrawal wait.
The first of four ESPORTS prints this cycle: Bitunix bid at $0.025647 against a KuCoin ask of $0.026610. ESPORTS showing up four separate times across four different venue pairs (Bitunix, KuCoin twice, Bitget, Binance Futures) in a single scan is itself the story — it points to a ticker with genuinely fragmented liquidity rather than one clean two-sided market, which is exactly the setup that keeps spreads open longer but also means every leg is thin. KuCoin's book on a low-cap name like this is workable for a few thousand dollars of size before slippage starts eating the edge. Verdict: executable in small size (under $3,000) as a quick in-and-out; not a trade to scale without checking live depth on both venues first.
Same ticker, different cross: Binance Futures at $0.021776 against KuCoin spot at $0.024350. Like the HANA trade above, this is a futures-to-spot structure, so the realized edge depends on how cleanly the futures leg can be closed or hedged. The redeeming feature is that Binance Futures is deep and liquid even on a small-cap perpetual, so buy-side fill risk is low — the constraint sits entirely on the KuCoin sell side, and on getting inventory across the two venues fast enough that the spread doesn't collapse first. Verdict: executable for a desk running a standing basis book between Binance Futures and KuCoin spot; not a cold trade for a single-shot retail arbitrageur.
AVAAI printed $0.009602 on Binance Futures against $0.010723 on Bitunix, its first of two appearances this cycle (the second, an 11.40% Bitunix-to-Gate-Futures cross, shows the same name mispriced against a third venue). Sub-cent tokens like AVAAI are notorious for wide effective spreads even outside of arbitrage windows, since tick-size and quoting granularity alone can produce apparent gaps that are mostly noise. Combined with Bitunix's thinner withdrawal and order-book profile, this is the lowest-confidence trade of the top five — real, but small, and worth confirming live depth before committing size. Verdict: sub-$2,000 size only, and treat the fill price as an estimate, not a guarantee, until the order actually clears the book.
A clear structural pattern runs through today's tape: Binance and Binance Futures priced the cheap side of every spread they appeared in (CHZ, HANA, AVAAI, and the second ESPORTS cross), while Coinbase, KuCoin, Bitget, Bitunix and Gate Futures consistently sat on the expensive side. That is the inverse of the classic Kimchi-premium shape, but the same mechanism — Binance functions as the primary global price-discovery venue for these tickers, and smaller or geographically distinct venues lag its book, letting a premium build until someone arbitrages it closed.
DOT's 10.67% spread (OKX Spot $0.804200 → Coinbase $0.890000) is the cleanest illustration of a well-known basis: Coinbase, as the dominant US-regulated venue, regularly trades at a premium to offshore spot venues like OKX during periods when US retail demand outpaces Asia-session liquidity. Unlike the small-cap names above, DOT is a top-20 asset with real depth on both sides, which makes this specific cross the most mechanically sound trade of the entire list even though it posted the lowest spread of the top five — a reminder that spread size and trade quality are not the same axis.
ESPORTS is the outlier worth flagging on its own: four appearances against four different counter-venues (Bitunix, KuCoin twice, Bitget) in one scan window means its price is not converging across exchanges the way a liquid asset's should. That is a liquidity-fragmentation signal, not a repeatable statistical edge — treat every ESPORTS print as independently thin rather than as one large tradable basis.
The top five splits cleanly into two trade types. CHZ and DOT are large and liquid enough to run real size — a $10,000-$25,000 clip is plausible without moving the book more than a few basis points — but both require moving inventory across venues, so profit is capped by how fast that transfer clears relative to how fast the spread decays. HANA, ESPORTS and AVAAI are the opposite: fast to notice, fast to fill in small size, but the book empties out after the first few thousand dollars, so scaling them means eating progressively worse fill prices rather than capturing the headline percentage.
Slippage on the small-cap leg is the real tax here, not the fee schedule. On a token like AVAAI or HANA trading in the $0.009-$0.040 range, the visible top-of-book size on a Tier-2 venue like Bitunix or Gate Futures is often a few hundred to low-thousands of dollars deep before the next level adds another 1-3% to the effective price — enough to eat half of an 11% headline spread on a $5,000 order alone. Position sizing rule of thumb for this tape: cap small-cap legs at roughly 20-30% of the visible top-3-levels depth on the thinner side of the trade, and reserve full clip sizes ($10k+) for the DOT/CHZ-tier names where Tier-1 venues sit on both legs.
Speed matters most on the withdrawal window, not the entry. A desk with pre-funded balances sitting on both exchanges can capture a spread in seconds and is largely indifferent to whether it decays in one minute or ten. A desk that has to withdraw from the buy venue before selling is racing the spread's natural decay against a transfer that, for tokens like HANA and AVAAI on Bitunix, is not guaranteed to be fast. That single structural difference is worth more to expected profitability than any amount of latency optimization on the order-entry side.
Worked example on the CHZ trade at a $10,000 clip: buying on Binance at $0.014940 with a 0.10% taker fee nets $9,990 of effective capital, or 668,675 CHZ. A flat ~5 CHZ network withdrawal fee trims that to 668,670 CHZ landing on Coinbase. Selling at $0.018300 grosses $12,236.67, and Coinbase's 0.60% taker fee brings that down to $12,163.25. Net profit: $2,163.25, or 21.63% — versus the 22.49% gross headline, meaning the full cost stack (two taker fees plus withdrawal) cost about 0.86 percentage points of edge.
Same math on the DOT trade at $10,000: OKX Spot buy at $0.804200 with a 0.10% taker fee gives 12,422 DOT of effective size. A ~0.1 DOT flat withdrawal trims that negligibly. Selling at Coinbase's $0.890000 grosses $11,055.52, and the 0.60% Coinbase taker fee brings net proceeds to $10,989.19 — a $989.19 profit, or 9.89% net versus 10.67% gross. DOT's lower headline number is actually the more reliable trade of the two: deeper books on both legs mean the realized fill is far more likely to match the quoted price than on a thin name like CHZ.
The single biggest risk across today's list is withdrawal timing on the small-cap names. Four of the top five trades (HANA, both ESPORTS crosses, AVAAI) route through Bitunix, Gate Futures or KuCoin on at least one leg — venues that, relative to Binance or Coinbase, have historically shown slower manual review on larger withdrawals and thinner published proof-of-reserves. A spread that looks like free money on paper can evaporate entirely if a withdrawal sits in review for 20-30 minutes while the price on the destination venue reverts.
Watch ESPORTS first — a ticker that fragmented across four venues today (Bitunix, KuCoin twice, Bitget) is a strong candidate to reopen a gap tomorrow simply because its price discovery is not converging. The Binance-Futures-to-KuCoin and Bitunix-to-Bitget crosses are the two pairs to have on a live monitor.
On the majors side, keep an eye on the OKX-Coinbase basis that drove today's DOT spread — that dislocation tends to widen again around US-session opens when Coinbase's retail flow outpaces overnight Asia liquidity on OKX, and it is the most mechanically reliable setup on this list given both venues' depth. The CHZ Binance-Coinbase gap is worth a standing alert as well; a 22% print is unusual enough that it may either fully close overnight or reopen if the underlying driver (likely a listing or liquidity event on one venue) hasn't resolved.
Best windows to watch: the Asia-to-US session handoff (roughly 00:00-04:00 UTC) for the Binance/OKX-cheap, Coinbase-expensive pattern, and late UTC evening for the Tier-2 venues (Bitunix, Gate Futures) where thinner order-management teams mean mispricings linger longer before getting arbitraged closed.
Sixty-nine windows, one clean 22% headline, and a tape that rewarded desks with pre-funded balances over desks racing a withdrawal clock. Size the small-caps carefully, trust the DOT-style majors crosses more than the sub-cent ones, and keep ESPORTS on a dedicated watch — four venues mispricing one ticker in a single scan doesn't happen by accident. Arbitrage Hunter — July 20, 2026.