🎯 Arb Desk Report
Forty-seven arbitrage opportunities crossed the scanner today, and the top of the board was not subtle. VELVET printed an 18.09% spread between Binance Futures and Gate Futures — buy side at $0.186600, sell side at $0.198377 — which is the kind of number that either means a fantastic fill or a market so thin that the second leg never actually clears at the quoted price. Both are plausible with a mid-cap perp like VELVET, and the desk's job today was separating the two.
Below VELVET, the board had real texture. NES showed up twice with opposite directionality — a 10.70% spread buying OKX and selling KuCoin, and later a 9.18% spread running the reverse route, buying KuCoin and selling OKX. That's a classic sign of a temporary price dislocation working itself out rather than a stable, repeatable edge — the market found the imbalance and closed part of it, then briefly overshot the other way. STORJ was the volume leader by event count, showing up four separate times across four different exchange pairs (Bitget-Binance Futures, Binance-OKX Spot twice, and Gate Futures-Bitunix), with spreads ranging from 6.35% to 8.26%. That kind of repeated appearance across multiple venues is usually a sign of genuine, structural pricing friction on a token that doesn't have enough cross-exchange market-maker coverage to keep books tight everywhere at once.
Rounding out the picture: BTR (6.96%, Bybit to Gate Futures), PROM (6.42%, Bitunix to Exchange51), and HEMI (6.20%, Gate Futures to Hyperliquid) all cleared the 6% threshold that separates 'interesting' from 'noise' on most fee schedules. Notably absent from today's feed: pump/dump volume and buy/sell pressure totals, which all read $0.0M — the volume-tracking layer on this feed didn't populate for the session, so everything below is priced off the spread data alone. Treat any sizing recommendation as a framework, not a hard number, until volume confirmation is back online.
🏆 Top 5 Arbitrage Opportunities
- VELVET — 18.09% spread. Buy Binance Futures at $0.186600, sell Gate Futures at $0.198377. This is the standout of the day by a wide margin, nearly double the next-best spread. An 18%+ gap between a top-tier venue like Binance Futures and a mid-tier one like Gate Futures almost never survives contact with real size — it's the market telling you liquidity is stacked unevenly between the two books. No volume figure came through on this feed, which is the single biggest caveat here: without a depth check on the Gate Futures side, this reads as executable only in small clips, likely under $5-10k notional before slippage eats the edge. Our take: real, but shrinking fast — this is a scalp-sized opportunity for whoever was already positioned on both venues, not a set-up-your-accounts-now trade. The window on spreads this size typically measures in single-digit minutes.
- NES — 10.70% spread. Buy OKX at $0.129400, sell KuCoin at $0.143251. NES showing up twice in the top rankings with opposite legs is the tell that this was a genuine two-sided arbitrage event rather than a one-way anomaly. OKX and KuCoin are both reasonably liquid mid-tier venues, which improves the odds this was actually fillable rather than a stale-quote artifact. Risk factors: KuCoin withdrawal queues for lower-cap tokens can run 10-30 minutes during busy periods, and if NES isn't already resting on both exchanges pre-positioned, the cross-chain/cross-exchange transfer time alone can erase a 10% edge. Our take: executable for pre-funded accounts, marginal for anyone needing to move capital first.
- NES — 9.18% spread. Buy KuCoin at $0.138810, sell OKX at $0.145200. The mirror-image follow-up to the trade above, and the price levels tell the story clearly: KuCoin's ask here ($0.138810) sits below OKX's earlier bid ($0.143251), meaning the market partially closed the first gap and then briefly overshot in the other direction. This is textbook mean-reversion arbitrage — the kind of pattern that rewards traders running automated inventory-neutral bots across both venues simultaneously, since neither single-leg trader could realistically catch both legs manually. Our take: this pair (OKX/KuCoin on NES) is worth a standing alert — two flips in one session on the same asset suggests a temporary liquidity imbalance that could recur.
- STORJ — 8.26% spread. Buy Bitget at $0.046150, sell Binance Futures at $0.048222. This is the best of STORJ's four appearances today and pairs a smaller venue (Bitget) against a top-tier one (Binance Futures) — the more common and more durable arbitrage shape, since Binance Futures depth is rarely the constraint. Risk here sits almost entirely on the Bitget leg: withdrawal processing and spot liquidity on mid-tier venues for a token like STORJ can be inconsistent, and Bitget occasionally throttles withdrawals during high-volatility windows. Our take: probably the most executable trade in the top five given the Binance Futures counterparty on one side — worth checking Bitget's live order book depth before sizing up.
- BTR — 6.96% spread. Buy Bybit at $0.031470, sell Gate Futures at $0.033280. Bybit-to-Gate Futures is a less common pairing in today's data, and at sub-7% this is squarely in the range where fees start to matter more than direction. It clears the bar for a fee-adjusted trade (see the profit math below) but with less margin for error than the spreads above it. Our take: fine for traders already holding balances on both venues, not worth a fresh transfer-and-wait cycle.
📊 Exchange Spread Patterns
The clearest pattern in today's data is Binance Futures acting as the high side of the trade repeatedly — it was the sell leg for both VELVET (vs. Gate Futures) and STORJ (vs. Bitget), and appeared again as the buy leg for STORJ against OKX Spot twice. That's consistent with Binance Futures carrying deeper open interest and tighter funding-driven pricing than the mid-tier venues it's paired against, which pushes its price to lead rather than lag during volatility. OKX and KuCoin formed a genuine two-way arbitrage pair on NES specifically — both directions triggered within the session, which is the strongest signal in the data that this exchange pair has a structural, recurring liquidity gap worth monitoring going forward, not just a one-off.
Gate Futures shows up on three separate trades today (VELVET sell side, BTR sell side, HEMI buy side), always on the side you'd expect from a venue with thinner books than its counterparty — sometimes the high side (VELVET, BTR), sometimes the low side against an even deeper book like Hyperliquid (HEMI). That inconsistency suggests Gate Futures pricing lag isn't directional, it's just slower to update across the board, which is a useful thing to know when deciding which side of a Gate Futures spread to trust. Hyperliquid's single appearance, on the sell side against Gate Futures for HEMI, is worth flagging on its own — Hyperliquid's on-chain perp pricing mechanics differ from centralized order books, so spreads involving it can close or widen on different timing than a pure CEX-to-CEX arb. Bitunix and Exchange51 rounded out the field on PROM, a pairing with less trading history in this dataset than the others, which argues for smaller size until a pattern establishes.
⚡ Speed vs Size Analysis
The VELVET spread (18.09%) and the two NES spreads (10.70% / 9.18%) sit in a different risk category than the STORJ, BTR, PROM, and HEMI spreads clustered in the 6-8% range, and the right playbook differs for each. Wide, rare spreads like VELVET's are almost always thin-liquidity events — the spread is large precisely because there isn't enough size on one side to close it quickly. Chasing these with anything beyond a probe-sized order (a few thousand dollars notional) risks moving the market against yourself mid-execution, turning an 18% paper edge into a fraction of that after your own trade eats through the book. These are speed trades: get in with small size, fast, or don't bother.
The narrower, more frequent spreads — STORJ's four appearances and the general 6-8% cluster — are the ones where size can actually scale, because a token showing repeated cross-venue dislocations usually has enough standing liquidity on both sides to support it. The tradeoff is time: without volume data confirming depth (today's feed came back empty on that front), the responsible move is to start with a small clip, confirm the fill quality on both legs, and scale up only after seeing your own execution doesn't move the quoted spread materially. As a rule of thumb for this kind of thin-data session: size positions so that your own order is under 5-10% of what you'd estimate as available book depth, and never assume a quoted spread survives a market order of more than a few thousand dollars on a mid-cap token without checking the book first.
💰 Profit Calculations
Walking through the STORJ 8.26% trade as the clearest example: buy on Bitget at $0.046150, sell on Binance Futures at $0.048222, gross spread 8.26%. Taker fees on both legs typically run 0.05-0.10% on Binance Futures and 0.06-0.10% on Bitget spot — call it roughly 0.20% combined round-trip for a conservative estimate. Withdrawal fees for moving STORJ off-chain between exchanges commonly run in the $1-3 flat-fee range depending on network congestion, which on a $1,000 notional trade is roughly 0.1-0.3% of position size — larger as a percentage on smaller trades, smaller on larger ones. Net that out and the STORJ trade still clears roughly 7.8-7.9% after costs on reasonably sized positions, which is a genuinely strong edge if the depth is there to support it.
Applying the same framework to BTR's 6.96% spread (Bybit to Gate Futures): after ~0.20% round-trip fees and a similar withdrawal-fee drag, net profit lands closer to 6.5-6.7% depending on position size — still solidly worth taking, but with noticeably less margin for slippage than the STORJ example. The general rule for this fee environment: spreads below roughly 1.5-2% aren't worth chasing at all once fees and withdrawal costs are accounted for, since a bad fill or a fee-tier miss can flip the trade negative. Everything in today's board cleared 6% or higher, meaning every single top-line opportunity had real theoretical margin — the open question on each one is execution quality, not math.
⚠️ Risk Alerts
Volume data was completely absent from today's feed (pump volume, dump volume, buy pressure, and sell pressure all read $0.0M), which means none of the spreads above come with a depth confirmation — treat every listed opportunity as a price signal only, not a guaranteed-fillable trade, until you've checked the live order book yourself. Withdrawal timing is the other major risk vector across the board: mid-tier venues like Gate Futures, Bitget, KuCoin, and Bitunix can all see withdrawal processing slow to 10-30+ minutes during high-volatility windows, which is more than enough time for a 6-9% spread to fully close before your second leg lands. Any trade that requires moving fresh capital between exchanges rather than using pre-positioned balances should be sized down accordingly or skipped entirely on the tighter spreads.
PROM's spread (Bitunix to Exchange51) deserves specific caution — Exchange51 has less trading history in this feed than the other venues referenced today, and unfamiliar or lower-tier exchanges carry elevated counterparty and withdrawal risk that a percentage spread alone doesn't capture. Hyperliquid's appearance in the HEMI trade is worth a separate note too: as an on-chain perp venue, its pricing and settlement mechanics don't map directly onto CEX withdrawal/deposit timing, so cross-venue arbs involving it should be modeled with on-chain confirmation times in mind, not standard CEX transfer assumptions.
🔮 Tomorrow's Setup
STORJ is the name to keep on a standing alert — four separate spread events across four different exchange pairs in a single session is unusual and suggests the token's cross-venue liquidity coverage is genuinely thin right now, not a one-day fluke. The OKX/KuCoin pair on NES also earned a watch flag after flipping direction twice today; pairs that oscillate like that in one session sometimes keep doing it for a few days before liquidity providers close the gap for good. Binance Futures is worth watching as a recurring high-liquidity anchor — any spread that has Binance Futures on one leg has historically been the more executable half of today's board, so new spreads involving it deserve priority attention over spreads between two thinner venues.
On timing: today's cluster of high-value spreads (VELVET, both NES prints) suggests dislocations are still concentrated around normal high-volatility windows — US market open and the Asia-Europe overlap tend to be when order books thin out enough for gaps like these to appear. Keep Gate Futures and Bitunix on the monitor list generally, since both showed up multiple times today as the slower-moving side of a spread. If volume/pressure data comes back online tomorrow, prioritize re-checking any spread above 8% against real depth before treating it as tradeable — today's board had the price signals right but the liquidity confirmation missing.
Sign Off
Forty-seven signals, one 18% headline, and a STORJ book that can't seem to keep its price straight across four venues — busy day on the desk. Watch the depth, respect the withdrawal clocks, and don't let a good spread talk you into a bad fill.
Arbitrage Hunter — August 25, 2026
◈ tags
#analysis#crypto#market#arbitrage#spreads#trading