Three Ways to Arbitrage Zcash (One Isn't Real)

Zcash had a loud year. Through 2025 and into 2026, ZEC ran hard, OKX relisted it, Binance kept it and added perpetual futures on top, Robinhood listed it, and it was widely reported to have briefly passed Monero as the largest privacy coin by market cap. When an asset suddenly trades on a dozen venues, across three of its own internal value pools, and through a few cross-chain bridges, the same question always shows up: where's the arbitrage?

People point at three different layers and call each one an arbitrage. The first — between Zcash's own shielded pools — isn't a trade at all, and it's worth knowing exactly why. The second, between exchanges, is real but thin. The third, across chains, is the widest and the most interesting. Let's walk all three, because the same property keeps shaping each one: ZEC is a privacy coin, and privacy is friction.

Angle 1 — The trade that isn't: shielded pools & the turnstile

Every ZEC in existence sits in one of Zcash's value pools — consensus-tracked buckets defined by the kind of address holding the coin. There's a transparent pool (public, Bitcoin-style t-addresses) and the shielded pools, which hide amounts and parties behind zero-knowledge proofs (a zk-SNARK — a proof that a transaction is valid without revealing what's in it). The shielded pools arrived in layers:

Value moves between pools by shielding (transparent → shielded) and deshielding (shielded → transparent), plus wallet-driven migrations like Sprout → Sapling → Orchard as the old pools get deprecated. Here's the rule that governs all of it — the turnstile: each shielded pool publicly tracks a running net total of value in minus value out, and consensus forbids that total from ever going negative. More can never leave a pool than entered it. The only way to cross a pool boundary is to transparently reveal the amount you're moving.

The turnstile exists for one reason: counterfeiting detection. Because shielded amounts are hidden, a bug that let someone forge ZEC inside a pool would be invisible by definition. But forged coins still have to exit eventually — and when they did, more would leave the pool than ever entered, turning the pool's public total negative. That negative is the alarm bell. This isn't hypothetical: a real counterfeiting bug once lived in Sprout's original proving setup, was caught, and was quietly fixed at the Sapling upgrade — and turnstile accounting was the tool used afterward to argue no one had actually exploited it.

So here's the punchline for anyone hunting an arb: there is no price spread between pools. A ZEC in Sapling and a ZEC in Orchard are the same coin, worth the same thing, redeemable 1:1. "Inter-pool arbitrage" is a misnomer — there's no gap to close. What actually can be played here is not price but metadata: the timing of a shield/deshield changes the size of your anonymity set (shield when the pool is busy and you hide in a bigger crowd; deshield and you publicly reveal the amount, leaking information), and only the transparent side has a public mempool where ordering games are even possible. The thing people picture as a trade is really a privacy decision.

The one genuinely observable dynamic is migration. By late 2025, about 29% of all ZEC sat shielded (per Coin Metrics), and the bulk of it had drained into Orchard as the newest pool — Sapling holds a thin slice, and Sprout is all but empty. That shift is real and trackable. It just isn't a trade.

Angle 2 — The real-but-thin trade: cross-venue (CEX/DEX)

The second arbitrage is the classic one: the same ZEC quoted at different prices on different venues, at the same moment. This one is real. It's just bounded hard by the fact that ZEC is a thin, fragmented, privacy-coin market. First, where it even trades in 2026:

That last pair of rows matters more than it looks. Because ZEC lives on its own UTXO chain (a Bitcoin-style ledger of unspent outputs, not Ethereum-style accounts), it has no native automated market maker — no Uniswap pool you can trade against on-chain. On-chain swap liquidity exists only through cross-chain protocols like THORChain and Maya. So cross-venue arbitrage for ZEC is really CEX-to-CEX, or CEX-to-native-swap — and every leg of it runs into the same three frictions.

• You usually have to deshield to deposit. Most exchanges accept only transparent (t-address) deposits, so a holder sitting in a shielded pool has to deshield first — which publicly reveals the amount and adds an extra on-chain hop before the funds are even movable.
• Finality is slow. A new Zcash block lands every 75 seconds, and exchanges want several of them to confirm before crediting a deposit. That stretches the round-trip into many minutes — minutes during which the spread you spotted can simply evaporate.
• The books are thin. Privacy-coin order books are shallow and scattered across venues, so spreads are naturally wider — which is whatcreates the nominal arbitrage — but depth is shallow, which is whatcaps how much you can actually move before your own trade closes the gap.

Compare that to the CEX-DEX arbitrage on Solana, where the whole game is landing a transaction inside a 400-millisecond slot. ZEC arbitrage isn't a speed game at all. The edge isn't being faster than the next bot; it's being willing to tie up capital across a multi-minute, deshield-and-reconfirm round-trip that most fast money won't touch.

Angle 3 — The widest trade: cross-chain & bridges

The third arbitrage is between ZEC on its own chain and ZEC's representations on other chains. This is the widest playing field and the one with the most graveyards. Start with the graveyards, because they teach the core lesson: wrapping a shielded UTXO coin is genuinely hard, and the custodial shortcuts tend to die.

renZEC is the cautionary tale. RenVM let you mint a wrapped ZEC on Ethereum and elsewhere — until RenVM disabled minting and wound down in late 2022, collateral damage from the FTX/Alameda collapse (Alameda had acquired Ren and ran its custody infrastructure). The token contract still exists on-chain, but it's dead and illiquid. That wasn't a price depeg so much as a stranding: the same event flattened RenVM's total value locked from over $1 billion to a few tens of millions, and renZEC went with it. The custodial pegs that remain — Binance-Peg ZEC on BNB Chain, the old Tokensoft WZEC on Ethereum — technically exist but carry thin-to-nil liquidity and full counterparty risk.

The model that actually works now is the opposite of wrapping: native cross-chain swaps, via THORChain (native ZEC went live recently, in phases) and its fork Maya (live a while longer). These don't mint a wrapped token at all. ZEC sits in a continuous liquidity pool — a constant-product AMM paired against the network's own settlement asset (RUNE on THORChain, CACAO on Maya) — and you swap into and out of it with a slip-based fee that scales with trade size. There's no price oracle. The pool only tracks the real market because arbitrage is the peg mechanism: when the pool's ZEC price drifts from the outside market, profit-seekers buy the cheap side and sell the rich one until it snaps back. NEAR Intents adds a third route, using a solver network to fill native ETH/BTC/SOL → native ZEC swaps across many chains (the Zashi wallet wired it in for private swaps).

Which makes the mechanics of the cross-chain trade clear. The arbitrageur is the one keeping THORChain's ZEC pool honest, and the friction bounding how tightly they can hold the peg is bridge latency. Settling a ZEC leg means waiting on Zcash confirmations — several blocks, each 75 seconds apart — plus the protocol's own settlement, so the window where the price can move against you is measured in minutes, and your capital is locked the whole time. There's a privacy cost layered on top: a Maya contributor has openly noted that cross-chain shielded ZEC swaps still leak metadata once value leaves the Zcash chain. So even the "native" route trades some of the privacy that was the whole point.

The map

Three layers, three different answers:

• Shielded pools (Angle 1): not a trade. All ZEC, 1:1, no spread. The only thing you're really choosing is how much privacy to keep.
• Cross-venue (Angle 2): a real trade, but a slow one — gated by deshield-to-deposit, 75-second finality, and thin books. Capital patience, not speed.
• Cross-chain (Angle 3): the widest and most active, where arbitrage literally is the peg for native-swap pools — bounded by bridge latency and a residual metadata leak.

Notice the single thread running through all of it. On a transparent, fast chain, arbitrage is a millisecond race. On Zcash, the defining property is privacy, and privacy shows up as friction at every layer — the turnstile that makes pools auditable but un-arbitrageable, the deshielding step that gates every deposit, the confirmation delay that widens every window, the metadata that leaks the moment value crosses a bridge. The arbitrage in ZEC is real in two of the three places people look for it. It's just slow, shallow, and patient by design — which is the most Zcash thing about it.