Zcash: Claude Opus Found in Hours What Auditors Missed for Four Years
A four-year-old bug in Zcash's privacy pool. Found by an AI the day after it launched. Patched in five days. Disclosed on day seven. By then the insiders knew, the exchanges had frozen withdrawals, and retail was rotating in while BTC dumped. The Holy Trinity is dead.

On May 22, 2026, Arthur Hayes posted three ticker symbols to his followers and called them his Holy Trinity. HYPE. ZEC. NEAR. Hyperliquid for trading. NEAR for AI infrastructure. Zcash for privacy. He was calling ZEC to $3,000. Multicoin Capital had been quietly accumulating since February. The privacy coin narrative was running.
On June 5, Hayes posted again. Arthur Hayes is the co-founder of BitMEX, one of crypto's first major derivatives exchanges, and one of the most followed traders in the industry. "Sadly due to the Orchard Pool exploit, I had to dump our entire $ZEC bag. While I think it's extremely unlikely of any minting, it cannot be formally cryptographically proved impossible. The privacy from AI, govt, big tech narrative demands perfection."
ZEC dropped 31% on the news. The Holy Trinity is dead.
Here is what happened in between.
Zcash is a privacy cryptocurrency. Unlike Bitcoin, where every transaction is publicly visible on the blockchain, Zcash lets users send funds through a shielded pool where the sender, receiver, and amount are all hidden. The privacy comes from something called zero-knowledge proofs - a branch of cryptography that lets you prove something is true without revealing the underlying information. You can prove you have the funds to complete a transaction without showing anyone your balance. The math does the verification, not a bank, not a human committee.
In May 2022, Zcash launched Orchard - its newest and most advanced shielded pool. The privacy technology underpinning it was state of the art. The world's top cryptographers had worked on it.
It had a bug.
Not an obvious bug. Not one that showed up in routine testing. A soundness bug - a flaw in the mathematical circuit that validates transactions. Soundness means the system should reject invalid proofs. A soundness failure means it might accept one. In a privacy pool where balances are hidden by design, that is about as bad as it gets. An attacker who found this bug could double-spend funds within Orchard, spend the same ZEC twice, and the network would accept both transactions as valid.
The bug existed from May 2022. Through the 2022 bear market. Through 2023. Through the 2024 halving. Through the 2025 bull run. Through Hayes calling ZEC to $3,000. Multiple audits did not find it.
Anthropic released Claude Opus 4.8 on May 28, 2026.
Taylor Hornby is an independent security researcher. He was conducting a paid protocol audit for Shielded Labs, one of Zcash's development organizations. On May 29, one day after Opus 4.8 launched, Hornby used the model alongside custom tooling to analyze the Orchard circuit. Within hours he had a working exploit. In local testing, it generated counterfeit ZEC inside Orchard without detection. The same tool, Shielded Labs confirmed, could have worked on the mainnet before the fix.
Four years of human audits did not find it. An AI found it in a day.
Hornby disclosed the vulnerability to the Zcash Open Development Lab that same evening. Within hours, ZODL engineers Daira-Emma Hopwood, Kris Nuttycombe, and Jack Grigg confirmed the flaw was real. The emergency response began immediately.
What followed was five days of private coordination that most of the market knew nothing about. Miners, exchanges, wallet providers, and node operators were quietly contacted from May 31. On June 2, Zebra 4.5.3 activated a soft fork at mainnet block 3,363,426, temporarily disabling all Orchard transactions while developers prepared the corrected circuit. Exchanges froze ZEC deposits and withdrawals without explaining why. Cake Wallet, which had only added Orchard support in January 2026, suspended ZEC functionality entirely.
On June 3 at 00:05 EDT, the NU6.2 hard fork activated at block 3,364,600. Orchard came back online with the patched circuit. ZODL CEO Josh Swihart called it the most ambitious network upgrade in Zcash's history. It was only the second security-driven protocol upgrade since the network launched in 2016.
ZEC was trading at $605 that day. Still rising.
Then on June 5 they told the world.
Here is a question nobody is formally asking. Between May 29 and June 5, a small group of people knew. Taylor Hornby knew. The ZODL engineers knew. Miners were privately coordinated from May 31. Exchanges were privately coordinated from May 31 - they froze ZEC deposits without explaining why, meaning retail holders could not sell even if they wanted to. During those seven days, ZEC was at $605 and Bitcoin was crashing toward $61,244. Retail money was rotating into ZEC because it looked like the safe trade while BTC dumped. The insiders watched it happen. Nobody is accused of anything here. But someone was on the right side of that information gap and the privacy chain they were protecting makes it difficult to trace who.
The Zcash Foundation confirmed there is no evidence the bug was ever exploited. Zcash's turnstile mechanism, which tracks total ZEC balance across all pools, confirmed the total supply remained intact throughout. The bug could allow double-spending within Orchard but could not create new ZEC from nothing. The turnstile held.
None of that resolves the real problem.
Because Orchard is a privacy pool, proving that nobody exploited the bug during four years of its existence is cryptographically impossible. The transactions are hidden by design. You can confirm the total supply did not change. You cannot confirm that individual transactions within the pool were not manipulated. Shielded Labs said prior use looked unlikely. They also said you should not take only their word for it. They are now proposing a network upgrade that would allow public verification of the ZEC supply going forward.
Udi Wertheimer, a prominent Bitcoin developer, put it more bluntly: "btw this isn't the first time a bug like this was discovered in zcash. last time it was disclosed after being a year+ in the wild and everyone lost faith and zcash went to zero for 7 years, until they found a new generation of buyers who doesn't know the history (that's you)."
He was referring to the Sprout vulnerability. In 2019, Zcash disclosed a similar counterfeiting flaw in its previous shielded pool that had existed since the network launched in 2016. It had been kept secret during remediation. The community lost confidence. ZEC went nowhere for years. Then a new cycle started. New buyers arrived with no memory of the first bug. Hayes called it the Holy Trinity. Multicoin accumulated quietly. ZEC hit $642 in May 2026.
Then the second bug arrived. Same type of flaw. Different pool. Same problem: you cannot prove it was not exploited.
Peter Todd, a veteran Bitcoin developer, posted after the disclosure that this was precisely why Zcash-style zero-knowledge privacy should not be integrated into Bitcoin. The cryptographic surface area is too large. The failure modes are invisible.
Hayes understood the argument. His ZEC thesis was privacy. Privacy coins work when the math is airtight. The math had a hole in it for four years. The probability of exploitation is low. The certainty is not available. For a trade built on cryptographic perfection, low probability is not good enough. He sold everything.
ZEC trades around $377-410 as of June 5, down 31% from before the disclosure, giving back the gains it had posted against a falling Bitcoin market. The bug is patched. The five-day response was competent. Taylor Hornby found a critical flaw that human auditors missed for four years and disclosed it responsibly. The Zcash team shipped an emergency hard fork in under a week.
None of that answers what happened between May 2022 and May 29, 2026.
Nobody can.
The Aftermath
ZEC down 31% as of June 5, 2026. Arthur Hayes sold his entire position. Shielded Labs is proposing a supply verification upgrade to allow public confirmation that no counterfeit ZEC exists within Orchard. The Zcash Foundation confirmed no evidence of exploitation and total supply intact. Peter Todd cited the incident as reason against adding ZK privacy to Bitcoin. This is the second security-driven hard fork in Zcash's history - the first followed the Sprout vulnerability disclosure in 2019.
COMMENTS