Arthur Hayes Pushes Back on “Exit Liquidity” Accusations, Cites “AI Jitters” Macro Thesis Behind HYPE, NEAR, and WLD Sales

Key Highlights

• On-chain investigator ZachXBT publicly questioned whether Arthur Hayes used his followers as exit liquidity after the BitMEX co-founder exited positions in NEAR, HYPE, ZEC, and WLD within roughly two weeks of publicly endorsing them.
• In an interview with Cointelegraph, Hayes maintained he reports his own trades rather than offering financial advice, and challenged critics to find any post where he told followers to buy or sell.
• Hayes attributes the sales to a macro thesis he calls "AI jitters" — the view that roughly $1.5 trillion in AI-related debt issued since 2022 has absorbed liquidity that would otherwise have flowed into crypto.
• He identified three specific risks driving the exits: rising energy costs tied to the Strait of Hormuz conflict, the possibility of Trump turning rhetorically against AI companies ahead of November, and a looming supply wall from SpaceX, Anthropic, and OpenAI listings expected around September.
• Hayes treated the Zcash (ZEC) sale separately, framing it as a trust-based exit following the discovery and patching of a bug in the Orchard privacy pool.
• Despite selling HYPE, Hayes maintains it remains one of the best products in crypto and says the sale reflected timing rather than a change in conviction, citing valuation multiples around 20x versus 60x for Coinbase and 45x for Robinhood.

Arthur Hayes, co-founder of BitMEX and Chief Investment Officer of Maelstrom, addressed accusations that he used his large public following as exit liquidity, in an interview with Cointelegraph's Giovanni Pigni. The exchange centred on on-chain investigator ZachXBT's claim that Hayes had publicly promoted NEAR, HYPE, ZEC, and WLD before exiting all four positions within a roughly two-week window.

The Accusation

The dispute originated from a post by ZachXBT on X, in which the investigator questioned the timing of Hayes's exits, listing the four tokens in the sequence he had sold them and asking how much exit liquidity had been generated from Hayes's followers in the preceding days. The sequence drew scrutiny because each sale followed bullish public commentary from Hayes that had brought retail attention to the relevant assets.

Hayes's Response: Reporting, Not Advising

Hayes drew a firm distinction between disclosing his own trades and providing financial advice, stating he manages money only for himself, has never claimed to be a financial advisor, and challenged critics to identify any instance where he instructed followers to buy or sell anything. He framed the disclosure of his trades as informational — what people choose to do with that information, he said, is their own decision.

On why he posts his trading activity publicly at all, Hayes was candid that his underlying incentive is to draw attention to his market thesis, regardless of whether his audience agrees with the position.

The "AI Jitters" Macro Thesis

Hayes framed the sales as the product of a broader macro shift rather than any change of view on the individual tokens themselves. He described developing what he called "AI jitters" — a conclusion that the liquidity which would normally flow into crypto markets has instead been almost entirely absorbed by the AI investment boom. He traced this to the scale of AI-related debt issuance, estimating roughly $1.5 trillion issued between 2022 and 2026, with the large majority concentrated from 2025 onward, and argued that this capital effectively never reached Bitcoin — which he sees as the underlying reason for crypto's underperformance since October 2025.

Three specific risks underpinned the decision to exit. The first concerns energy prices: the longer the US-Iran conflict restricts flows through the Strait of Hormuz, the more commodity inventories will need rebuilding once a resolution is reached, structurally pushing oil and gas prices higher. Higher energy costs raise the cost of running AI infrastructure, which Hayes argues could slow AI adoption and undermine the earnings narrative currently supporting AI-related valuations.

The second concerns US politics. Hayes suggested that President Trump may find it politically useful to turn rhetorically against AI companies ahead of the November elections, tapping into voter anxieties around job displacement and the impact of data centre construction. He characterised this as likely rhetorical posturing rather than substantive policy, but noted that markets tend to react to political statements at face value regardless of underlying intent.

The third concerns an anticipated supply wall. Hayes expects SpaceX, Anthropic, and OpenAI to list at valuations exceeding a trillion dollars around September, creating an absorption challenge as existing lockups expire around the same time as these new listings. He was particularly blunt in describing the structure of the SpaceX offering, characterising it as a low-float, high fully-diluted-valuation setup more reminiscent of a speculative token than a traditional equity listing. His overall framing was that avoiding being forced to sell into a falling market outweighs the opportunity cost of selling early, even if prices subsequently rise further.

Zcash: A Different Kind of Exit

Hayes treated the ZEC sale as distinct from the macro-driven exits, describing it instead as a response to a trust violation. Following the discovery and subsequent patching of a bug in Zcash's Orchard privacy pool, Hayes said he was unwilling to maintain the position without formal verification that no unauthorised token minting had occurred as a result of the bug. While he credited the researcher who identified and patched the issue with likely having handled it correctly, he held to a standard of requiring certainty before holding a position of that size, while leaving open the possibility of re-entering at higher prices if the matter is formally resolved.

Still Bullish on Hyperliquid

Despite exiting his HYPE position, Hayes was emphatic that the sale reflected timing rather than any deterioration in his view of the underlying project. He described Hyperliquid as one of the strongest products in crypto, citing its product-market fit and its mechanism of returning revenue to token holders.

On valuation, he noted that Maelstrom's original entry was around 9x earnings based on circulating supply, with current multiples sitting closer to 20x. He compared this favourably to Coinbase at roughly 60x earnings, Robinhood at roughly 45x, and CME at roughly 25x, concluding that on a pure multiples basis, HYPE remains undervalued relative to comparable financial infrastructure businesses.

The Post-AI-Bubble Case for Bitcoin

Looking beyond his immediate caution, Hayes outlined two scenarios for if the AI investment cycle ultimately corrects sharply. In the first, emergency monetary easing in response to the correction benefits Bitcoin as a new narrative cycle begins. In the second, a financial crisis emerges as over-leveraged second- and third-tier AI companies struggle to service their debt. In either scenario, Hayes expects Bitcoin to outperform AI-related equities once the dust settles — though not necessarily before that point. He framed his outlook as one where, following a significant AI correction, investor attention rotates elsewhere, and Bitcoin benefits from both that rotation and any resulting monetary stimulus.

His framework for evaluating altcoins going forward remains narrow: clear revenue-generating mechanisms, genuine paying customers, and profit returned to token holders through buybacks or token burns — the model he believes Hyperliquid currently exemplifies.

The Unresolved Question

The exchange leaves a genuine tension unaddressed. Hayes's macro reasoning is internally coherent, and he has discussed elements of this thesis publicly before the sales occurred. What drew scrutiny was the compression — four publicly endorsed positions, all exited within roughly two weeks of those endorsements.

His defence — that he reports his trades rather than advising on them — accurately reflects how he frames his public communications. It does not, however, fully address the underlying asymmetry ZachXBT pointed to: the reach a large, influential account generates on the way into a position compared with the speed at which that same account can exit. Readers are left to weigh both elements — a coherent macro case for reducing risk exposure, and a timing pattern that understandably invites the question raised.