The capital rotation that many expected to lift Bitcoin never materialized. Instead, fiat liquidity poured into artificial intelligence, absorbing the bulk of new money that might have entered crypto markets. That assessment came directly from Arthur Hayes, the former BitMEX CEO, who revealed in a recent interview that he has sold all of his altcoin holdings.
In the original report, Hayes confirmed he liquidated positions in HYPE, NEAR, and Worldcoin, among others. The move signals a shift in macro thinking from one of the most closely followed voices in crypto, who previously outlined bullish scenarios tied to global liquidity cycles.
Hayes now sees a structural problem with the AI narrative that threatens to undercut risk assets broadly. He pointed to three issues that could unwind the speculative capital parked in AI. The first is energy pricing. The AI arms race demands enormous power consumption, and rising costs threaten the profitability assumptions behind many AI projects. The second concern is US policy risk—a sudden regulatory reversal or clampdown on AI firms could instantly reprice the sector. The third warning focuses on the liquidity vacuum created by the upcoming IPOs of Anthropic and OpenAI. Those mega-listings, Hayes argues, will absorb enormous amounts of institutional capital, draining money from other high‑beta assets, including crypto.
A recent legislative fight in Washington underscores the unpredictable regulatory environment. Banks are already maneuvering to kill a landmark crypto bill, and the same political forces could easily shift their focus to AI. If lawmakers reverse course on AI development or impose heavy restrictions, the capital backing AI plays could flee—and that reallocation, Hayes warns, would not automatically benefit Bitcoin.
The AI trade has also reached deep into crypto’s own markets. AI-themed NFTs have been top performers recently, with $X@AI BRC-20 NFTs racking up $17.8 million in weekly sales. This kind of speculative overflow shows how AI hype has soaked up attention and capital that might otherwise have flowed into layer-1 tokens, DeFi protocols, or Bitcoin. When that narrative cools, the unwinding of AI‑linked crypto bets could accelerate, dragging down tokens that are only loosely connected to artificial intelligence.
The Bathwater Risk for Bitcoin
Hayes delivered a stark warning about what happens if the AI trade corrects sharply. In his view, Bitcoin will not act as a safe harbor. Instead, it gets “thrown out with the bathwater.” That framing challenges the common thesis that Bitcoin decouples during risk‑off events by attracting flight‑to‑safety flows. In a liquidity squeeze driven by AI disillusionment, institutional money might simply exit all speculative sectors simultaneously, leaving no advantage for the largest digital asset.
This outlook matters because it contradicts the narrative that Bitcoin is still waiting for a macro liquidity trigger. If the wall of money that went into AI is withdrawn not because of a broader recession but because of sector‑specific issues, the effect could be a harsh re‑rating of assets that have been priced against the AI boom. Traders holding altcoins that benefited from the AI narrative—whether through direct exposure or through ecosystem associations—may find that correlation runs both ways.
What the Exit Tells Markets Now
Hayes moving to cash on altcoins deserves attention because his previous macro calls have often aligned with large liquidity shifts. He was early in mapping out the impact of Treasury General Account draws and the RRP facility on crypto prices. His exit from HYPE, NEAR, and Worldcoin is not a routine portfolio rebalance; it is a directional bet that the AI‑driven liquidity cycle is peaking, and that the unwind will damage the broader altcoin complex.
Yet the timing remains uncertain. AI‑related capital flows have been staggering. In the tokenized asset space, a single week saw Bullish buy Equiniti for $4.2 billion and real‑world assets cross $20 billion on‑chain, illustrating how massive institutional liquidity is moving into digital infrastructure. Some of that momentum may continue to feed AI ventures before the turn. For now, Hayes has already walked away, and the risk is that others follow before the peak announces itself.
His move also raises questions about the specific altcoin tokens he dropped. NEAR and Worldcoin both sit at the intersection of AI and blockchain. If even AI‑crypto hybrids are being discarded, it suggests Hayes expects a systemic drain, not a rotation within the altcoin bucket. For traders still long AI‑themed tokens, the signal is clear: the exit door is open, and the liquidity backing those positions may prove thinner than it appears.
What remains unclear is whether a peak in AI enthusiasm would bring new flows back to more established crypto assets, or simply leave the entire ecosystem smaller as institutional money moves to bonds, IPOs, or cash. Hayes clearly believes the latter. His sell‑everything approach points to a market view where the AI trade’s end is not just a sectoral rotation but a broader de‑risking event. The coming months will test whether that judgment proves prescient or premature.
