Woofun AI reports that Arthur Hayes, appearing on the Bankless podcast, attributes the current muted performance of the crypto cycle to a massive diversion of capital into artificial intelligence infrastructure. Hayes argues that marginal fiat dollars, which historically flowed into Bitcoin and Ethereum, are now being absorbed by AI investments offering superior relative returns. He highlighted that while Bitcoin's ascent to $125K was explosive, the move felt subdued when compared to AI assets like SK Hynix, which delivered 10x returns. Hayes characterizes this dynamic as a classic opportunity cost issue where investors relentlessly chase the 'fastest horse,' which is currently the AI sector.

Institutional data substantiates this suction effect with alarming precision. The Bank for International Settlements (BIS) reported in January 2026 that AI investment is fundamentally reshaping corporate balance sheets on an unprecedented scale. Firms are aggressively utilizing debt and private credit to fund these buildouts, effectively diverting marginal capital away from other asset classes. J.P. Morgan Asset Management estimated that 2026 capital expenditure for the five largest US hyperscalers reached roughly $697 billion, representing an increase of $173 billion since the start of the year.

Furthermore, AI capex for these specific firms surged from 33% of operating cash flow in 2023 to an estimated 93% in 2026. J.P. Morgan noted that this extreme concentration carries significant risk, as the market demands concrete evidence of real AI demand to justify such spending, potentially leading to severe volatility if revenue fails to support expenditure levels.

Woofun AI data shows that Hayes posits a specific mechanism for market recovery: if AI spending continues to outrun revenue, a correction could release trapped capital, rotating it back into hard assets and crypto. He views current BTC and ETH prices as strategic entry points, specifically noting that Ether is trading 30% below its 200-week moving average. His proposed strategy involves riding the AI bubble into cash positions and executing a buy order at the bottom once capital begins to rotate out of the sector. This approach relies on the premise that the current valuation gap is temporary and driven purely by liquidity constraints rather than fundamental flaws in the crypto thesis.

Michael Saylor, referenced on The Pomp Podcast, echoed this sentiment by framing Bitcoin's weakness as a temporary capital queue rather than a structural failure. He cited $500 billion in capital being raised by companies like SpaceX, Anthropic, and OpenAI specifically for AI data center construction, creating a powerful suction effect on global liquidity. Saylor emphasized that only 1-2% of this massive capital inflow originates from Bitcoin, underscoring the dominance of fiat and equity funding in the current AI boom. He predicts a 12-to-24 week cycle where hot money will rotate back into Bitcoin as AI deals mature and lockup periods expire. Saylor estimates that this suction activity will ease by the end of 2026, arguing that falling Bitcoin prices actually make the asset more attractive for institutional re-entry.

Both arguments converge on the conclusion that crypto assets are currently undervalued because capital is trapped elsewhere, with a rotation expected once the AI cycle peaks or corrects.

However, this thesis hinges entirely on the assumption that AI will either crash or that capital will naturally rotate out due to diminishing returns. If AI continues delivering robust returns, the anticipated catalyst may never arrive, and crypto prices could reflect a continued drain of liquidity for an extended period. The market remains in a state of suspended animation, waiting for the resolution of the AI capex cycle to determine the next major directional move for digital assets.