BitMEX Founder Hayes: The U.S. May Print Money to Rescue the Market
Arthur Hayes, co-founder of BitMEX, stated that under the "AI Safety First" policy framework, the U.S. government may intervene to support credit risks related to AI infrastructure through one of two paths in the future.
The first path involves the government directly taking on the "buyer" role currently played by institutions like Anthropic, OpenAI, and SpaceX in the computing power market—meaning that when these companies can no longer procure computing resources at their current scale, government departments would step in to fulfill procurement needs, preventing a sudden shrinkage in demand in the computing power market.
The second path is for the Federal Reserve to print money to provide assistance to insurance institutions holding AI-related debts. Hayes pointed out that the trigger condition for this scenario is a decline in computing power demand, leading to defaults on related debts when the original buyers are absent. At that point, insurance companies, as significant holders of AI infrastructure debt, would face pressure on their asset side, necessitating the Fed to expand its balance sheet to cover the losses.
Hayes believes that regardless of which path is ultimately taken, both are essentially the same thing—supporting the construction of AI infrastructure and the resulting credit expansion through monetary issuance, with the only difference being whether the government intervenes directly as a "buyer" or indirectly as a "central bank."
The underlying market logic of this judgment is that capital expenditure for AI infrastructure currently relies mainly on corporate bond issuance and capital expenditures from major cloud computing companies. Once the growth rate of computing power demand slows and buyer power weakens, insurance companies and bond investors holding related debts will be the first to feel the pressure; if the Fed intervenes with money printing assistance, it would imply a rise in expectations for liquidity easing, potentially benefiting risk assets (including crypto assets) due to expectations of monetary expansion, while traditional creditors would face relatively less pressure.
Source: Public Information
ABAB AI Insight
As a co-founder of BitMEX, Arthur Hayes has long analyzed the market through the framework of "macro liquidity driving risk assets." His representative views include multiple predictions post-2020 pandemic that Fed balance sheet expansion would drive up risk assets like Bitcoin, and after the 2023 Silicon Valley Bank incident, he argued that the Fed would indirectly print money to rescue the banking system through tools like BTFP. The subsequent market trends aligned with his predictions, making his "money printing drives asset prices" framework highly regarded in the crypto community.
The core logic of Hayes' current judgment is to compare the construction of AI infrastructure with the "too big to fail" rescue logic after the 2008 financial crisis—currently, the computing power procurement of AI labs like OpenAI and Anthropic is supported by capital expenditures from Nvidia chips and cloud service providers like Microsoft/Oracle, along with corporate bond issuance. If the revenue growth of these labs fails to keep pace with the scale of computing power procurement, the pressure of debt repayment will transmit to insurance companies, pension funds, and other fixed-income investors holding related corporate bonds, while the Fed's response tools typically involve balance sheet expansion or providing liquidity convenience, rather than direct default resolution.
This logic resonates with the precedent of the 2008 takeover of Fannie Mae and Freddie Mac by the U.S. Treasury, as well as the unprecedented purchase of corporate bond ETFs by the Fed in 2020—when the market pricing mechanism for a certain type of asset fails due to a sharp reduction in buyers, the government or central bank often replaces market buyers through direct acquisition or liquidity injection. The AI infrastructure is currently in a phase of rapid capital expenditure expansion, but commercial revenue has not yet fully materialized, showing similarities to the asset-liability structure during the expansion period of "two houses."
Essentially, this represents a combination of capital concentration and regulatory changes—if the government or central bank ultimately intervenes in AI infrastructure credit risks, it means that the market financing risks of the private sector are partially transferred to public balance sheets, and the credit expansion of AI infrastructure will gradually transition from "corporate credit" to "quasi-sovereign credit." The underlying mechanism is that when a certain type of asset is defined as a "national security priority," its financing costs and risk pricing will deviate from pure market logic, instead being dominated by policy backing expectations.
ABAB News · Law of Cognition
When losses are large enough, private risk becomes public credit.
Printing money does not solve problems; it merely changes the holder of the problem.
Once "safety first" is established, market pricing power yields to policy backing.