According to CoinMeta, Wu said that in his latest article “Safety First”, he argues that the emphasis by American AI companies such as Anthropics and OpenAI on “safety first” and the slowdown in AGI development are not primarily due to safety concerns, but rather because of insufficient demand for AI at current prices. If training expenditures decline, it will directly weaken the demand for computing power in data centers and chips. Hayes noted that although the AI laboratories themselves are not yet profitable, their demand for computing power has supported over $1 trillion in investment-grade debt, as well as hundreds of billions of dollars in lower-rated debt and loans. He further believes that if the demand for AI computing power falls short of expectations, related debts could be impaired, and risks could be transmitted through the insurance and reinsurance systems controlled by private equity. In either case, he thinks the ultimate outcome will be more money being injected into the market, which in turn will drive up the prices of Bitcoin and some crypto assets.