Does the slowdown in AI development mean that computing power demand has peaked? The answer may not be so pessimistic.

AI giants have unusually signaled a "slowdown in development," leading to a sell-off in the global AI and semiconductor sectors. However, some market participants believe that the short-term emotional shock is not enough to reverse the investment logic for AI infrastructure.

On September 14, AI and semiconductor stocks generally declined. In Japanese and South Korean stock markets, SK Hynix fell more than 6%, and Samsung Electronics fell 4%. In pre-market trading in the US, memory chip stocks also weakened collectively, with Micron Technology falling nearly 5%, Western Digital falling nearly 4%, and SanDisk falling more than 5%.

Behind the market sell-off is a rare collective statement from AI giants recently. Anthropic CEO Dario Amodei published a lengthy article on the X platform last Saturday, calling on AI companies to proactively slow down the development of cutting-edge models and announcing that Anthropic will unilaterally introduce an independent third-party evaluation mechanism.

OpenAI CEO Sam Altman subsequently expressed his support, and xAI founder Elon Musk also posted that "Dario is right." This rare consensus among the three AI giants quickly impacted market sentiment and sparked investor concerns about the AI capital expenditure cycle.

However, some investors believe that "slowing down development" does not mean that the AI capital expenditure cycle has reversed. The demand for computing power, data center construction, and infrastructure such as electricity, storage, networks, and cooling remains. If the pace of model development slows down, it may actually give the industry more time to digest previous investments and promote the commercialization of AI applications.

AI giants unusually "hit the brakes," leading to a revaluation of chip stocks.

In its article, Amodei warned that within the next 6 to 12 months, AI agents could gain the ability to "control the entire internet," with potential losses reaching hundreds of billions of dollars. Previously, Anthropic's threat intelligence report disclosed that its Claude model had been used in weapons development, cyberattacks, surveillance, and fraud.

These warnings quickly resonated within the industry. Altman stated that the risk of human extinction posed by AI was "unacceptable," while Musk publicly supported Amodei's views.

After the news reached the capital markets, the AI industry chain was the first to feel the pressure. Takayuki Miyajima, senior economist at Sony Financial Group, said that weekend comments about slowing down AI development would put selling pressure on Japanese AI and semiconductor-related stocks, and the uncertainty in the Middle East further amplified risk aversion.

AI assets, which were already at high levels, are particularly sensitive to changes in expectations. Charu Chanana, chief investment strategist at Saxo Markets Singapore, said that the valuations of AI and chip stocks are based on both strong demand and expectations of continued rapid technological progress. "Even a possible delay is enough to trigger profit-taking."

Slowing down R&D does not mean that computing power demand has peaked.

However, some investors believe that the market may equate "slowing down AI development" with "cutting investment in AI infrastructure."

Billy Leung, an investment strategist at Global X Management in Sydney, believes that the three CEOs' support for slowing down development will not directly change investment in chips, power, and data center infrastructure; on the contrary, it may lengthen the entire development cycle.

"If commercialization and AI applications continue to grow, while the pace of new capability development slows slightly, the industry may shift from 'spending money to build' to monetizing existing assets."

Gary Tan, portfolio manager at Allspring Global Investments in Singapore, also believes that the statements may create short-term pressure, but are unlikely to undermine the long-term AI investment theme. He points out that the AI industry is still in a relatively early stage, and given the rapid evolution of technology, other participants in the ecosystem may not be willing to slow down in tandem.

In other words, what the market really needs to focus on is not whether AI R&D will temporarily slow down, but whether AI capital expenditure will substantially contract as a result . If projects in data centers, power, networks, and storage are still progressing, then the long-term logic of chip demand has not fundamentally changed.

The real test: Can the massive investment in AI deliver returns?

Compared to the AI security controversy itself, the deeper concern of the capital market lies in whether the massive investments in AI infrastructure over the past few years can ultimately translate into sufficient commercial returns.

Sebastien Mallet, London-based portfolio manager at T. Rowe Price, stated that AI "will change the world" does not mean that every related investment will yield an attractive return. As capital investment continues to expand, investors are shifting their focus from "how much computing power AI needs" to "how much money that computing power can ultimately generate."

At the same time, stricter security regulations may also create new investment opportunities . Chanana believes that areas such as cybersecurity and AI surveillance may see additional demand, while infrastructure companies such as storage, networking, cooling, and power equipment may continue to benefit from projects already underway.

She pointed out that introducing more security safeguards will not make the demand for computing power and AI applications disappear out of thin air; on the contrary, it may make the development path of the AI industry more prudent and sustainable.

Of course, there are also very different voices in the market. Michael Burry, an investor known for his accurate prediction of the 2008 U.S. housing crisis, wrote on the X platform that the recent security warnings in the AI industry may just be "hype and bluff" intended to cover up the real and out-of-control slowdown in growth.

Therefore, for the AI industry, the core issue may have shifted from "whether AI will continue to develop" to whether the massive capital expenditures on AI can be sustained, and when computing power investments can truly translate into profits . This will determine whether the current AI infrastructure boom is merely an emotional pullback or enters a phase of valuation and business model repricing.

Risk warning and disclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.