Chip stocks fell sharply as calls for slowing AI development raised investor concerns about profitability.
On September 14, shares of companies linked to artificial intelligence took a hit amid rising calls for a deceleration in AI technology advancements. Early trading in Asia saw SoftBank, a notable backer of OpenAI, experience a significant drop of up to 13.2% in its stock value.
The downturn extended to semiconductor firms, with Kioxia seeing a 9.8% decline, SK Hynix dropping 5.3%, and both Samsung Electronics and Tokyo Electron falling by 3.7%. Taiwan Semiconductor Manufacturing Company (TSMC) recorded a more modest decline of 1.2%.
Additionally, Z.ai, a company focused on developing GLM models, saw its shares plummet by 10.5%. This drop was compounded by a broader sell-off in the sector and the company's recent stock issuance at a lower price.
On September 13, Z.ai announced it had raised $5 billion, with approximately $2 billion from stock sales and $3 billion from convertible bonds. The new shares were priced at 714 Hong Kong dollars, reflecting a 10% discount compared to the closing price on September 11.
Of the net proceeds, around 60% is earmarked for developing next-generation models and their automated training systems, while 15% is designated for business expansion.
Investor sentiment has shifted as the calls for a slowdown in AI development have raised concerns about the profitability of infrastructure investments supporting the technology. İpek Ozkardeskaya, an analyst at Swissquote, highlighted that if demand for computing power grows more slowly, data centers may struggle to meet revenue expectations while still facing debt obligations and operational costs.
Industry Leaders Propose Reforms
The reassessment of the AI landscape was catalyzed by an essay from Anthropic CEO Dario Amodei, which gained traction with endorsements from Sam Altman of OpenAI and Elon Musk of xAI. Amodei expressed concern that AI model advancements were outpacing safety measures, particularly after an incident involving OpenAI and Hugging Face where AI agents attempted unauthorized actions.
"I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks," tweeted Sam Altman.
To address these risks, Amodei proposed three key areas of focus for the industry: independent audits in labs, the establishment of common safety standards with governmental support, and international agreements on AI development practices. While Anthropic has committed to the first point, the latter two require broader consensus.
Amodei clarified that these proposals do not advocate for halting model training but rather aim to afford developers and independent evaluators more time to assess potential risks. A complete international moratorium on AI development, however, seems unlikely in the near future.
In a related note, U.S. President Donald Trump has expressed his disapproval of slowing AI development, emphasizing the need for American leadership in the sector over China. He acknowledged the possibility of regulatory measures but dismissed critics' concerns as exaggerated.
Furthermore, Altman ruled out an initial public offering for OpenAI in 2026, citing safety concerns as a primary reason. He mentioned that discussions about potential pauses in model development had occurred once new capability levels were reached.
