On September 14, shares of companies involved in artificial intelligence saw a drop due to calls for a slowdown in AI technology development. Yahoo Finance reported that during early trading in the Asian market, shares of SoftBank, an investor in OpenAI, fell by as much as 13.2%.
Source: Yahoo Finance.The decline also affected semiconductor manufacturers, with Kioxia shares decreasing by 9.8%, SK Hynix by 5.3%, Samsung Electronics and Tokyo Electron by 3.7%, and TSMC by 1.2%.
Z.ai, a developer of GLM models, experienced a 10.5% drop. This decline followed a general sell-off in the sector and was exacerbated by the company's recent stock issuance at a discount.
On September 13, Z.ai announced it raised $5 billion, with around $2 billion coming from stock sales and another $3 billion from convertible bonds. The shares were issued at 714 Hong Kong dollars, which is 10% lower than the closing price on September 11.
Approximately 60% of Z.ai's net proceeds will be allocated for the development of next-generation models and their fully automated training systems, while 15% is earmarked for business expansion.
The calls for a slowdown in AI development have led investors to reconsider the profitability of investments in the infrastructure that supports it.
Swissquote analyst İpek Ozkardeskaya explained that if the demand for computing power grows at a slower pace, data centers may fail to meet revenue expectations. Moreover, these operators will still have to cover debts, pay rent, and fulfill electricity procurement obligations.
She noted that companies that built data centers using borrowed funds with the expectation of rapid demand growth, along with the lenders of such projects, are particularly vulnerable.
Proposals from AI Company Leaders
The reassessment of expectations was prompted by an essay by Anthropic CEO Dario Amodei, which received support from Sam Altman of OpenAI and Elon Musk of xAI.
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.
Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon. https://t.co/1YhhIybZX7
— Sam Altman (@sama) September 12, 2026
Amodei pointed out that the capabilities of models are advancing faster than the means to ensure their safety, citing the increasing involvement of AI in creating the next generation of systems as a concern.
Another worry arose from an incident involving OpenAI and Hugging Face where AI agents attacked external targets and attempted to breach their evaluation system. Amodei warned that more powerful neural networks with such behavior could cause significantly greater harm.
To mitigate risks, the Anthropic chief proposed three areas of focus:
- Independent audits within labs. External specialists should continuously access development processes, check adherence to safety measures, and report incidents.
- Common standards for companies. Developers should agree on safety standards and model development speed limits with government support.
- International agreements. Countries need to discuss joint measures considering the complexity of enforcement verification.
Anthropic has already committed to implementing the first point, while the others require agreements with other companies and governments.
Amodei emphasized that this proposal does not imply halting model training but aims to give developers and independent reviewers more time to assess risks.
A full international pause on AI development is deemed unlikely in the near future.
U.S. President Donald Trump has not supported the calls to slow AI development. Speaking to reporters in Ireland, he stated that maintaining American leadership over China remains a priority.
Trump acknowledged the possibility of protective measures but described the concerns raised by critics as exaggerated, claiming they discuss scenarios that are unlikely to materialize.
It is worth noting that on September 13, Altman ruled out an IPO for OpenAI in 2026, citing AI safety concerns. He remarked that the company had also discussed potential pauses in model development upon reaching new capability levels.