SoftBank shares fell by more than 10% on Monday after senior figures in the artificial intelligence industry backed calls for a coordinated slowdown in the development of the most advanced AI systems. The sell-off highlights growing investor concern that safety risks, regulatory intervention and geopolitical tensions could affect the pace and profitability of the sector.
The Japanese investment conglomerate is a major investor in OpenAI, making its share price particularly sensitive to changes in expectations around AI development. Other technology and semiconductor stocks across Asia also declined as investors assessed warnings that the race to build increasingly powerful models may be moving faster than safety research.
Why SoftBank shares fell
The immediate catalyst was a series of public warnings from leading AI executives and researchers. Anthropic chief executive Dario Amodei called for companies to “pace the frontier”, arguing that a coordinated slowdown would give developers more time to understand and manage the risks associated with advanced systems.
Amodei has raised concerns about “recursive self-improvement” — a scenario in which AI systems assist in creating more capable successors. He argued that this possibility could outpace humanity’s ability to understand or control the technology and should therefore be approached cautiously.
OpenAI chief executive Sam Altman and xAI chief executive Elon Musk publicly supported the warning. Altman also said OpenAI would not proceed with an initial public offering during 2026, according to an interview published by Fortune, as the company continues to focus on safety.
Warnings from AI researchers
The debate intensified after Anthropic researcher Jacob Coxon resigned and warned that AI systems could escape human control. Another Anthropic researcher, Evan Hubinger, said he believed there was a significant risk that AI could cause catastrophic harm within the next decade.
Those comments have not established that such an outcome will occur. They do, however, demonstrate the depth of disagreement within the industry over how quickly frontier AI should advance and how much testing should take place before more capable models are deployed.
Technology stocks across Asia also declined
SoftBank was not the only company affected. Semiconductor and technology shares fell in several Asian markets, including:
- SK Hynix, which declined 5.3%.
- Samsung Electronics, down 2.8%.
- Tokyo Electron, which fell 0.9%.
- Kioxia Holdings, down 6%.
The declines suggest that investors were considering wider consequences for the AI supply chain, including demand for advanced memory, chip-making equipment and computing infrastructure. A slower development cycle could reduce the speed at which companies invest in data centres and specialised hardware, although the effect on future earnings remains uncertain.
Could regulation slow AI development?
Morningstar analyst Dan Baker said SoftBank’s decline may reflect concerns that regulators could intervene to reduce the risk of worst-case outcomes. Investors may also be assessing whether additional examples of AI systems behaving unpredictably would encourage governments to impose stronger controls.
The warnings do not represent a new European Union law or a formal decision by the European Commission. They are industry and research statements rather than binding regulatory measures. Any future restrictions would require action by relevant national governments or international institutions and would depend on the specific risks being addressed.
For European policymakers, the debate is relevant to ongoing discussions about AI safety, transparency, accountability and the enforcement of existing technology rules. The EU AI Act already establishes obligations for different categories of AI systems, but the source material does not indicate that the latest comments have changed the Act or triggered a new EU regulatory decision.
The race with China adds geopolitical pressure
Amodei also called for tighter restrictions on China’s access to advanced AI chips and semiconductor-manufacturing equipment. His argument was that access to high-performance chips would play a major role in determining China’s ability to develop increasingly capable AI systems.
He acknowledged that a coordinated international slowdown would be difficult if China chose not to participate. China’s state-backed Global Times criticised the position as an attempt to restrict the country’s technological development. At the same time, President Xi Jinping has proposed a BRICS open-source AI zone intended to deepen cooperation on AI models and training.
The disagreement illustrates the tension between AI safety and strategic competition. Governments and businesses are seeking to limit dangerous uses of the technology while also trying to avoid falling behind in a rapidly developing global market.
What happens next for investors and policymakers?
The immediate market reaction does not establish that AI investment is ending or that a global development pause has been agreed. It shows instead that investors are beginning to price in a broader range of risks, including:
- Potential regulation of advanced AI models.
- Higher safety and compliance costs.
- Delays to commercial launches.
- Restrictions on semiconductor exports and manufacturing equipment.
- Uncertainty over the future demand for computing infrastructure.
SoftBank’s exposure to OpenAI and other technology ventures means the group may remain particularly sensitive to changes in sentiment. Semiconductor companies could also be affected if demand forecasts change, although their performance will depend on the wider technology cycle as well as AI spending.
In Europe, the next key questions concern how existing AI rules are enforced and whether governments seek additional international cooperation on frontier-model safety. No such new measure has been announced in the information available for this report.
Conclusion
SoftBank shares plunged as investors responded to warnings that the AI industry may need to put caution before speed. The comments from Anthropic, OpenAI and other technology figures have reopened debate about control, safety and competition with China, but they have not produced a formal development freeze or new European regulation. The main takeaway is that AI companies and their investors may increasingly have to account for safety obligations and geopolitical constraints alongside the promise of rapid growth.
