AI-linked shares fall as industry leaders intensify calls for slower development

Safety warnings reach markets

Shares tied to artificial-intelligence investment fell across Asia on September 14 after leaders of major technology companies supported calls to slow the development of the most capable systems. The Guardian reported that SoftBank declined 13%, South Korea's chip-heavy Kospi index dropped 3% and Taiwan Semiconductor Manufacturing Company lost 1.2%. Nasdaq futures were also pointing lower before the US market opened.

The immediate trigger was a renewed public intervention by Anthropic chief executive Dario Amodei, followed by expressions of support from senior figures associated with OpenAI, Google DeepMind and SpaceX. The executives did not announce a coordinated halt to investment. Their statements nevertheless prompted investors to consider whether additional safety work, external testing or future regulation could slow the pace at which expensive computing infrastructure generates revenue.

A policy divide is taking shape

The market reaction comes as governments pursue different responses to advanced AI. President Donald Trump has rejected a slowdown, treating technological leadership over China as a strategic objective. In Britain, meanwhile, lawmakers and regulators are continuing to define safeguards for high-impact systems. That divergence leaves companies facing pressure to scale quickly in the United States while preparing for more structured oversight in Europe and parts of the United Kingdom.

Official policy work shows that the safety debate is broader than speculative warnings about future systems. A UK commission reporting on September 10 recommended stronger, adaptable rules for AI used in healthcare, with accuracy, human oversight and equal standards of care among its central principles. Although medical AI differs from frontier general-purpose models, the commission illustrates how governments are moving toward risk-based supervision rather than relying entirely on voluntary corporate promises.

The European Commission's ethics advisers have also called for anticipatory governance of systems combining AI and neurotechnology. Their September 8 recommendations emphasised protection of sensitive data, democratic accountability and institutional capacity to oversee emerging infrastructure. Neither official initiative ordered AI developers to stop work, but both demonstrate that calls for independent evaluation are landing in an active regulatory environment.

What investors will watch

The next test is whether the selloff lasts beyond a single session. Technology valuations rely heavily on continued demand for chips, data centres and model training, so any durable reduction in capital spending would affect suppliers well beyond the companies developing models. Investors will watch for concrete commitments from AI laboratories, government responses to the executives' intervention and evidence that safety programmes are delaying deployments. Until those emerge, the decline is best understood as an early repricing of regulatory and execution risk, not proof that the global AI investment cycle has reversed.