For years, the investment case for artificial intelligence has rested on one assumption above almost everything else: the technology will keep moving faster.
On Monday, investors were forced to consider what happens if the people building it decide that it shouldn’t.
AI-linked stocks fell sharply across Asia after some of the industry’s most influential executives backed calls to slow the development of increasingly powerful artificial intelligence systems.
The most dramatic move came from SoftBank Group.
Shares in the Japanese technology investor fell as much as 13.2% during trading, according to Reuters, before recovering some of those losses. SoftBank finished the session down 10.7%, according to AP.
The reason SoftBank attracted so much attention isn’t difficult to understand.
Its fortunes have become deeply connected to the AI boom — and particularly to OpenAI.
But the selloff stretched far beyond one company.
Memory-chip maker SK Hynix fell more than 6%, Samsung Electronics dropped more than 4%, Kioxia declined sharply and Taiwan Semiconductor Manufacturing Company also traded lower.
Suddenly, an argument that had largely belonged to AI researchers, regulators and safety specialists was showing up on trading screens.
The AI industry’s strange new message
The catalyst came over the weekend.
Anthropic CEO Dario Amodei published an essay titled We Must Pace the Frontier, arguing that companies developing the world’s most capable AI systems should deliberately slow the rate at which those systems become more powerful.
This wasn’t a vague request for companies to “develop AI responsibly.”
Amodei wants independent safety evaluators to receive unusually deep access to frontier AI companies. He also wants major laboratories to establish common safety standards and ultimately explore international coordination around particularly dangerous uses of artificial intelligence.
The remarkable part was what happened next.
OpenAI CEO Sam Altman backed the basic idea. Elon Musk, whose xAI is itself competing aggressively in frontier AI, also expressed support.
Three executives whose companies are involved in one of technology’s most expensive races were effectively acknowledging that the race may be moving too quickly.
Investors noticed.
Why SoftBank suddenly became the centre of the story
SoftBank is an unusually useful proxy for enthusiasm around OpenAI.
Masayoshi Son has made artificial intelligence central to the company’s strategy, committing enormous amounts of capital to OpenAI and the infrastructure expected to support the next generation of AI systems.
That works spectacularly well if AI capabilities, adoption and spending continue accelerating.
A deliberate slowdown introduces a question markets haven’t had to seriously price before.
What happens to the AI investment boom if frontier laboratories intentionally take longer between major capability jumps?
That doesn’t mean demand for AI chips suddenly disappears. It doesn’t mean OpenAI stops developing models. And it certainly doesn’t mean the artificial-intelligence boom is over.
Amodei himself isn’t arguing that AI research should stop.
The proposal is about buying time — slowing the rate of capability advancement while safety testing and safeguards catch up.
But markets don’t need an industry to stop growing to become nervous.
Sometimes uncertainty is enough.
The chip stocks matter more than SoftBank
The more interesting part of Monday’s reaction may actually be what happened elsewhere.
SK Hynix, Samsung, TSMC, Kioxia and other companies exposed to the AI infrastructure buildout also came under pressure.
That’s important because the AI boom isn’t simply a story about ChatGPT, Claude or Gemini.
Behind every new frontier model sits an enormous physical supply chain: GPUs, advanced memory, semiconductor manufacturing, networking equipment, electricity and increasingly enormous data centres.
Investors have poured money into those companies partly because the computational appetite of AI has appeared almost limitless.
If model developers voluntarily reduce the speed of the capability race — or regulators eventually force them to — markets will inevitably start asking whether infrastructure forecasts need to change too.
That question may prove premature.
Training powerful models still requires extraordinary amounts of computing capacity, and slower capability releases don’t necessarily mean less research or lower inference demand.
But Monday’s trading demonstrated something important.
Investors are finally beginning to think about the possibility.
This isn’t just an AI safety story anymore
Until now, warnings about advanced AI have often existed in a separate world from the financial story surrounding the technology.
Researchers debated alignment.
Governments debated regulation.
Silicon Valley debated artificial general intelligence.
Markets mostly watched revenue growth and ordered more Nvidia shares.
Those worlds are starting to collide.
Recent AI safety incidents have made the discussion harder to dismiss as hypothetical. Amodei specifically pointed to increasingly autonomous AI systems and the emerging possibility of AI helping to build subsequent generations of AI.
AP reported Monday that the latest warnings have reignited the long-running debate over whether increasingly autonomous systems could eventually threaten critical infrastructure or be misused at enormous scale.
There remains considerable disagreement among researchers about how probable the most extreme scenarios actually are.
But financial markets don’t require certainty.
They price risk.
And AI risk has suddenly become something investors can attach numbers to.
There are other reasons technology stocks are nervous
It’s important not to turn Monday’s market move into a cleaner story than it really is.
AI safety wasn’t the only pressure facing markets.
Oil prices were climbing amid geopolitical tensions. Investors were also watching interest-rate expectations and broader weakness across technology shares.
SoftBank itself carries risks beyond AI safety, including the scale and structure of its investments.
So it would be simplistic to say that a handful of comments from AI executives single-handedly erased billions of dollars in market value.
Markets rarely work that neatly.
What the reaction does show is how sensitive the AI trade has become to anything that challenges the assumption of uninterrupted acceleration.
For several years, investors have worried about whether AI companies could move fast enough.
Now they’re confronting a stranger possibility.
What if the companies themselves decide they’re moving too fast?
The AI boom has entered a different phase
None of this means the AI boom is ending.
If anything, the scale of Monday’s reaction demonstrates how economically important artificial intelligence has become.
A safety essay published by the CEO of Anthropic can now influence sentiment around semiconductor manufacturers thousands of kilometres away.
Comments from Sam Altman can matter not only to OpenAI employees or ChatGPT users but to shareholders in companies supplying the infrastructure underneath it.
That is a remarkable change from even a few years ago.
Artificial intelligence has moved from being a technology story to an economic one.
And that changes the stakes of the safety debate.
If Anthropic, OpenAI and other frontier laboratories genuinely begin pacing development, investors will eventually have to decide what slower capability growth means for valuations built around extraordinary expansion.
If they don’t slow down, another question emerges: were these warnings serious enough to justify the concern in the first place?
Either way, the market has been introduced to a new kind of AI risk.
For years, Wall Street worried that the artificial-intelligence boom might eventually slow down.
It probably didn’t expect the people building the technology to be the ones reaching for the brakes.
