Malaysia has spent the past two years telling the world it wants to be the place where artificial intelligence gets built. The company now preparing the largest stock listing in history has quietly told its future investors that the same building spree carries a risk almost nobody has priced in.
According to CNBC, Anthropic, the maker of the Claude chatbot, will list public backlash against AI and the data centres that power it as a formal risk factor in its upcoming initial public offering. The prospectus is expected in the coming weeks, and the framing matters. A company only names something a risk in that document when it believes it could genuinely dent the business.
Table of Contents
A record listing with a warning attached
The scale is hard to overstate. Investors are targeting a valuation of around US$2 trillion when Anthropic goes public, reportedly as early as October, which would make it the biggest IPO ever, ahead of SpaceX's US$1.77 trillion debut earlier this year, Fortune reports. The company is running at more than US$65 billion in annual revenue, and that revenue is tied directly to how much computing power it can secure. Less compute means slower growth.
That is where the backlash comes in. A Gallup survey conducted in March found that 71 percent of Americans oppose an AI data centre being built in their local area, with 48 percent strongly opposed. Opposition ran higher than it did for nuclear plants. When a public turns against the physical footprint of AI, permits slow, projects stall, and the compute that companies like Anthropic depend on gets harder to build.
Why this lands in Malaysia
Malaysia sits on the opposite side of that equation. While many Americans are fighting data centres in their backyards, Malaysia has been actively courting them. The government has approved RM144.4 billion worth of data centre projects since 2021, according to the Ministry of Investment, Trade and Industry, and Johor has become one of Southeast Asia's fastest growing hubs for them. Kuala Lumpur has since tilted its policy to favour AI data centres specifically, even as it worries about the water and electricity they draw.
The bet is straightforward. If the world needs somewhere to build AI capacity, Malaysia wants to be that somewhere, complete with a RM2 billion Sovereign AI Cloud and a broader national push to move from merely using AI to actually building it. But a bet on being the supplier only pays off if demand holds. Anthropic's own risk disclosure is a reminder that the demand rests on public tolerance and market sentiment in places Malaysia does not control.
The risk is not the one everyone was watching
Earlier this month we looked at how the AI boom is partly financing itself, with chipmakers bankrolling the very customers who buy their chips. That is a supply-side fragility. What Anthropic is flagging is different, and arguably more stubborn. It is a social and political one. Hardware can be re-engineered. A public that has decided it does not want AI infrastructure nearby, or investors who cool on the whole trade, cannot simply be patched.
None of this means Malaysia's strategy is wrong. Being early to a real shift is how smaller economies win. But the smart version of that bet plans for demand wobbling, not only for it climbing. The largest IPO in history is about to be sold on a story of endless AI growth. The fact that its own filing hedges that story is the detail Malaysian planners should read twice.
Image(s) courtesy of Kevin Ache and Martijn Oudkerk on Unsplash.
