Jardine chief: Asia's data centers hinge on US policy toward China

The billions of dollars going into Asian data centers rest on a narrow base: a handful of US technology companies and Washington's mood toward China. That is according to the head of Jardine, one of Hong Kong's largest conglomerates. He believes the ability of American giants to operate globally could shift sharply because of increasingly unpredictable US policy. Bloomberg published the statement without details on specific projects, so what was said and the surrounding context are separated below.

Why the risk is concentrated in a few names

The Asian data center market grows mainly on orders from large American cloud providers and on supplies of accelerators from American manufacturers. Regional developers and infrastructure funds buy land, negotiate power, build halls and procure cooling. But the tenant on the other side of the contract is most often one of a few global players. If that tenant is restricted from operating in a particular jurisdiction or market, a site with a decade-long payback period sits idle.

This model has an unpleasant feature. The larger the campus, the greater the dependence on a single tenant. Diversification is possible, but demand for AI capacity today comes precisely from these companies.

Where US policy hits Asia

Washington's chip policy has changed more than once in recent years. Export restrictions on advanced accelerators for China have been in place since 2022. In 2025, the administration rescinded the diffusion rule, which divided countries into tiers of chip access. Licenses for shipping scaled-down accelerators to China were required, then permitted again. Malaysia, where one of the region's notable data center clusters has grown up around Johor, introduced a permit regime for re-exports of American AI chips. For regional operators, each such decision means recalculating risk: supply, customer base, compliance.

The Hong Kong conglomerate looks at this from an investor's perspective. For it, the question is not whether demand for compute will rise or fall, but who will be able to serve that demand, and on what terms.

The main risk factors, as they follow from the statement and context:

• Investment volume: billions of dollars in Asian data center infrastructure.
• Concentration: a limited number of American tech giants as the key tenants and suppliers.
• Source of uncertainty: the unpredictable US stance on China, not market demand.
• Weak link: owners of sites with long-term contracts and fixed capital expenditure.
• What the available excerpt does not include: specific amounts, project names and timelines.

Practical takeaways for investors and operators

Lease contracts should spell out what happens when new export restrictions or sanctions are introduced, and who bears the cost of downtime. It helps to model in advance the scenarios in which the main tenant loses access to some of its equipment or to one of the region's markets.

Diversifying the customer base takes longer than construction, so it needs to start at the design stage. Regional clouds, telecom operators and enterprise customers help, even though they do not provide the scale of hyperscalers. Power, land and permits remain assets that can be repurposed for a different tenant.

Finally, Washington's licensing decisions and the responses of regional countries should be tracked as part of operational reporting, not as background news. In a market where building a single campus costs hundreds of millions of dollars, political risk is a line item just like energy.