CoreWeave is heading into Asia-Pacific for the first time, and it picked Indonesia as its entry point. The AI cloud company announced plans to build three new data centers there, adding up to 360 megawatts of contracted IT power once they come online in 2028.
Unlike expansion deals that lean on local partners, CoreWeave will own and operate all three facilities itself. That gives labs, startups, and enterprise customers running AI workloads in Southeast Asia access to the company’s full cloud platform, rather than a scaled-down regional version of it.
The move also tracks with where AI compute demand has been heading lately. Companies and governments across Asia increasingly want processing power sitting close to their data and users, partly for latency reasons and partly because of data locality rules that matter for regulated industries. Since CoreWeave has built its US and European business around exactly that kind of demand, Indonesia looks like a natural next step rather than a surprising one.
Sachin Jain, CoreWeave’s chief operating officer, framed the choice around the country’s own momentum in tech investment. He pointed to Indonesia’s push into AI education and digital infrastructure as a factor that’s already paying off, noting the country has become one of Southeast Asia’s fastest-growing destinations for global tech investment.
Beyond the physical buildout, CoreWeave says it plans to build a local team to run the new facilities. That also means investing in Indonesian technical talent instead of importing an entire operations staff, which fits with how Indonesia has framed its own digital ambitions: stronger infrastructure, more domestic AI capability, and a bigger role in the regional digital economy.
The Indonesia sites add to a footprint that’s already grown fast. As of March 31, CoreWeave operated 49 data centers worldwide, with more than 1 gigawatt of active power and over 3.5 gigawatts contracted for AI workloads. Along the way, the company has also picked up some notable performance marks, including record MLPerf benchmark results and the top ranking in both versions of SemiAnalysis’s ClusterMAX evaluation.
