SoftBank breaks ground on US$15B Stargate Japan data center campus

SoftBank broke ground on a US$15 billion data center campus in Japan, the first major deployment under the Stargate Japan partnership that will bring OpenAI's infrastructure to Japanese soil. The campus is planned to include GPU clusters optimised for AI training and inference, with OpenAI services delivered through the facility once operational. Completion of the first phase is targeted for 2027, making this a medium-term infrastructure commitment rather than an immediate change to available AI services.

**What this means for Japanese data residency requirements.** Japan's APPI and FSA guidelines impose data localisation requirements on certain categories of sensitive data, but Japan has not historically mandated that AI inference occur within Japanese borders — unlike China's Cybersecurity Law requirements. The significance of the Stargate Japan facility is primarily commercial rather than regulatory: it provides Japanese financial institutions, healthcare providers, and government agencies with contractual certainty that OpenAI's inference occurs within Japan, which satisfies board-level governance requirements even where not strictly legally mandated.

**Scale and timeline context.** US$15 billion over the planned build-out period implies a facility comparable to hyperscaler-tier Japanese data centers. For reference, AWS's Japan region infrastructure has been built over a decade at roughly $2–3 billion per year; $15 billion from SoftBank/OpenAI suggests a compressed 3–5 year build that will meaningfully expand available AI compute capacity in Japan. The timing aligns with Japan's government-driven AI capacity investment that began with the GENIAC programme.

**Impact on APAC AI infrastructure supply.** Japan's Stargate facility adds significant GPU compute capacity to APAC's AI infrastructure, alongside similar investments in Singapore and Australia. The aggregate effect is a reduction in inference latency for APAC users of OpenAI services and an increase in available compute for AI model training within the region. This benefits all APAC enterprises using OpenAI APIs, not just Japanese customers, through improved regional edge infrastructure.

**AIMenta's editorial read.** For Japanese enterprises evaluating OpenAI deployment, the Stargate campus announcement is a directional commitment — meaningful for long-term planning but not yet operational. For practical 2025–2026 deployment decisions, Japan-region Azure (which already hosts OpenAI's enterprise tier) provides the data residency option available today. Plan around what's available now; incorporate the Stargate capacity into your 2027+ infrastructure assumptions.