According to IDC's latest data published this month, domestic AI infrastructure spending will reach over $5.5 billion in 2026, growing at least 18% year over year — after a seven-fold expansion between 2022 and 2025.
But the headline number obscures the more significant story: by 2028, AI infrastructure spending will exceed non-AI infrastructure spending in Japan for the first time in history.
From Government Catalyst to Enterprise Engine
The initial surge was government-driven. Under the Economic Security Promotion Act, Tokyo accelerated large-scale GPU server deployments and national cloud initiatives. Many of these programmes are now reaching completion, permanently raising Japan's domestic AI compute capacity.
The next wave, however, is enterprise-led. IDC forecasts enterprise AI infrastructure spending to grow 5% year over year in 2026, with investment shifting toward business-critical domains including sales and marketing optimisation, customer service transformation, and R&D acceleration.
The market is expected to sustain a five-year compound annual growth rate of 13% through 2029, underscoring the structural and long-term nature of this expansion.
Building Sovereign AI Capability
Japan isn't simply expanding compute capacity. It is building sovereign AI capability — a deliberate strategic choice that distinguishes its approach from the rest of the region.
The investment encompasses:
- Domestic high-performance GPU clusters
- National-scale data centre expansion
- AI-optimised infrastructure environments
- On-premises enterprise AI platforms
SoftBank's late-2025 acquisition of DigitalBridge exemplifies this strategy. By securing control over extensive global data centre assets, Japanese capital has moved closer to the physical foundations of global AI deployment.
The Enterprise Opportunity
For vendors operating in the Japanese market, IDC's analysis points to a fundamental shift in what enterprise buyers want. Hardware supply models are no longer sufficient. Vendors must evolve toward ecosystem-based capabilities that include facilities integration, lifecycle support, managed AI capabilities, and infrastructure optimisation.
Japanese enterprises are also pursuing a distinctive approach to AI models. Rather than chasing scale, firms like NTT with its tsuzumi model and startups like Sakana AI are building lightweight, efficient models designed for on-premises deployment.
What This Means for APAC
Japan's AI infrastructure buildout has implications beyond its borders. As the country becomes a major hub for AI compute capacity, neighbouring markets — particularly in Southeast Asia — stand to benefit from improved connectivity and potential infrastructure partnerships.
The 2028 tipping point — when AI infrastructure overtakes non-AI spending — will mark a new era not just for Japan, but for how the entire region thinks about the foundations of AI competitiveness.