LOGO

The global digital infrastructure market is rapidly expanding, fueled by the Artificial Intelligence (AI) boom and surging demand for cloud services. The Asia-Pacific region, in particular, has seen a flurry of major Data Center (DC)-related deals, with transactions totaling over 800 billion yen completed in September 2025 alone. As major US tech giants like Amazon, Google, and Meta continue to pour in investments, the importance of infrastructure supporting AI and cloud foundations is escalating.

Oracle Intensifies Offensive with $300 Billion AI Cloud Contract

Oracle’s latest move has captured significant attention across the industry. In September 2025, the US-based company signed a colossal cloud supply agreement with OpenAI, valued at approximately 44 trillion yen ($300 billion) over five years. This deal positions Oracle’s cloud infrastructure to support the massive training and inference processing demands of the generative AI developed and operated by OpenAI.

The two companies are already developing data centers with a power capacity equivalent to 4.5 gigawatts in the US, playing a part in the “Stargate Project” which also includes SoftBank Group, among others.

Responding to the explosive growth in AI demand, Oracle is aggressively accelerating its data center investments. In the quarter from June to August 2025, the company announced that its remaining performance obligations (RPO) had reached $455 billion, a 3.3-fold increase from the previous quarter, stating it had “closed multiple billion-dollar deals with three major customers.” Furthermore, a data center financing package of approximately 5.64 trillion yen, led by US banks J.P. Morgan and Mitsubishi UFJ Financial Group, is now active, paving the way for the construction of massive facilities in Wisconsin and Texas. These sites are expected to be primarily dedicated to OpenAI’s operations, forming the core of Oracle’s AI infrastructure strategy.

Expanding into Asian Markets: The New Infrastructure Race to Power the AI Society

Meanwhile, the major cloud vendors continue their aggressive expansion in the Asian market. Bain Capital sold its Chinese data center portfolio for around 588 billion yen, and Singaporean sovereign wealth fund GIC and the Abu Dhabi Investment Authority (ADIA) collectively invested over 230 billion yen into Malaysia’s Vantage Data Centers.

Across Asia, US tech leaders including Amazon and Google are investing an estimated 23 trillion yen annually in data centers. Establishing the foundational infrastructure to support AI and cloud growth has become the central focus of this competitive landscape.

Moving forward, data centers are set to transition from being mere server hubs to becoming a “new societal infrastructure supporting the AI society.” Companies prioritizing energy efficiency and the integration of renewable energy will hold the key to sustainable growth. The movements of global players like Oracle in this foundational build-out for the AI and cloud era will remain a crucial trend to watch.

The explosive growth of generative AI has led to a skyrocketing demand for the computing resources necessary to develop and run AI models. Amid this surge, news that OpenAI is moving to invest in its own data centers (DCs) has captured significant attention. This move signals a change not just in AI development structure, but also a shift in the leadership of IT infrastructure itself.

This development is set against the backdrop of strategic investments by NVIDIA in OpenAI. NVIDIA, which controls the supply of GPUs, is forging closer ties with AI development companies to secure computational power as AI demand expands. By owning its own data centers, OpenAI appears to be aiming for end-to-end optimization, from model development straight through to deployment and operation.

The Shift of Power: From Hyperscale Clouds to AI Companies

Historically, the data center landscape has been dominated by hyperscale cloud vendors like Amazon (AWS), Microsoft (Azure), and Google Cloud. These companies deploy hundreds of megawatts of DC capacity globally, utilizing proprietary server designs and specialized cooling technologies. For instance, AWS is expanding its self-built facilities in Northern Virginia and Singapore, while Microsoft is growing its footprint in Ireland and Japan. Google is also continuing its expansion across Europe and the U.S.

The move by AI model developers like OpenAI to own their data centers represents a new trend. Previously dependent on the environments provided by cloud vendors, the evolution of AI models is driving a greater need for AI-specific designs that prioritize high-density computing and cooling efficiency. Going forward, it is anticipated that the specialized requirements of OpenAI, not the general specifications of the hyperscale clouds, will increasingly take precedence.

Establishing “Compute Sovereignty” as the Next Competitive Edge

Many cloud vendors adopt a “hybrid strategy”: they initially utilize leases or co-location agreements, then transition to expanding self-designed facilities as they mature. OpenAI is likely following this path, aiming to optimize availability, efficiency, and cooling performance through its own DCs. This is not merely an infrastructure investment; it is a critical step toward establishing “compute sovereignty”—a factor that will determine future AI competitiveness.

As lower-cost AI models, particularly from Chinese contenders, begin to emerge, OpenAI is aggressively strengthening its unique technological foundation. This deep commitment, extending beyond the performance of the AI model to the very infrastructure that runs it, is poised to fundamentally reshape the industry structure. With the main driver of data center investment potentially shifting from cloud providers to AI companies, the future of next-generation infrastructure is now squarely in the spotlight.