News

Analyzing the financial structure of AI startups and the resulting expansion of data center investment.

Friday, November 28, 2025 at 08:53 AM

An article explores the fundamental reasons behind the persistent increase in data center infrastructure spending, focusing on how the revenue models of capital-intensive AI startups drive demand for semiconductor and AI hardware suppliers.

Context

A November 28 Toyo Keizai report highlights how massive capital injections into AI startups are fueling an unprecedented data center investment boom. This trend is exemplified by Microsoft, which has earmarked roughly $80 billion for AI infrastructure capex in fiscal year 2025. Even larger projects are on the horizon, such as the "Stargate" initiative, which could channel up to $500 billion into building new GPU-dense data centers. This immense spending reflects the race to build the physical foundation for next-generation AI, though it also stokes investor fears of an AI bubble. This capital expenditure wave flows directly to the AI supply chain. Nvidia continues to be a prime beneficiary, reporting $41.1 billion in data center revenue for its second quarter of fiscal 2026 as its Blackwell platform ramps up. The demand extends beyond GPUs, benefiting foundries like China's Hua Hong Semiconductor. The company is expanding its 12-inch wafer capacity to 20,000 wafers per month by 2025 to produce essential power management and other chips needed for the data center build-out.

Related Companies

Microsoft
Microsoft
MSFT
US
Nvidia
Nvidia
NVDA
US
Hua Hong Semiconductor
Hua Hong Semiconductor
1347
CN