The billion-dollar infrastructure deals powering the AI boom

The burgeoning AI industry requires unprecedented computing power, leading to estimated infrastructure spending of $3 to $4 trillion by the end of the decade, straining power grids. Major deals include Microsoft's initial $1 billion investment in OpenAI, which evolved into massive Azure commitments. This model has been replicated, with Amazon investing $8 billion in Anthropic. Oracle secured massive, future-looking deals with OpenAI, including a staggering five-year, $300 billion compute power agreement. Nvidia is also heavily involved, investing in customers like OpenAI ($100 billion in GPUs for stock) to maintain demand for its scarce GPUs. Hyperscalers like Meta plan massive capital expenditures, such as $600 billion in U.S. infrastructure through 2028, including building large data centers like Hyperion. Furthermore, a highly ambitious, though potentially stalled, $500 billion 'Stargate' project involving SoftBank, OpenAI, and Oracle was announced. This massive capital expenditure, projected to exceed nearly $700 billion among hyperscalers in 2026 alone, is causing nervousness among bankers, contingent on tech companies proving these investments yield returns.

(Source：TechCrunch)