Apollo Global Management

Nvidia shifts AI advantage from chips to capital

Nvidia is using its large cash reserves and credit capacity to extend its AI advantage beyond chip technology by financing AI infrastructure and partners. The company said it will provide up to $105 billion to support a large OpenAI data-center project in Ohio, including a $1.5 billion equity investment in SB Energy and commitments for power and lease support. Nvidia previously invested $30 billion in OpenAI and has arranged a broader pact with Wall Street firms to enable up to $500 billion in GPU financing. The strategy leverages strong free cash flow (reported at $48.5 billion in the latest quarter), dividend increases, and an $80 billion buyback to accelerate AI buildout and create financial moats around its systems.

Inflation Eases as Intel, Nvidia Drive AI Financing

Cooling inflation data and major AI-related financing moves drove market gains last week. U.S. inflation metrics (CPI and PPI) came in cooler-than-expected, easing pressure on the Federal Reserve. Intel raised capital via a large stock offering that was increased from $15 billion to $20 billion, while Nvidia announced partnerships with six major asset managers on a $500 billion financing initiative to make AI compute an investable asset class. The developments reinforced momentum in AI and lifted related semiconductor stocks, even as some indices diverged in weekly performance.

China risk threatens Nvidia’s $500B AI financing plan

Nvidia has struck agreements with six major asset managers — Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR — to mobilize roughly $500 billion to finance AI data centers and GPU clusters. Under the plan Nvidia will guarantee that GPUs used as collateral retain value, agreeing to cover up to 25% of any shortfall on liquidated collateral to help create a secondary market for aging GPUs and sustain demand for older hardware. Analysts warn the structure depends on GPUs preserving resale value — rapid depreciation or a surge of low-cost Chinese compute could crash collateral markets and push investor yields into the 11%–17% range — and introduces “wrong-way” risk because Nvidia’s obligations rise if demand weakens. Nvidia points to its CUDA software, continuous updates, and prior financial support for customers to extend chip productivity.