NVIDIA Partners Target $500 Billion AI Factory Financing Market
NVIDIA teams with Apollo, BlackRock and other investors to mobilize over $500 billion for AI factories, with residual support capped at 25% per project.
Summary
On August 12, 2026, NVIDIA announced independent financing platforms with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR, designed to mobilize more than $500 billion in aggregate third party capital over time for AI infrastructure. The sum is not NVIDIA revenue, one fund or a commitment to one customer. Qualified AI labs, enterprises and AI clouds could finance AI factories at scale, while each institution independently evaluates customers, demand, utilization, cash flow and residual value. NVIDIA supplies the platform and may provide project specific residual value support for up to 25% of an opportunity, which it says is below other compute financing arrangements and does not replace underwriting.
NVIDIA defines a DSX AI factory as accelerated computing, networking, systems software, AI frameworks and a global developer ecosystem supporting language, vision, speech, biology, physical AI and robotics. Its common architecture runs across major clouds, systems makers and enterprises, allowing capacity to shift among customers, clouds and operators. NVIDIA says CUDA upgrades improve existing installations' performance, efficiency and ownership costs. The Ampere based A100, introduced in 2020, remained commercially active six years later for training, fine tuning, inference and high performance computing, with multiyear capacity commitments extending its expected economic life toward a decade.
NVIDIA cites one year H100 rental prices rising from about $1.70 per GPU hour in October 2025 to $2.35 in March 2026, while cross provider on demand median pricing climbed from roughly $2.00 in October 2025 to $2.70 in June 2026. Reported B200 cloud rates are $5.30 to $7.05. NVIDIA argues that demand from frontier labs, startups, enterprises, clouds and countries makes compute revenue producing and redeployable. Capacity will proceed where financing partners validate customer economics.
Positives
- More than $500 billion in third party capital could expand AI infrastructure access for qualified AI labs, enterprises and AI clouds.
- H100 one year rental pricing rose from about $1.70 per GPU hour in October 2025 to $2.35 in March 2026.
- A100 systems introduced in 2020 remained commercially active six years later, with commitments potentially extending their economic life toward a decade.
- NVIDIA DSX supports language, vision, speech, biology, physical AI and robotics workloads through one common architecture.
- Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR will independently underwrite customers, utilization, cash flow and residual value.
Risks & concerns
- More than $500 billion is a long term mobilization target, not NVIDIA revenue, a single fund or committed financing for one customer.
- Up to 25% of an opportunity may receive NVIDIA residual value support, leaving the company with limited project specific exposure.
- Access to affordable, large scale financing remains uneven among AI companies, enterprises and cloud providers despite strong compute demand.
- Each project must pass independent assessments of demand, utilization, cash flow and residual value before capacity receives financing.
- Circular financing and market absorption are explicit concerns, although NVIDIA says independent underwriting and customer economics will constrain construction.