Nvidia, Wall Street Firms Aim to Mobilize $500B for AI

On Aug. 10 Nvidia and six major financial firms signed memorandums of understanding to build independent financing platforms intended to mobilize more than $500 billion of third‑party capital over time to fund AI infrastructure.

The group includes Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR. The capital target represents aggregated outside investment that would be deployed over years, not a single committed fund from Nvidia.

Each financial partner will evaluate and underwrite financing opportunities independently. Lenders and asset managers will review customers’ demand forecasts, expected utilization, cash flow and projected residual value of equipment before committing capital. Nvidia has offered to provide residual‑value support of up to 25% on individual deals to help backstop asset value, but that support would be assessed case by case.

The platforms are designed to enable large‑scale leasing, revenue‑sharing and structured finance arrangements for GPU compute and related data‑center equipment. Nvidia’s chief executive framed the effort as a shift in the company’s business toward creating infrastructure that investors can back. Nvidia CEO Jensen Huang wrote on X, “NVIDIA has reached an important milestone. We began by building chips; today, we are helping create a new class of productive, investable infrastructure: AI factories. In AI, compute is revenue. NVIDIA compute is uniquely suited for this role.”

The participating firms bring large asset bases and financing experience. BlackRock manages roughly $15.3 trillion, Blackstone about $1.3 trillion, Apollo about $1.05 trillion and Brookfield over $1 trillion. Goldman Sachs and KKR will contribute investment‑banking, private‑capital and infrastructure financing capabilities. The partners expect to serve customers ranging from hyperscale cloud providers to smaller AI labs that need GPU capacity without large up‑front capital spending.

Several sizable AI infrastructure projects have already attracted external funding. Nvidia plans a $1 billion investment in South Korea’s Naver. Brookfield has a nonbinding term sheet to provide up to $9 billion for expanded AI capacity in South Korea. In Australia, SharonAI signed a six‑year collaboration covering 72 megawatts of data‑center capacity and up to 40,000 of Nvidia’s Grace Blackwell GB300 GPUs in a contract with a value up to $4.88 billion. HIVE Digital Technologies agreed to a three‑year AI cloud contract worth about $220 million involving Bell Canada and Cohere.

Market data cited by the firms point to large financing needs for AI infrastructure. Wall Street estimates global AI‑related debt issuance could reach about $570 billion in 2026, while data‑center capital expenditures are tracking near $850 billion. Financial analysts and lenders have noted that certain AI accelerators retain commercial value for multiple years; Nvidia’s A100, introduced in 2020, remained in commercial use several years later, which market participants say supports equipment redeployment and residual value assumptions. Goldman Sachs has included chip longevity in its AI capital‑spending outlook.

Rental rates for high‑end Nvidia chips have risen, reflecting demand. One‑year H100 rental rates moved from around $1.70 per hour in October 2025 to roughly $2.35 in March. Newer B200 chips have been rented at rates near $5.30 to $7.05 per hour. Some industry participants are testing decentralized compute networks that allow owners to rent GPU capacity on demand.

The financing platforms would not be controlled by Nvidia. Partner firms will retain final authority over which customers and projects receive funding, and Nvidia’s residual support would be considered on a project‑by‑project basis rather than pledged across the full $500 billion target. The companies say the structure is meant to complement independent underwriting and existing capital markets activity.

The memorandums are subject to final agreements and detailed terms remain to be negotiated.