Artificial Intelligence
SpaceX Lands Roughly $40 Billion NVIDIA GPU Lease Commitment

Key PointsSpaceX frames the Apollo-led lease as the largest committed GPU financing ever…
View more
- SpaceX frames the Apollo-led lease as the largest committed GPU financing ever assembled.
- The deal spreads SpaceX compute funding across Apollo Goldman BlackRock and KKR rather than one lender.
- SpaceX claims sub-year payback on its compute investments as demand for Grok accelerates.
SpaceX (SPCX ) said on October 26, 2026, that it has reached a binding commitment to enter into an approximately $40 billion master finance lease for GPUs and related equipment with an entity managed by Apollo Global Management (N7I.DE ). The lease will deliver the next generation of NVIDIA Vera Rubin chips for SpaceX’s terrestrial compute buildout, SpaceX said in an announcement hosted on Apollo’s press page. The company described the commitment as an important step in its plan to build what it calls a Super Intelligence future and characterized the transaction as the largest committed GPU financing.
Committed Parties, Conditions and Timeline
The financing has been committed by a number of financing parties led by Apollo and Goldman Sachs (GS ), with Goldman Sachs serving as sole structuring agent, according to the announcement. Funds and accounts managed by BlackRock and Goldman Sachs Asset Management, together with capital accounts advised by KKR, have also made anchor commitments in the transaction.
Entry into the lease remains subject to certain specified closing conditions and is expected to close prior to year-end 2026. Lease schedules are expected to commence upon delivery of the GPUs and related equipment in the first half of 2027. The announcement states that additional information regarding the terms of the lease will be available in a SpaceX Current Report on Form 8-K filed with the U.S. Securities and Exchange Commission and on the company’s investor relations website.
In commentary the announcement attributes to SpaceX’s chief financial officer, the company said it aims to deploy the largest compute infrastructure in the world and that, as demand for compute continues to accelerate, it is seeing payback periods of less than one year on its SI infrastructure investments. “This lease diversifies our funding sources and addresses a notable portion of the funding required to deliver on our growth plans as we aim to get additional compute online at record speeds to power Grok for enterprises and consumers, as well as serve existing and future cloud services customers,” the commentary states.
Financing Platforms and SpaceX’s Compute Program
The commitment follows an August 10, 2026, announcement in which NVIDIA (NVDA ) said it had signed memorandums of understanding with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to establish independent compute financing platforms intended to mobilize more than $500 billion of third-party capital for the buildout of AI infrastructure over time. NVIDIA said the agreements were designed to create dedicated pools of capital at significant scale for its customers, including frontier AI labs, enterprises and AI clouds, and characterized the platforms as turning its compute and full-stack AI infrastructure into an investable asset class for global capital. The chipmaker said the partnerships remained subject to execution of final agreements.
“Modern compute has emerged as a scarce, mission-critical asset class with compelling investment characteristics that is positioned to drive significant long-term economic growth and productivity gains,” Apollo President Jim Zelter said in that announcement. NVIDIA founder and chief executive Jensen Huang said the partnerships were helping create what he called a new class of productive, investable infrastructure: AI factories. Goldman Sachs chairman and chief executive David Solomon said the firm’s investment and distribution roles reflected confidence in NVIDIA’s leadership and pointed to what he described as a new opportunity to create a market for credit backed by NVIDIA compute.
SpaceX has separately detailed an orbital AI-compute initiative on its Starmind product page. The company says its AI1 satellites carry a localized compute payload of up to 250 kilowatts peak and 175 kilowatts average on a vehicle with a 75-meter wingspan and a deployed height of 30 meters, capturing solar power in sun-synchronous orbit and beaming results back through high-bandwidth laser links to the Starlink constellation. SpaceX presents space-based compute as the fastest way to scale given what it describes as the growing difficulty of finding electrical power sources on Earth, and says the satellites use proven technology and are more streamlined than the Starlink satellites it already manufactures at scale. The company says it is building the Gigasat Factory in Bastrop to manufacture AI satellites at scale, with production planned to start as soon as late 2027, and that it plans to produce advanced AI chips at a new fab facility as a joint project with Tesla. The October 26 announcement specifies that the leased Vera Rubin chips are for SpaceX’s terrestrial compute buildout.












