Nvidia’s $366 Billion Commitment Backlog Redefines Its Revenue Trajectory

Nvidia disclosed $366 billion in multiyear artificial intelligence infrastructure commitments Wednesday, a figure that dwarfs its quarterly revenue and signals a fundamental shift in how the chipmaker’s business should be valued. The commitments, detailed alongside fiscal second-quarter results that beat estimates, jumped from $119 billion just three months earlier — driven almost entirely by a surge in supply and capacity procurement to $279 billion from virtually zero a year ago.

Chief Financial Officer Colette Kress attributed the increase primarily to memory procurement agreements secured across Nvidia’s supplier network to meet demand “for the next several years.” The company also carries up to $108.5 billion in phased guarantees tied to those arrangements, with $3.5 billion in maximum gross exposure currently recognized. Those guarantees represent potential liability rather than current losses.

The scale of the backlog reframes the $96.2 billion in quarterly revenue — itself a 106% year-over-year increase — as merely the near-term manifestation of a far larger contracted pipeline. Data Center revenue rose 117% to $89 billion. Adjusted earnings of $2.22 per share exceeded the $2.09 consensus. Gross margins held at 75%.

A separate agreement with SB Energy to support the PORTS-Pike Technology Campus in Ohio adds another dimension. Nvidia infrastructure deployed at the site, leased long-term to OpenAI, could involve approximately 1.5 million GPUs per generation — translating to $150 billion to $200 billion in Nvidia revenue per upgrade cycle over a 20-year horizon. Including both arrangements, maximum gross guarantee exposure stands at $108.5 billion.

The stock initially fell 1.59% in regular trading to $209.66 before rebounding to roughly $217 after hours, a 3.5% recovery. Investors appeared to digest the guarantee exposure before refocusing on the commitment trajectory.

CEO Jensen Huang framed the moment as an inflection point: “AI has reached its inflection point. It’s doing useful work. Its tokens are productive and profitable. Now, compute is revenue, and demand is accelerating.” He cited a shift from a single lab driving buildout a year ago to “a golden age of new AI labs and startups, multiple frontier labs scaling in parallel, a thriving open-model ecosystem and physical AI coming online.”

Nvidia guided third-quarter revenue to $108 billion, plus or minus 2%, with gross margins of 74%, plus or minus half a percentage point. Shipments of Hopper data-center products to China represented less than 1% of Data Center revenue. The company’s next-generation Vera Rubin platform is in full production.