Nvidia's CEO Jensen Huang recently suggested the company could achieve 70% year-over-year growth, but this statement is better interpreted as a reflection of supply capabilities rather than a new official revenue forecast. Speaking at Goldman Sachs' technology conference on September 10, Huang expressed confidence in securing enough supply to support such expansion, yet Nvidia did not accompany his remarks with a formal fiscal 2028 revenue projection.
This distinction carries significant weight for a company valued at nearly $5.27 trillion. Huang also reiterated his belief that global spending on AI infrastructure could reach $3 trillion to $4 trillion by 2030. His argument extends beyond selling chips; he emphasizes Nvidia's role in delivering complete computing systems, networking, and software for data centers, with financial backing that helps specialized cloud operators secure essential resources like land, power, and buildings.
According to the official conference transcript, Huang identified land, power, and completed data-center shells as primary constraints. He noted that Nvidia is tracking gigawatt-scale projects across cloud providers, AI developers, and regional operators. This framing turns the "70%" claim into an execution test grounded in physical inputs, not merely an extrapolation of chip orders.
Strong Financial Foundation
Nvidia's recent performance provides a solid baseline. The company reported fiscal second-quarter revenue of $96.22 billion, up 106% year-over-year. Data Center revenue alone reached $89.0 billion, accounting for 92.5% of total revenue and growing 117%. For the current quarter, Nvidia guides to $108 billion in revenue, plus or minus 2%, with a gross margin target of 74% plus or minus half a percentage point.
As of September 11, NVDA shares closed at $218.29, essentially flat for the session. The company's market capitalization of roughly $5.27 trillion equates to about 13.7 times its annualized second-quarter revenue run rate of $384.9 billion. This multiple underscores how much future expansion is already priced into the stock.
The Financing Lever
Huang's financing strategy may be more impactful than his headline spending numbers. He described a $500 billion platform that he expects to be mostly asset-backed, with Nvidia's investments helping specialist "neoclouds" become distribution channels for its architecture. The model works in a loop: Nvidia helps a cloud operator finance capacity, the operator buys Nvidia systems, and contracted customers rent the compute. However, this also shifts some customer and project risk onto Nvidia's balance sheet.
Nvidia's latest Form 10-Q quantifies this exposure. Supply and capacity commitments jumped from $119 billion in the prior quarter to $279 billion as of July 26. The company also reported $99 billion in equity investments and another $25 billion in equity-investment commitments. Additionally, Nvidia has guarantees capped at $105 billion on 20-year data-center leases supporting OpenAI, plus an option to support about 3.8 additional gigawatts.
The filing warns that some AI-cloud agreements require Nvidia to purchase capacity if operators cannot sell their commitments to third parties. Delays in power, permits, construction, or customer financing could therefore hurt twice: revenue would arrive later while capacity or guarantee obligations remain. This is the strongest counterargument to treating every financed gigawatt as demand already earned.
Partners in Focus
CoreWeave and Nebius illustrate the potential and the risks. Nvidia invested $2 billion in CoreWeave at $87.20 per share, targeting more than five gigawatts by 2030. It separately invested $2 billion in Nebius, which also plans over five gigawatts of Nvidia systems by 2030. These partners broaden Nvidia's market reach, but their construction progress, utilization, and contracted revenue must be monitored independently of Nvidia's chip backlog.
The 70% growth case will become measurable well before 2030. Quarterly revenue must keep approaching or exceeding Nvidia's published guidance while gross margin stays near the mid-70s. Delivered power and data-center capacity must catch up with purchase commitments, and operating cash flow must remain comfortably ahead of investments and guarantees. If these three conditions hold, Huang's supply confidence has financial substance. If commitments keep rising faster than deployed, revenue-producing capacity, the growth story becomes more capital-intensive than the valuation allows.
