Results Review - Nvidia beats, Guides Q3 Revenue to $108bn and Sees 70% FY2028 Growth

Nvidia delivered another broad beat in Q2 FY2027:

• Revenue doubled yoy to $96.2bn, ahead of the $92.3bn consensus.
• Non-GAAP EPS reached $2.22 versus the expected $2.09.
• Q3 revenue guidance of $108bn exceeded consensus of ~$104.9bn and implies ~12% qoq growth.
• Data Center revenue rose 117% yoy to $89bn.

Key takeaways

• The 70% FY2028 revenue growth outlook is constrained by available capacity. Management said Nvidia currently has enough supply to support ~70% growth. Based on customers’ forecasts, demand could support growth closer to 100% if more capacity were available. That does not make 70% a guaranteed floor, but it suggests the outlook does not capture all the demand Nvidia says it can see.

• The spending and compute requirements behind that demand are unusually large. Management highlighted:
  • More than $2tn of cloud-industry backlog.
  • Nearly $800bn of capital spending by the five largest hyperscalers in 2026.
  • An expected increase in that spending to $1.3tn in 2027.
  • A potential 15–100x increase in compute usage as AI agents reason, plan and use tools repeatedly.

These are management estimates, but they explain why Nvidia believes the infrastructure cycle still has room to expand.

• Nvidia is helping convert demand into physical capacity. The company is working with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR on platforms intended to mobilize more than $500bn of third-party capital. Nvidia’s own future commitments have reached $366bn, while disclosed maximum gross guarantee exposure stands at $108.5bn. These arrangements may accelerate new capacity, but they also make counterparty risk and cash conversion more important.

• “AI has reached its inflection point.” Huang’s argument is that AI compute is moving from experimental spending toward revenue-producing infrastructure. Vera Rubin is already in full production, but the claim still needs to be tested: customer utilization and AI monetization must grow fast enough to absorb the capacity now being financed and built.

Key debates

• Is 70% FY2028 growth really the demand outlook—or simply the limit of what Nvidia can supply?

• Will agentic AI deliver the 15–100x compute multiplier management expects?

• Does Nvidia’s $500bn financing push expand the market—or bring customer risk closer to its own balance sheet?

• Can customers turn $1.3tn of hyperscaler capex into enough AI revenue to keep the cycle accelerating?

Source:

1. Nvidia: Company Press Release