Nvidia Q2 2025: Record Revenue Signals AI’s Global Power Shift

Santa Clara, August 27, 2025 — Nvidia announced second-quarter revenue of 46.7 billion dollars, a 56 percent increase from the same period last year, driven by unprecedented demand for artificial intelligence chips. The company reported adjusted earnings per share of 1.05 dollars, slightly ahead of analyst expectations.

The data center division, which has become the backbone of Nvidia’s business, accounted for 41.1 billion dollars of revenue. This represented a 56 percent year-over-year increase, but results came in a little below Wall Street forecasts. Investors viewed this as evidence of both extraordinary demand and the limits imposed by market constraints.

Nvidia also authorized a 60 billion dollar stock buyback, underscoring its confidence in long-term growth and reinforcing the company’s strong cash position.

China and the H20 Dilemma

Despite record-setting revenue, Nvidia’s stock declined by around three percent in after-hours trading. The pullback reflected concerns over uncertainty in China, where sales of Nvidia’s H20 chip have stalled.

Chief Executive Jensen Huang confirmed that Nvidia did not record any H20 chip sales to China during the quarter. The company’s forecast for the third quarter of 54 billion dollars excluded any contributions from China. Analysts estimate that if export issues are resolved, Nvidia could add 2 to 5 billion dollars in revenue as early as next quarter.

The H20 chip was developed specifically to comply with United States export rules, but sales remain frozen until further regulatory clarity is achieved. Nvidia’s dependence on global markets, particularly China, highlights the intersection of technology and geopolitics in the artificial intelligence race.

AI Momentum and Market Questions

Nvidia’s results underline its role as the single most important supplier of AI infrastructure. The company’s Blackwell architecture, which powers next-generation AI models and enterprise systems, has already become the benchmark for performance in training and inference.

At the same time, investor commentary suggests unease about the pace of growth. Some analysts argue that Nvidia’s valuation is showing signs of excess, comparing parts of the AI boom to the early stages of the dot-com era. The debate is no longer about whether Nvidia is profitable, but whether growth at this scale can be sustained without disruption from competition, supply bottlenecks, or regulation.