Chipmaking giant NVIDIA has delivered quarterly results that surpassed expectations, easing concerns about a potential AI bubble. The world’s most valuable publicly traded company reported third-quarter revenue of $57.01 billion, a 62% year-over-year increase and above analysts’ consensus estimate of approximately $54.9 billion. Adjusted earnings per share came in at $1.30, beating the anticipated $1.25. Net income also rose by 65% to $31.91 billion.
NVIDIA forecasts fourth-quarter revenue of around $65 billion—significantly higher than the analyst expectation of $61.66 billion. Its shares climbed more than 4% in after-hours trading, following a 3% gain during regular market hours, offering reassurance to a market unsettled by questions over the sustainability of AI-driven capital spending. The company’s GPUs currently power nearly all major AI developers and cloud providers, including Microsoft, Amazon, Google, Meta, Oracle, and OpenAI. With a market capitalization of roughly $4.5 trillion, NVIDIA’s performance not only shapes the semiconductor sector but also influences broader tech valuation sentiment.
CEO Jensen Huang highlighted “extraordinary demand” for the company’s newly launched Blackwell architecture, particularly the GB300 Blackwell Ultra chip, noting that cloud-based GPUs are selling out across major suppliers. NVIDIA reaffirmed that computational demands for both training and inference continue to grow at a compound rate as organizations deploy increasingly sophisticated foundation models and agent-based applications. One notable weakness this quarter was China: U.S. export controls prohibit NVIDIA from shipping its most advanced Blackwell chips to the region. Although the company secured a license to sell the downgraded H20 alternative, CFO Colette Kress disclosed it generated only $50 million in revenue.
The Data Center segment—the core engine of NVIDIA’s AI business—posted $51.2 billion in revenue, up 66% year-over-year and well ahead of the expected $49 billion. Of that, $43 billion came from compute products, primarily GPUs, while $8.2 billion stemmed from networking hardware used to interconnect large-scale clusters comprising thousands of chips. Kress noted that the Blackwell Ultra is now the company’s best-selling chip family, surpassing earlier architecture iterations. She added that NVIDIA now has visibility into AI infrastructure projects involving over 5 million GPUs across hyperscale cloud providers, national governments, industrial conglomerates, and supercomputing centers. Gaming revenue reached $4.3 billion, a 30% increase, while professional visualization sales surged 56% to $760 million, driven by demand for NVIDIA’s AI-optimized DGX Spark workstations. The smaller automotive and robotics segment reported $592 million in revenue, up 32%.
NVIDIA’s earnings come amid growing investor skepticism about AI valuations, which recently triggered a four-day decline in the S&P 500 as markets questioned whether tech giants would sustain heavy investments in AI hardware. Microsoft, Alphabet, Amazon, and Meta have collectively raised their 2025 capital expenditure outlook to $380 billion—primarily for data center expansion—intensifying scrutiny over whether these investments will yield returns. However, NVIDIA’s strong results and forward guidance suggest the build-out cycle is far from peaking. Additionally, the company repurchased $12.5 billion of its own stock and paid $243 million in dividends. Huang revealed that NVIDIA already has $500 billion worth of AI chip orders booked for 2025 and 2026, a figure Kress said “could” grow further as additional deals materialize.
