NVIDIA's earnings break AI anxiety: Q1 revenue up over 70%, Jensen Huang says the inflection point of proxy AI has arrived, and customers are increasing their investment in computing power

AI (Artificial Intelligence) chip leader NVIDIA surpasses expectations with revenue and provides an optimistic outlook, easing some of the AI-related anxiety.

On February 25th, after the US stock market close, NVIDIA (Nasdaq: NVDA) released its Q4 FY2026 financial report ending January 25th. During the quarter, revenue reached $68.127 billion, up 73% year-over-year, beating market expectations of $65.8 billion; net profit under US GAAP was $42.96 billion, up 94%; Non-GAAP diluted earnings per share were $1.62, up 82%, exceeding market expectations of $1.53.

Additionally, NVIDIA provided guidance for the next quarter, forecasting FY2027 Q1 sales of a staggering $78 billion, with a +/- 2% variance, surpassing market expectations of $72.8 billion. NVIDIA stated that this forecast does not include data center revenue from China.

For the entire FY2026, NVIDIA achieved total revenue of $215.938 billion, up 65% year-over-year; GAAP net profit was $120.067 billion, up 65%; Non-GAAP diluted earnings per share were $4.77, up 60%.

NVIDIA CEO Jensen Huang said, “Computing demand is growing exponentially — the inflection point for AI is here. Today, the Grace Blackwell chip with NVLink is the king in inference — reducing the cost per token by an order of magnitude — and Vera Rubin will further extend this lead. Adoption of intelligent agents by enterprises is soaring. Our customers are racing to increase AI computing power — these compute factories are fueling the AI industrial revolution and future growth.”

Summary of NVIDIA’s latest quarterly performance. Source: Financial report

In FY2026, NVIDIA returned $41.1 billion to shareholders through stock buybacks and cash dividends. As of the end of Q4, the company still has $58.5 billion remaining in authorized stock repurchase programs. NVIDIA will pay a quarterly dividend of $0.01 per share to all shareholders registered as of March 11, 2026, on April 1, 2026.

On February 25th, NVIDIA (Nasdaq: NVDA) stock rose 1.44%, closing at $195.62 per share, with a market cap of $4.78 trillion. After the earnings release, NVIDIA’s stock price briefly surged over 3% in after-hours trading, then the gains narrowed.

Jensen Huang: “We are about to reach an agreement with OpenAI”

During the earnings call, NVIDIA CFO Colette Kress stated that the next-generation Vera Rubin series chips are on track for mass production and shipment starting in the second half of this year. The company “delivered the first Vera Rubin samples to customers earlier this week.”

However, Kress confirmed that NVIDIA has not sold any H200 chips to Chinese companies: “Although the US government approved limited exports of H200 products to China, we have not received any revenue from these sales, and we do not know if H200 will be allowed into China.”

On January 15th, Foreign Ministry spokesperson Mao Ning held a routine press conference. An AFP reporter asked whether the US President Trump recently stated that the US government would approve the sale of H200 AI chips to China, but would extract 25% of the total sales as a fee. What is China’s response? Mao Ning replied, “China has repeatedly stated its position on US chip exports to China and tariffs.”

NVIDIA has been heavily investing in AI industry companies. During FY2026, it invested $17.5 billion in private companies and infrastructure funds, mainly supporting early-stage startups. The company stated in its financial report that these investments “may not be profitable in the short term, or possibly never,” and that “we cannot guarantee a return on investment.”

Regarding the massive investment plan in OpenAI, NVIDIA noted that “there is no guarantee the deal will be completed.” In September last year, NVIDIA announced plans to gradually invest up to $100 billion in OpenAI, but the agreement has yet to be finalized.

Huang Huang said on the call, “We are continuing efforts to reach a partnership agreement with OpenAI, and we believe we are close to an agreement. We are pleased to maintain a collaborative relationship with this once-in-a-century company and have been working with them since their inception.”

Data center revenue up 75%, with large-scale clients contributing over half of revenue

By business segment, NVIDIA’s data center segment once again hit a record high, accounting for 91% of total revenue. In Q4 FY2026, data center revenue was $62.3 billion, up 75% year-over-year and 22% quarter-over-quarter, beating market expectations of $60.7 billion. For FY2026, total data center revenue increased 68% year-over-year to $193.7 billion.

Kress said that demand from hyperscale cloud providers like Amazon and Microsoft remains strong, with hyperscale data centers “still our largest customer category,” accounting for over 50% of data center revenue.

Kress emphasized, “Since the beginning of this year, analyst expectations for capital expenditure by the top five cloud providers and hyperscale clients in FY2026 have increased by nearly $120 billion, approaching a total of $700 billion. These clients collectively contribute over 50% of our data center revenue. We still expect that the shift of traditional data center workloads to GPU-accelerated computing, along with AI empowering existing hyperscale workloads, will contribute to about half of our long-term opportunities.”

Recently, NVIDIA and Meta announced a multi-year, multi-billion-dollar chip procurement agreement, with Meta deploying millions of NVIDIA chips.

In other segments, gaming and AI PC (Artificial Intelligence Personal Computer) revenue for Q4 reached $3.7 billion, up 47% year-over-year, driven mainly by strong demand for Blackwell; full-year revenue hit a record $16 billion, up 41% from last year.

However, due to strong holiday season demand and natural inventory reduction, this segment declined 13% quarter-over-quarter. Regarding potential impacts from the global memory shortage, Kress said the company expects supply constraints to negatively affect NVIDIA’s gaming business starting in “Q1 FY2027 and beyond.”

Additionally, professional visualization revenue was $1.3 billion, up 159% year-over-year, mainly driven by Blackwell’s strong demand. Automotive and robotics revenue was $604 million, up 6% year-over-year.

FactSet data shows that before the earnings release, out of 74 analysts, 68 rated NVIDIA as “Buy” or “Outperform,” five rated it “Hold,” and one rated it “Sell.” The average target price among analysts is $260.42.

Since the beginning of this year, NVIDIA’s stock has become less volatile. Deepwater Asset Management partner Gene Munster wrote in a blog that the recent disconnect between NVIDIA’s news and its stock performance is due to investors debating whether AI trading is about to end or just beginning — the real debate centers on growth prospects for 2027 and 2028.

(Source: The Paper)