Nvidia Beats Q1 With $81.6B Revenue, Q2 Guided to $91B

Nvidia (NASDAQ: NVDA) beat analyst expectations on every key measure in its fiscal Q1 2027 report released on Wednesday. The company also gave Q2 guidance that was nearly $4 billion higher than what the Street thought it would be, which is the clearest sign yet that spending on AI infrastructure is not slowing down.

The company also said it would buy back $80 billion worth of its own shares and raised the payment it pays every three months from $0.01 to $0.25 per share. After hours, shares went up and down a lot, but they ended the day slightly down.

This response doesn't say much about the quarter itself, but it does show how much good news is already built into a stock that has grown to a $5 trillion market value.

The Numbers

Expected EPS was $1.77, but it was $1.87. Instead of the expected $79.19 billion, they made $81.62 billion. Adjusted gross margin: 75%, which is the same from one quarter to the next but up 14.2 basis points year-over-year, which is what was expected. Analysts expected sales to be $87.36 billion in Q2, but the company expects sales of $91 billion, plus or minus 2%.

The success was driven by the data center segment. Analysts had predicted that Nvidia's Data Center would make $73.47 billion, but instead it made a record $75.25 billion, a 92% increase year-over-year. In the same quarter last year, the business made $39.11 billion. That means that the company has grown by two times in twelve months.

Half of the data center's income came from hyperscaler customers, and the other half came from AI clouds, industrial, enterprise, and sovereign customers. This spread of customers lowers the risk that came with having just one big customer like in previous quarters.

Jensen Huang's Framing — and the Reporting Restructure

CEO Jensen Huang talked about the quarter in a way that goes beyond quarterly results: "The construction of AI factories is speeding up at an incredible rate. This is the largest expansion of infrastructure in human history." Agentic AI has arrived and is now doing useful work, creating real value, and quickly spreading across businesses and fields.

Nvidia changed its reporting structure at the same time, reducing the number of market platforms from five to two: Data Center and Edge Computing.

Hyperscale includes public clouds and big internet service providers for consumers. The new "ACIE" sub-market includes AI clouds, industrial, and enterprise use cases. PCs, game consoles, workstations, AI-RAN base stations, robotics, and cars are all examples of products that can use agentic and physical AI.

The reform is important. That means Nvidia sees itself as a two-platform business in the future: the data center, where intelligence is made, and the edge, where intelligence is used. The smaller category of Edge Computing will be the one to watch for the next stage of Nvidia's growth as agentic AI goes from the cloud to the device.

The China Absence: $4.6 Billion Missing From the Comparison

There is a number in the quarter that needs special attention. In Q1, no Data Center Hopper products were sent to China. In the same quarter last year, $4.6 billion worth of products were sent there. Analysts didn't see that gap in revenue because they had already set their plans for China Data Center compute to zero. The estimate for Q2 also assumes that the Data Center will not make any money from China.

Huang, the CEO of Nvidia, went with President Trump to China last week, which led to rumours that chip export restrictions might be eased. There was no big breakthrough. The company is functioning with one of the world's biggest technology markets, which used to bring in a lot of money, fully cut off from its target audience.

Nvidia is seeing 92% year-over-year growth in its data centers, even though China is a $4.6 billion headwind. This shows how strong demand is in other parts of the world.

U.S. export controls and China's active push to create its own AI hardware alternatives, including hardware from Huawei and others, have changed how Nvidia interacts with China. This isn't a temporary problem; it's a lasting shift in the market that Nvidia is more than making up for by growing in other areas.

The Competitive Landscape

The results landed in a week when the competitive narrative around Nvidia intensified from multiple directions. Cerebras held its IPO last Thursday, bringing a direct inference competitor to public markets.

Amazon disclosed that its Trainium chip segment now has an annual revenue run rate exceeding $20 billion with triple-digit year-over-year growth, and announced a 2 gigawatt capacity deal with OpenAI and a 5 gigawatt commitment from Anthropic for Trainium chips. Google unveiled its TPU 8i and TPU 8t at Google I/O — the former for AI inference, the latter for model training — alongside a multigeneration, multigigawatt Anthropic TPU commitment.

These are real changes in the economic landscape, not just hypothetical threats. But Steve Sosnick, chief analyst at Interactive Brokers, said what the market really thought: "It's quite something to yawn at 75% gross margins." As the only platform that runs in every cloud and powers every frontier model, Nvidia's ecosystem continues to produce a margin profile that custom silicon competitors have not yet reached on a large scale.

Dan Ives of Wedbush explained the bigger picture on X: "Another strong and impressive quarter from Nvidia." Demand keeps going up faster than anyone on Wall Street thought it would. Hugely important print and guide for Jensen, who is known as the "Father of AI." It will have an effect on the whole tech world.

Most people expected Q2 to bring in $87.36 billion, so the $91 billion range is the number that means the most for positioning. It means that Nvidia's income is still growing faster than the previous quarter, even though the previous quarter was already a record.

At current prices, the $80 billion return supports growth in the value of each share. The 25-cent payment shows that the company is confident that free cash flow will last. The China absence is already included in most predictions, so it doesn't add to the downside.

Amazon Trainium, Google TPU, and Cerebras are all real competitors, but Nvidia's numbers show that they are not yet cutting into their profit margins. The stock's lacklustre response after hours shows how much good news was already priced in. This is not a sign that the quarter was disappointing.