SANTA CLARA, California — Nvidia delivered another blockbuster quarter and offered one of its strongest indications yet that the artificial intelligence spending boom is far from over, forecasting revenue growth of about 70% in its next fiscal year as demand for AI computing continues to outstrip the company’s ability to supply it.
The chipmaker reported $96.2 billion in revenue for its fiscal second quarter, up 18% from the prior quarter and 106% from a year earlier. It forecasts $108 billion in revenue for the current quarter, above Wall Street’s average estimate of $104.19 billion.
More strikingly, Nvidia said it expects revenue to grow by about 70% in fiscal 2028, a rare long-range forecast from a company that typically provides guidance only for the coming quarter. Analysts had expected roughly 44% growth for the period.
The forecast comes as investors question whether the multiyear surge in spending on data centers, advanced chips and generative AI can continue. Nvidia’s answer is clear: it expects demand to remain exceptionally strong, supported not only by the largest cloud companies but by AI labs, startups, governments, industrial companies and a new category of AI-focused cloud providers.
“AI has reached its inflection point. It’s doing useful work. Its tokens are productive and profitable. Now, compute is revenue,” Nvidia Chief Executive Jensen Huang said in the company’s earnings release. “Demand is accelerating.”
Nvidia’s shares rose nearly 5% in extended trading after the results, recovering from an initial decline as investors absorbed the company’s larger-than-expected outlook.
A quarter that more than doubled sales
Nvidia’s quarterly revenue of $96.22 billion beat analysts’ estimate of $92.17 billion, according to LSEG data cited by Reuters. Adjusted earnings were $2.22 a share, above the $2.10 Wall Street forecast.
Net income under generally accepted accounting principles rose to $59.69 billion, up 126% from $26.42 billion a year earlier. GAAP diluted earnings per share were $2.46, compared with $1.08 a year earlier.
The company’s gross margin was 75%, up from 72.4% in the same quarter a year ago. Operating income reached $63.73 billion, a 124% increase from the prior-year period.
The numbers are extraordinary even by Nvidia’s recent standards. Before the generative AI boom, the company was primarily known for graphics processing units used in gaming, professional visualization, and specialized computing. It is now one of the world’s most important suppliers of computing infrastructure that powers large language models, AI image generation, robotics, data analysis, and autonomous systems.
The latest quarter shows that the company’s growth remains concentrated in its data-center business, but it also highlights its widening reach across cloud infrastructure, AI software, networking, robotics and high-performance computing.
| Metric | Fiscal Q2 2027 | Year-over-year change |
|---|---|---|
| Revenue | $96.22 billion | +106% |
| Data Center revenue | $89.0 billion | +117% |
| GAAP net income | $59.69 billion | +126% |
| GAAP diluted EPS | $2.46 | +128% |
| Non-GAAP diluted EPS | $2.22 | +120% |
| GAAP gross margin | 75.0% | +2.6 percentage points |
Data centers drive the boom
Nvidia’s Data Center division produced $89 billion in quarterly revenue, up 18% from the previous quarter and 117% from a year earlier. That unit alone accounted for about 92% of the company’s total revenue in the period.
The business supplies graphics processors, central processors, networking equipment, and software used to train and run AI models.
The biggest buyers include cloud providers such as Amazon Web Services, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure. But Nvidia said demand is broadening beyond the handful of large technology companies that dominated the early stages of the AI buildout.
Huang said that a year ago, “one lab alone” was driving much of the growth in AI infrastructure. Now, he said, the market includes multiple frontier AI labs, startups, an expanding open-source model ecosystem and applications in physical AI, including robotics and autonomous systems.
Nvidia expects AI labs to represent about one-quarter of its total business next year, Reuters reported.
The company also said so-called neo-cloud firms, including CoreWeave and Nebius, are expected to finish the year with more than 8 gigawatts of Nvidia GPU capacity, compared with 3 gigawatts at the end of last year.
These companies lease AI computing capacity to enterprises and developers that may not want to build their own massive data centers. Their growth is important to Nvidia because it expands the customer base beyond established cloud giants.
Huang described this broader category as including regional AI firms, startups, enterprises and specialized cloud providers. CNBC reported that he referred to the group as “ACIE,” saying it could eventually become larger than the traditional cloud computing market.
Vera Rubin moves into production
Nvidia’s next-generation Vera Rubin platform has entered full production and will represent about one-fifth of Data Center revenue in the current quarter, which ends in October.
The platform combines the Vera CPU with Nvidia’s Rubin graphics and networking technologies. The company says it is designed for the next stage of AI computing, especially AI agents that can perform multistep tasks, make decisions, and interact with software tools.
Nvidia said rack-scale Vera Rubin systems are operating with partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius.
The company also announced Vera, which it described as its first CPU built for AI agents.
The rollout matters because Nvidia’s investors are watching whether the company can successfully transition customers from one generation of AI hardware to the next. Each major platform cycle presents both an opportunity and a risk: the new product can increase sales and performance, but manufacturing complexity and supply constraints can delay deliveries.
So far, Nvidia is signaling confidence.
“We expect Vera Rubin to mark the fastest product ramp in Nvidia’s history,” Chief Financial Officer Colette Kress said, according to reporting by Investopedia.
The company is also expanding its product range beyond chips. It is selling complete systems, networking technology, software libraries and reference designs for what it calls AI factories, large-scale computing facilities designed to produce AI models and services.
That broader strategy helps Nvidia capture more of the value from AI infrastructure spending. It also makes customers more dependent on Nvidia’s technology ecosystem.
AWS partnership adds 2 million GPUs
Nvidia announced an expansion of its partnership with Amazon Web Services that will add 2 million more Nvidia graphics processors across AWS infrastructure in 2027 and 2028.
The scale of the agreement illustrates the enormous capital investment now flowing into AI computing.
Cloud providers are buying advanced chips not only to run their own AI services but also to rent computing capacity to companies building AI products. AI models require enormous amounts of processing power to train and, increasingly, to serve responses to users in real time.
The AWS deal is one of several signs that the largest technology companies remain willing to make multiyear commitments to AI infrastructure.
Nvidia has also worked to make financing more available for AI data-center projects. During the quarter, it announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to develop independent computing-finance platforms that could mobilize more than $500 billion in third-party capital over time, subject to definitive agreements.
Such financing arrangements could help AI labs, cloud providers, and enterprises fund expensive data-center projects. They could also deepen concerns among some investors about whether the industry is building more capacity than end users will ultimately need.
Nvidia’s management argues that AI computing is moving from experimentation into productive use, creating a stronger economic basis for continued spending.
Supply constraints remain a risk
Despite its bullish forecast, Nvidia warned that shortages of memory components and rising costs will limit how quickly it can expand.
Kress said the company is supply-constrained, and that memory shortages will continue to pressure margins.
Nvidia expects third-quarter gross margins of about 74%, plus or minus 50 basis points. It expects margins to fall further, reaching a low point of approximately 71% to 72% in the fourth quarter, before recovering.
That is still an unusually high margin for a hardware company. But the decline matters because investors have become accustomed to Nvidia’s exceptional profitability.
The challenge is partly driven by the enormous demand for high-bandwidth memory, a specialized type of memory used alongside advanced AI chips. AI systems need vast amounts of memory to handle increasingly large models and datasets.
Nvidia has said it is working with suppliers to secure components and expand capacity, but the global supply chain is under pressure.
“Our demand is much greater than 70%,” Huang said during the earnings call, according to CNBC. “Our supply allows us to confidently deliver 70%.”
That statement captures both the opportunity and the constraint. Nvidia believes customer demand could support growth even higher than the company’s forecast, but it cannot guarantee that enough components, manufacturing capacity, power, and data-center space will be available.
China revenue remains uncertain
Nvidia did not include any China Data Center compute revenue in its third-quarter outlook.
China has been a source of uncertainty for the company because U.S. export restrictions have limited the sale of Nvidia’s most advanced chips to Chinese customers.
In May, Washington cleared about 10 Chinese companies, including Alibaba, Tencent and ByteDance, to buy Nvidia’s H200 AI chip, Reuters reported. But deliveries were delayed, and the company’s sales outlook remains unclear.
The lack of China revenue in the forecast is a conservative move. It reduces the risk that Nvidia will miss its guidance if shipments remain constrained by export controls or regulatory delays.
At the same time, it demonstrates that Nvidia’s growth is being driven by demand outside China. The company has customers across the United States, Europe, Japan, South Korea and other regions that are building national, corporate and commercial AI infrastructure.
Nvidia announced partnerships to support AI infrastructure projects in South Korea and Japan, as well as 35 new AI high-performance computing supercomputers under development across Europe.
Global expansion is strategically important. Governments increasingly see AI infrastructure as an economic and national-security priority, and they are investing in what is often called sovereign AI, domestic computing capacity that reduces dependence on foreign providers.
A revenue trajectory without precedent
Nvidia’s 70% fiscal 2028 growth forecast surprised investors because it was far above consensus expectations and because the company rarely gives a detailed annual growth outlook so far in advance.
Based on Wall Street’s consensus forecast of $396 billion in fiscal 2027 revenue, a 70% increase would imply roughly $673 billion in fiscal 2028 sales, CNBC calculated.
That would put Nvidia ahead of Apple and Alphabet in revenue based on current Wall Street projections and behind only Amazon among major U.S. technology companies, CNBC said.
Those calculations are not Nvidia’s official full-year revenue forecast. They are an illustration of what the company’s stated growth rate could mean if the fiscal 2027 consensus proves accurate.
Still, they show the scale of the shift. Nvidia has gone from a specialized semiconductor company to a central supplier for an industry spending hundreds of billions of dollars on AI data centers.
Its rise has also made it a bellwether for the broader technology market. A strong Nvidia forecast can boost confidence that AI investment is producing real demand. A weaker forecast could trigger concerns that the spending cycle is peaking.
The latest results offered the former message.
What investors will watch next
The company’s forecast has strengthened the case that AI infrastructure investment will continue into 2027 and 2028. But it has not ended the debate over how durable the boom will be.
Investors will watch several issues:
- Whether Vera Rubin ramps smoothly and meets performance and supply expectations.
- Whether memory shortages become more severe or begin to ease.
- How quickly Nvidia’s gross margins recover after the expected late-year decline.
- Whether cloud providers and AI labs continue to spend at the same pace.
- Whether new customers, including enterprises and governments, meaningfully diversify Nvidia’s sales.
- Whether China sales resume and contribute to future growth.
- Whether AI companies can generate enough revenue from products and services to justify the industry’s capital spending.
Nvidia’s current-quarter revenue forecast of $108 billion suggests that demand remains intense.
The company’s longer-range outlook is even more consequential. By forecasting 70% growth in fiscal 2028, Nvidia is making a public bet that AI computing is not a short-lived technology cycle.
It is saying the buildout has moved beyond hype, beyond a handful of early adopters and into a global competition to build the infrastructure behind a new generation of software, services and machines.
Whether that bet proves right will shape not only Nvidia’s future, but the trajectory of the entire AI economy.
