Jensen Huang Says Nvidia’s AI Boom Could Keep Surging 70% Next Year

Nvidia CEO Jensen Huang said the chipmaker expects revenue to keep climbing at a stunning pace next year, arguing that the company remains so deeply embedded in artificial intelligence infrastructure that demand for its systems is still far from peaking. Speaking at the Goldman Sachs Communicopia + Technology conference on Thursday, Huang repeated his view that Nvidia can grow revenue by 70% year over year, a projection that would put the company on course for roughly $680 billion in sales if analysts’ current estimates for the present fiscal year hold.

The comments come as investors continue to debate how long Nvidia can sustain its extraordinary run. Competitors are multiplying across the AI stack, from cloud giants designing their own chips to model makers and startups trying to carve out alternatives. Huang’s message was that Nvidia is not just selling processors anymore; it is supplying the full computing backbone of the AI economy, and that position still appears to be expanding.

What Huang said at Goldman Sachs

Huang used the conference to reinforce the same bullish forecast Nvidia first outlined after its most recent record quarter. He told the audience that the company remains confident in its earlier guidance for approximately 70% annual revenue growth next year and suggested the evidence is already visible in customer demand.

One of the clearest signals, he said, is the pace of orders for a high-end Nvidia system built around 36 Grace CPUs and 72 Blackwell GPUs. According to Huang, sales of that product are rising 27% month to month, underscoring how quickly customers are still increasing their AI infrastructure purchases.

Huang said Nvidia believes it can grow revenue by 70% year over year and that the company feels confident about that outlook because of how deeply its hardware and software are embedded across the AI market.

That confidence matters because Nvidia’s latest guidance has been treated as a test of whether the company’s explosive growth can continue after multiple quarters of record-setting results. Analysts currently expect Nvidia to finish its fiscal year with about $400 billion in revenue, which makes the company’s outlook for next year all the more striking.

Why Nvidia says it still has room to run

The simplest answer, Huang argued, is that Nvidia sits at the center of nearly every major AI effort. He described the company as a foundational platform for the industry, saying that model developers, cloud providers, and AI startups all rely on Nvidia’s hardware to train and run their systems.

That includes major names such as Anthropic, OpenAI, and Google, as well as open-weight model developers. In Huang’s view, Nvidia is not merely competing in the AI market; it is helping define the market’s technical base. If AI growth continues to spread across enterprises, startups, and cloud operators, Nvidia stands to capture spending at every layer.

He also pointed to the company’s broader reach, from memory suppliers to data center construction. Nvidia has visibility into the buildout of data centers around the world, including land acquisition, power availability, and shell construction before the buildings are filled with servers and networking gear.

How broad is Nvidia’s footprint?

It is broad enough, Huang suggested, that the company can track where future demand is likely to appear before it fully shows up in revenue numbers. He said Nvidia monitors gigawatts of planned power capacity, physical data center shells, and the activity of cloud operators, original equipment manufacturers, and AI-native startups.

That ecosystem view gives Nvidia an unusually detailed picture of the AI infrastructure pipeline. Unlike many semiconductor suppliers that depend on a narrower set of customers, Nvidia now sells into a market that spans public cloud giants, neocloud providers, enterprise buyers, and venture-backed startups spending heavily on compute.

The competition Nvidia is facing

Nvidia’s rise has not gone unchallenged. The company is facing pressure from nearly every direction in the AI chip market, including hyperscalers such as Amazon, Microsoft, and Google, each of which is investing in custom silicon. Anthropic and OpenAI are also building more of their own infrastructure, while newer rivals such as Cerebras and startups like Etched are attempting to win business with specialized designs.

That competitive backdrop is why Huang’s comments drew attention. The longer Nvidia’s growth persists, the more skeptics ask whether the company’s dominance can survive as customers look for cheaper, more efficient, or more vertically integrated alternatives. For now, Huang’s answer is that demand remains so intense that the market can absorb both Nvidia and its challengers.

He also framed Nvidia’s business in a way that goes beyond chip sales. The company increasingly sells entire systems, networking components, and software support, making it more difficult for rivals to attack a single product category and take share across the full stack.

What exactly is Nvidia selling now?

Nvidia is still known for GPUs, but Huang wants investors to think about the company differently. He joked that the old image of a consumer graphics chip priced at a few hundred dollars no longer reflects what Nvidia builds today. In his telling, the company now sells large-scale computing systems whose value is measured in millions of dollars rather than in single-chip pricing.

His example was a machine composed of dozens of Grace CPUs and Blackwell GPUs linked together with Nvidia’s NVLink networking technology. Huang said a single such system can cost about $8.5 million and includes millions of parts as well as enormous power requirements. That sort of purchase is closer to infrastructure spending than to traditional semiconductor buying.

The shift helps explain why Nvidia’s revenue can grow so quickly. Customers are not just buying chips for one server; they are building large AI clusters that require racks, networking, cooling, land, power, and software integration. In that environment, Nvidia’s role expands from component supplier to platform provider.

How the company’s product mix has changed

At the beginning of its history, Nvidia’s chips were mainly associated with gaming PCs. Today, the company is better understood as a critical supplier to AI training and inference systems. That transformation has made the business much larger and much more central to global technology spending.

• Consumer GPUs helped establish Nvidia’s brand.
• Data center GPUs created the first major wave of AI demand.
• Integrated systems now drive much of the company’s revenue growth.
• Networking and software have become important parts of the offering.

This evolution is one reason Huang believes Nvidia can continue to grow even as the broader AI market matures. The company is no longer dependent on one type of buyer or one product cycle.

What are circular deals, and why do they matter?

Circular financing concerns arise when a company invests in customers or partners that then spend that capital back on its own products. In the history of technology booms, that kind of arrangement has sometimes been seen as inflating demand and delaying a reckoning when real end-user revenue fails to catch up.

Nvidia has been asked whether some of its recent investments in AI companies could resemble that pattern. Huang brushed aside the criticism with a joking response, arguing that the company only makes small investments and then sees much larger amounts of business return in the form of product purchases.

Huang said Nvidia watches the spreadsheets closely and makes investments only when it believes there are real contracts and actual customer revenue behind the company receiving capital.

He added that he has already seen about $100 billion in contracts tied to the ecosystem around these deals and insisted that Nvidia is avoiding speculative risk. The CEO’s core message was that the investments are backed by commercial demand, not just financial engineering.

Why investors are still watching this closely

Even if Huang is right that the contracts are real, the concern is not going away. AI startups are raising enormous sums and spending heavily on cloud computing, chips, and model development. If that spending slows, the growth rate of infrastructure suppliers could cool as well.

In other words, the market is still trying to determine whether AI spending represents the start of a durable new industrial cycle or an overheated phase that will eventually normalize. Nvidia is benefiting from the answer being unresolved.

How much of AI does Nvidia really control?

Nvidia does not control AI itself, but it does occupy a remarkable share of the infrastructure that makes AI possible. Huang’s remarks highlighted the company’s reach across training, deployment, networking, and data center buildouts. That footprint gives Nvidia visibility into future demand and helps it capture spending from customers at multiple stages of adoption.

The company’s advantage is not just technical performance. It also benefits from ecosystem lock-in, software compatibility, developer familiarity, and a head start in scaling large systems. Those factors can make it difficult for even well-funded rivals to dislodge Nvidia quickly.

At the same time, the technology industry has a long history of leaders being displaced faster than they expected. Huang acknowledged that reality in broad terms, even while making the case that Nvidia’s current position is unusually strong. The biggest question is whether customers eventually learn to use AI more efficiently, reducing the amount of hardware and token spending required for each new model or application.

What happens next for Nvidia?

For now, the company is still benefiting from an AI buildout that shows few signs of stopping. Data centers are expanding, model developers keep spending, and cloud providers remain in an arms race to support ever-larger workloads. As long as that continues, Nvidia’s revenue trajectory may remain extraordinary.

But the next stage of the market could look different. As AI products become more mature, companies may focus on reducing costs, improving efficiency, and squeezing more value from each compute dollar. That could eventually reduce the pace of infrastructure purchases, even if overall AI use keeps growing.

Huang’s case on Thursday was essentially that this moment has not arrived yet. In his view, the company is still seeing enough demand across enough parts of the AI economy to justify another year of extreme growth. Whether investors believe that will depend on how long the current spending cycle lasts and how fast competitors can narrow the gap.

Key signals from Huang’s remarks

Several details from the conference stood out as especially important for investors and industry watchers:

1. Nvidia reaffirmed its unusually aggressive 70% revenue growth outlook for next year.
2. Demand for its Blackwell-based systems is still accelerating at the product level.
3. The company says it has visibility into the broader AI infrastructure pipeline worldwide.
4. Management insists its investments in AI companies are tied to real contracts.
5. Competition is intensifying, but Nvidia believes it remains the default platform for AI computing.

Those points together explain why Nvidia remains one of the most closely watched companies in the market. Its growth has become a proxy for the health of the entire AI industry, and Huang’s comments suggest the company believes the sector is still in the early phases of its expansion.

The bigger picture

Nvidia’s story is no longer just about semiconductors. It is about the infrastructure layer of artificial intelligence, where the cost of scaling models has created a new class of system-level spending. Huang’s confidence reflects the company’s belief that it is better positioned than any rival to benefit from that spending, not only because of its chips but because of the ecosystem built around them.

Still, the market’s enthusiasm will continue to meet skepticism. Every technology boom eventually faces questions about durability, efficiency, and competition. For the moment, though, Nvidia appears to have enough visibility, enough customer demand, and enough influence across the AI supply chain to support another very large year.

That is the central takeaway from Huang’s appearance: despite a crowded and fast-changing market, Nvidia believes the AI surge is not fading. If anything, the company is betting that its role in it is still growing.

Huang’s broader argument was simple: Nvidia is not just participating in the AI boom, it is helping power the entire buildout from chips and networking to data centers and model deployment.