Glonce

Nvidia's Jensen Huang reiterates 70% revenue growth guidance

businessThe Glonce editorsTechCrunch🔥65Glonce rating65Impact65Novelty40Relevance95Surprise35Players90Credibility80
Nvidia's Jensen Huang reiterates 70% revenue growth guidance

Nvidia founder and CEO Jensen Huang spoke at the Goldman Sachs Communacopia + Technology conference on Thursday, September 10, addressing persistent doubts about whether the company's run of AI-driven growth can continue as competition for GPUs and AI chips builds from several directions at once: hyperscalers Amazon, Microsoft and Google are each building their own chips, so are AI labs Anthropic and OpenAI, and newly public Cerebras and startups such as Etched are chasing the same demand. Huang used the appearance to repeat the guidance Nvidia first gave last month, on the earnings call where it reported yet another record quarter: he said the company could grow revenue about 70% year over year next fiscal year, and that he is confident in that number. Analysts expect Nvidia to close its current fiscal year with about $400 billion in revenue, so 70% growth would put next year's total at around $680 billion.

Huang's case for that confidence is that Nvidia sits inside nearly every part of the AI industry. “Nvidia runs every model. Every single lab can use us,” he said, naming Anthropic, OpenAI and Google alongside open-weight models. He called the company a “foundational platform of the AI ecosystem, foundational platform of the AI industry,” and said Nvidia tracks “every single gigawatt of land, power, shell around the world” through data its neocloud, OEM, cloud and AI-native company partners report back. To show how far Nvidia's hardware has scaled since the days when GPUs were sold mainly for PC gaming, he said what he still calls “one GPU” now costs $8.5 million rather than $399: an NVLink-connected system built from 2 million parts that draws 250,000 kilowatts, and Nvidia ships thousands of them. Separately, he said a system pairing 36 Grace CPUs with 72 Blackwell GPUs is seeing 27% month-to-month sales growth; it is not clear from his remarks whether that is the same system as the $8.5 million one or a different product.

Huang also addressed criticism of Nvidia's so-called circular deals, in which the company invests in AI businesses that then buy its hardware; the article notes that similar schemes famously contributed to the downfall of a previous generation of internet build-out suppliers, such as Lucent Technologies. His response was blunt: “Well, it’s not circular because we put a little bit of money in, and a lot of money comes back.” He offered a hypothetical ratio to make the point: “I look at the spreadsheet, we put in $1 and $100 comes back in. Is that circular? If that is, let’s do more of that.” Huang insisted Nvidia only invests once it has confirmed a company already holds real, revenue-generating contracts, saying he has personally seen $100 billion worth of such contracts and adding, “I’m not taking any risks. … I need a sure thing.”

The article closes by noting that much of AI's current growth is coming from AI-native startups that raise large sums and spend most of that money on their own AI use, a pattern Huang himself acknowledges, and by observing, in its own words, that big things in the tech industry tend eventually to be disrupted; it expects infrastructure and token use to become more efficient as the industry matures. No calendar date is given for when Nvidia's current fiscal year ends, and the $400 billion current-year estimate is attributed only to unnamed analysts, not to Huang.

Key facts

  • At the Goldman Sachs Communacopia + Technology conference on Thursday, September 10, Nvidia CEO Jensen Huang repeated the guidance he first gave last month: revenue could grow about 70% year over year next fiscal year.
  • Analysts expect Nvidia to close its current fiscal year with about $400 billion in revenue; 70% growth would put next year's total at around $680 billion.
  • Huang said what he calls “one GPU” now costs $8.5 million rather than $399, describing an NVLink-connected system built from 2 million parts that draws 250,000 kilowatts; separately, a system pairing 36 Grace CPUs with 72 Blackwell GPUs is seeing 27% month-to-month sales growth.
  • He defended Nvidia's disputed “circular deals,” in which it invests in companies that buy its hardware, saying “we put a little bit of money in, and a lot of money comes back” and that he has seen $100 billion worth of real revenue-generating contracts before any such investment.
  • Huang argued Nvidia is a “foundational platform” for the whole AI industry, saying every major lab, including Anthropic, OpenAI and Google plus open-weight model makers, runs on its hardware, even as some of those same companies build competing chips of their own.

Why it matters

Huang's remarks are not new guidance: they are the same 70% growth figure Nvidia already gave investors last month, restated in public a month later and under closer scrutiny. The repetition matters because it comes as competition for AI chips intensifies from every side, hyperscalers building their own silicon, AI labs doing the same, and both a newly public rival in Cerebras and startups such as Etched chasing the same demand. Huang's underlying argument, that Nvidia sits inside essentially every model any lab runs, is his stated reason for expecting growth to hold regardless of that competition. It also functions as his direct answer to questions about Nvidia's circular deals, the practice of investing in AI companies that then buy Nvidia hardware, which the article links to the kind of financing arrangements blamed for the collapse of earlier telecom-era suppliers like Lucent Technologies.

Who it affects

Investors and analysts weighing Nvidia's stock get a direct, on-the-record restatement of the number the company is asking them to trust, plus new detail on pricing and system-level sales growth. Nvidia's competitors, the hyperscalers building their own chips, the AI labs doing the same, and rivals such as Cerebras and Etched, are the implicit target of Huang's dominance argument. Nvidia's own ecosystem, the neoclouds, OEMs, cloud providers and AI-native companies he says report data back to Nvidia, and the AI labs he names, Anthropic, OpenAI and Google, plus open-weight developers, are cited as the evidence for that argument. Anyone tracking so-called circular deals in AI, where a chipmaker's investment in a customer becomes revenue for itself, gets Huang's own defense of the practice, backed by his $100 billion contracts claim.

How to use it

There is no product or price list here for a reader to act on directly, but the numbers are useful as a reference point. Huang's own implied math gives about $680 billion in next-fiscal-year revenue if the roughly $400 billion current-year estimate and the 70% growth guidance both hold, a figure worth checking Nvidia's actual results against once they are reported. The system-level figures, an $8.5 million, 2-million-part, 250,000-kilowatt NVLink system on one hand, and a separate system pairing 36 Grace CPUs with 72 Blackwell GPUs growing sales 27% a month on the other, give a rough scale reference for anyone budgeting large AI infrastructure purchases, though the source does not say whether the two descriptions refer to the same product.

How solid is it

Every number here traces to one on-the-record appearance by Huang, at a named conference, rather than to a filing or an independent analyst report. The 70% growth figure is guidance Nvidia already disclosed last month on its earnings call, so this is a reaffirmation, not a new data point. The $400 billion current-fiscal-year estimate is attributed only to unnamed analysts, without a firm or report named. The $100 billion in contracts Huang says back Nvidia's investments, and the $1-in-$100-out ratio he uses to describe them, are his own assertions, not independently verified in the source. The article also never confirms whether the $8.5 million GPU system and the 36-Grace-CPU/72-Blackwell-GPU system are the same product or two different ones, so that detail should be read as two separate data points.

Risks and caveats

Huang's own wording is that Nvidia “could” grow 70% and that he is “confident,” not that it will happen; the implied $680 billion figure is the article's own calculation from an analyst estimate, not a number Huang stated himself. The circular deals Huang defends, where Nvidia's investment in a company becomes that company's spending on Nvidia hardware, are explicitly compared in the article to the arrangements blamed for the collapse of an earlier telecom-era supplier, Lucent Technologies. Huang himself acknowledges that much of AI's current growth comes from AI-native startups spending most of their raised capital on their own AI infrastructure, a pattern that depends on continued fundraising rather than on end-customer revenue. The article's own closing view is that big technology dominance tends eventually to be disrupted, and that infrastructure and token use should become more efficient as the industry matures, both of which cut against the demand growth Nvidia's guidance assumes.

“I think we could grow 70% year over year. We’re confident about that.”

— Jensen Huang