Commentary from Nvidia’s latest earnings shows how it’s helping fuel the market’s fast growth while contributing to industry-wide supply constraints with its own financial muscle in service of rapidly increasing revenue and profits. And the channel stands to benefit.
It’s a situation in which the company is playing a growing role as it increasingly flexes its financial muscle to convert demand into sales with influential buyers such as AI labs, which is opening new opportunities in the channel while contributing to supply shortages.
[Related: Analysis: Does Qualcomm’s AWS Server Chip Deal Signal A Channel Shakeup?]
Nvidia sent these signals back in late August during its second-quarter earnings call, with its CEO and founder, Jensen Huang, saying that the vendor “needs a lot more” supply.
The company had reported a 106 percent year-over-year increase in revenue to $96.2 billion for the period that ended in June, and Nvidia CFO Colette Kress said annual revenue could double next year if it wasn’t for significant supply constraints felt by its suppliers.
“Our entire supply chain is challenged, and it’s everybody. Everybody is really running flat out,” Huang told Wall Street analysts.
While there have been calls for a slowdown in AI development this week by some of the industry’s most influential voices, including OpenAI CEO Sam Altman, due to rising concerns about safety and security, analysts have not yet seen any impact on hardware demand.
“While the governance and safety debate is worth watching, we do not believe it changes the near-term infrastructure outlook,” wrote Brad Gastwirth, global head of research and market intelligence at Southborough, Mass.-based Circular Technology, on Tuesday.
Chris Kapusta, vice president of advisory and transformation at Dallas-based Nvidia systems integration partner GDT, told CRN that he too hasn’t seen any newfound cautiousness from his AI infrastructure customers, which includes roughly 20 neoclouds. (U.K. solution provider giant Softcat announced Thursday that it acquired GDT in a $1.05 billion “enterprise value” deal.)
“I will say we haven’t seen full enterprise adoption at scale of what’s even out there today, so I still think there’s pent-up demand to even bring some of the functionality and capabilities that exist today on the market into a number of enterprise and clients,” he said.
Richard Rudometkin, vice president of global sales at Northbrook, Ill.-based Nvidia systems integration partner International Computer Concepts, said that demand for AI infrastructure from his customers is “nonstop,” even without financial assistance.
“I don’t think that the need or the desire is going to go anywhere,” he said.
Nvidia Forecasts 70 Percent Of Supply-Constrained Growth In 2027
During Nvidia’s second-quarter earnings call last month, the company surprised Wall Street by disclosing its revenue growth forecast for next year: a whopping 70 percent.
This raises the prospect that Nvidia—which financial analysts on average estimate will make $411 billion in revenue for its current fiscal year, according to Yahoo Finance—could reach nearly $700 billion in sales in 2027. Some analysts even think Nvidia could top that.
What was even more remarkable was the company’s claim that it could double revenue next year if it wasn’t constrained by the global production capacity of its suppliers.
“Incredibly, we are seeing demand acceleration, even at our scale,” Kress said on the call. “Customers’ forecasts point to our growth doubling next year. However, as I mentioned earlier, we expect to grow approximately 70 percent as we are supply constrained.”
This suggests that Nvidia could leave demand unfulfilled for more than $100 billion in AI infrastructure gear, which includes its GPUs as well as other kinds of chips it designs, such as CPUs and DPUs, that are becoming more pervasive in its systems. That is a little more than what Nvidia made in revenue for the entire second quarter this year.
“The unconstrained [forecast] would be a lot higher [if it wasn’t for supply issues],” Huang said. “We grew 100 percent year over year this year. The unconstrained is significant, and so we’re just going to have to go work hard to get more capacity.”
For channel partners, supply constraints have translated into long lead times, sometimes stretching out to as long as a year. This has also caused fast-rising prices for systems, and some partners said they are now bracing for price increases from Nvidia itself.
“Nothing feels sustainable about the current supply chain,” said GDT’s Kapusta.
A director-level employee at a large solution provider with a U.S. presence said shortages are impacting AI infrastructure deployments with enterprise customers who don’t want to put their workloads in the cloud for data sovereignty reasons.
“How do I explain to an enterprise that it might be literally eight to 12 months before they even see infrastructure, let alone put it into production?” said the director, who asked to not be named to speak candidly.
Not all GPU-accelerated systems are the same in terms of delivery times however, with Rudometkin at International Computer Concepts noting that some vendors, like Supermicro, carry certain models that can ship on short notice.
“They’ve done a really good job of planning, and they’ve taken [on the] risk. They’ve brought product in, but you know it’s paying off for them because if we have a project and we go to Dell or we go to MSI or [someone else], they’re telling us 12 to 16 weeks, and Supermicro is saying, ‘I can ship two weeks from Friday,’” he said.
Nvidia Is Getting Hit Hard By The Memory Shortage
The biggest supply constraint Nvidia cited is memory chips, showing the magnitude and unwavering nature of a component shortage that has consumed the tech industry for more than a year due to the AI data center boom.
In the earnings call, Kress acknowledged Nvidia’s paradoxical relationship to the memory shortage as the company at the center of the AI infrastructure buildout.
“Memory scarcity today is being driven in large part by the AI buildout itself, and unlike a component that simply raises our cost with no offset benefit, tighter memory supply is a symptom of the same demand surge that’s driving our own growth,” she said.
But Kress explained that the memory shortage isn’t just contributing to a massive supply-demand gap heading into next year. It’s also driving significantly higher prices for such components, much higher than even Nvidia had expected.
“As you are already aware, we are experiencing extreme pricing conditions in memory. The magnitude of the price increase has exceeded our prior expectations and are headed even higher into next year,” she said.
Nvidia is facing these issues despite holding strategic relationships with the industry’s three memory suppliers—Samsung, SK Hynix and Micron—which are working closely with the company to “further increase the capacity our road map requires,” according to Kress.
However, Kapusta of GDT said there have also been industry-wide shortages of SSDs, CPUs and even power supply components. This, according to the solution provider executive, is hurting the ability of enterprise customers to procure hardware.
“It’s just tough to get these components,” he said.
Nvidia Goes Beyond Equity Investments To Support Demand
While Nvidia continues to see high demand from hyperscalers, neoclouds and AI labs, the company is seeing a growing need to keep demand flowing with financial maneuvers that go beyond its go-to strategy of making equity investments.
The moves show how the AI infrastructure giant is helping fuel the market’s rapid growth while contributing to industry-wide supply constraints with its own financial muscle in service of rapidly increasing revenues and profits for Nvidia. And it’s bringing new opportunities to the channel.
With a growing war chest of cash, Nvidia has been ramping up investments over the past few years in companies ranging from publicly traded firms to startups, but the company has more recently started to use its financial muscle to provide support in other ways.
This has mainly come in the form of various backstops Nvidia is providing to lower the risk for investors involved in AI data center projects.
For instance, the company agreed last month to provide a guarantee of up to $105 billion for a massive data center in Ohio that will be leased to OpenAI, with payments by Nvidia set to trigger in case the frontier AI lab defaulted on the lease or failed to make lease payments, according to a filing with the U.S. Securities and Exchange Commission.
The AI infrastructure giant may also put up as much as $125 billion in “residual-value support” for partnerships with several Wall Street heavyweights, including Goldman Sachs and Blackstone, which plan to pool $500 billion in capital for AI infrastructure buying.
To Kapusta of GDT, Nvidia’s financial moves have opened new opportunities for the channel, noting that his company has already been a beneficiary because some of its neocloud customers have received equity investments to help them acquire AI hardware. And he thinks the new alternative funding mechanisms will pave the way for even more deals.
“The neocloud market’s a big market. There’s a lot of room for the channel to play here,” he said. “You can’t run any of this without data center space and power cooling racks, everything else, so this allows other services and product sales and things that go along with all of this that that benefit the channel as well.”
How Nvidia Is Justifying Financial Backstops
The growing moves by Nvidia to finance deals for its customers and partners has attracted heightened scrutiny, with “Big Short” investor Michael Burry, who took a short position in the AI infrastructure giant last year, calling the activity a “Wall Street stunt.”
On Nvidia’s latest earnings call, Kress justified these moves by saying that “frontier labs have extraordinary demand for training and inference compute” but are “growing faster than what their balance sheets and credit profiles can support.”
“They have rapidly growing customer demand yet still lack the decades-long infrastructure contracts and investment-grade financing capacity needed to secure the AI factory infrastructure independently,” she said in August.
But rather than saying these AI labs are limited by their financial capabilities, Kress characterized the situation by saying they are “limited by compute.”
“Nvidia is needed to help power this flywheel,” she said, with “more compute” translating into “more intelligence, more users and more revenue.”
Kress defended the use of Nvidia’s financial horsepower to help customers buy its chips and systems against critics who call such moves “circular financing.”
“We see it differently. We’re going through a major computing platform shift, the creation of one of the most important technologies in human history, and these are once-in-a-generation companies,” she said in prepared remarks on the earnings call. “The technology leadership is proven, and their customer traction and usage are skyrocketing. We expect them to become the largest technology companies in history.”
As a beneficiary of Nvidia’s financing frenzy, GDT’s Kapusta said the moves make “complete sense” because these sorts of deals are meant to facilitate existing demand, not create it.
“What I think Nvidia has quickly realized is the market has to catch up to meet that demand, and with funding for some of these newer neoclouds sometimes being challenging, this is a way to help secure that at a more successful rate and help accelerate the capacity of GPU- and AI-driven infrastructure that’s coming online,” he said.
