Nvidia Supply Concerns Ease, but Long-Term Challenges Remain
Questions persist regarding artificial-intelligence demand and geopolitics
Chip maker Nvidia’s NVDA -2.43%decrease; red down pointing triangle blowout results this past week showed that strong demand will likely carry its business to new heights in the coming quarters—but questions persist about how long the boom will last.
Nvidia eased concerns of a short-lived AI rally by handily beating estimates and giving another glowing outlook, citing an expected boost in supplies. The company still isn’t shipping close to demand, Chief Executive Officer Jensen Huang said in an interview with The Wall Street Journal, suggesting there is more upside even after a pair of record-breaking quarters.
Analysts, though, have raised questions about the supply chain and the long-term strength of the current boom, while tech executives have said they might cut back on the purchases of AI chips if they aren’t translating into new business.
“All major tech companies are going to plow money into this—not just big tech, but also software and services companies and of course the venture [capital] world,” said Dylan Patel of SemiAnalysis, an industry research firm. “The question is, how long will they continue to plow money in?”
In the near term, demand for Nvidia’s products could hardly be hotter. Customers are racing to install chips that underlie artificial-intelligence systems such as OpenAI’s ChatGPT. Companies are increasingly convinced that AI is indispensable for their growth, and analysts estimate that Nvidia has a market share of more than 70% in AI chips.
Like its chips, Nvidia’s shares have been sought-after by investors, having already more than tripled this year, propelling it to a valuation north of $1 trillion. The stock finished higher this past week despite falling later in the week following its earnings report Wednesday.
Analysts at UBS forecast Thursday that Nvidia’s current supply agreements could take its revenue to as much as $25 billion a quarter, not factoring in the expectation that supply will grow. The path to get to around that amount in fairly short order “seems pretty clear to us,” UBS said in a note.
That figure would be a large leap from the $13.5 billion of revenue for its recently completed quarter and the $16 billion it forecast for its current quarter. Just three years ago, Nvidia was averaging around $3 billion in revenue a quarter.
Some analysts still see potential problems ahead in Nvidia’s supply chain. The company designs chips but relies on contract manufacturers—primarily Taiwan Semiconductor Manufacturing Co.—to produce them. That makes Nvidia dependent on others’ ability to increase production at times of high demand.
TSMC has been expanding its capacity, including with a $2.9 billion investment announced in July to expand advanced manufacturing that knits together numerous chips in one package. Nvidia uses that packaging in many of its advanced AI chips, and the company reassured investors Wednesday that it had secured more of it.
As its supplies grow, though, Nvidia might have to fend off companies ordering more than they need, which could mask weaker long-term demand. “For the time being, demand seems unbounded,” Susquehanna International Group analysts said in a note. “However, we worry about double-ordering and rationalization into 2024.”
China, which historically accounts for 20% to 25% of Nvidia’s sales in its data center division, is another wild card. Intent on curtailing Chinese access to advanced AI, the Biden administration last year blocked exports of Nvidia’s most advanced chips without a license. Further restrictions are under consideration, and Chief Financial Officer Colette Kress told analysts Wednesday that a broad ban on sales of AI chips to China would cost the U.S. industry a permanent loss of opportunities in one of the world’s biggest markets.
There is concern about the sustainability of demand in the longer term, too. Many technology transitions in the past have come with a boom of investment in new infrastructure followed by a lull in market uptake.
Big companies spending billions of dollars on AI chips will need to generate profits from them to justify further investments, according to analysts. Some see echoes of the 1990s rollout in the U.S. of fiber-optic cable, which only became commercially viable years later when broadband internet demand rose.
“End customers have to be able to derive business models off of this, whether it’s driving revenue or saving costs,” said Stacy Rasgon, a semiconductor analyst at Bernstein Research. There was some evidence of those models starting to materialize, he said, such as the Copilot AI assistant that Microsoft is testing for its Office productivity software. Microsoft said in July that it would charge $30 a user monthly for the tool.
But given that AI investments for the most part have yet to translate into bumper sales, some consumers of Nvidia’s chips have started to raise concerns about costs. Meta Platforms, the owner of Facebook, plans to spend big on building up its AI computing capabilities into next year, but executives said last month that it wasn’t yet clear how strongly users would adopt AI-driven features.
“The scale of the adoption of those products is ultimately going to inform how much capacity we need,” Susan Li, Meta’s chief financial officer, said on a call with analysts.
Frank Slootman, the chief executive of the cloud-computing company Snowflake, said this past week that a lot of the uses of AI today are experimental and aren’t bringing in money, adding that Nvidia’s chips are expensive.
“We cannot unleash AI and have no business model to pay for it,” he told analysts.
“A lot of the use cases will focus on what are we getting for this,” he said. “This is not just fun and games and planning your next trip to Yellowstone.”