NVIDIA CEO Jensen Huang speaks at the COMPUTEX Forum event in Taipei, Taiwan on June 4, 2024.
Ann Wang | Reuters
The past two years have been fun for Nvidia investors, but things have been a rollercoaster lately.
Nvidia, the biggest beneficiary of the artificial intelligence boom, has seen its market capitalization expand about ninefold since the end of 2022. But after hitting an all-time high in June and briefly becoming the world's most valuable publicly traded company, Nvidia lost nearly 30% of its market capitalization in the seven weeks since, shedding about $800 billion in market value.
The stock is currently in an uptrend, down about 7% from its all-time high.
Chipmakers are scheduled to report quarterly results on Wednesday, and stock price movements are a top focus on Wall Street: Any signs that demand for AI is waning or that big cloud customers are cutting back a bit on costs could lead to a big drop in revenue.
“They're the most important stocks in the world right now,” EMJ Capital's Eric Jackson said on CNBC's “Closing Bell” last week. “If they fail, it's going to be a big problem for the whole market. I think they're going to unexpectedly rise.”
Nvidia's earnings report comes just weeks after other big tech companies reported results, and the company's name was ubiquitous on analyst calls, with Microsoft, Alphabet, Meta, Amazon and Tesla all investing heavily in Nvidia's graphics processing units (GPUs) to train AI models and handle massive workloads.
Nvidia's revenue has more than tripled year over year over the past three quarters, with most of the growth coming from its data center business.
LSEG said analysts are expecting a fourth consecutive quarter of triple-digit growth, but that growth rate is expected to slow 112% to $28.7 billion, with year-on-year comparisons becoming tougher from here on out, and growth is expected to slow in each of the next six quarters.
Investors will be particularly interested in Nvidia's outlook for the October quarter, when the company is expected to show growth of about 75% to $31.7 billion. Optimistic guidance suggests Nvidia's deep-pocketed customers continue to show a willingness to loosen their purse strings for AI builds, while a disappointing outlook could raise concerns that infrastructure spending is overspending.
“Given the rapid growth in hyperscale capital spending over the past 18 months and the strength of the near-term outlook, investors are often left questioning the sustainability of the current capital spending trajectory,” Goldman Sachs analysts, who recommend buying the stock, wrote in a note last month.
Much of the optimism ahead of the report was driven by comments from major customers about how much they continue to pay for data centers and NVIDIA-based infrastructure. The company's shares have risen 8% in August.
Last month, the CEOs of Google and Meta enthusiastically endorsed the pace of their companies' expansion, saying underinvestment was a bigger risk than overinvestment. Former Google CEO Eric Schmidt recently told Stanford University students in a since-deleted video that he was hearing from big tech companies “we need $20 billion, $50 billion, $100 billion worth of processors.”
But while Nvidia's profit margins have expanded recently, the company still faces questions about the long-term return on investment it gets from customers buying devices in bulk, costing tens of thousands of dollars each.
During NVIDIA's most recent earnings call in May, CFO Colette Kress presented data points suggesting that cloud providers, which account for more than 40% of NVIDIA's revenue, will generate $5 in revenue for every $1 spent on NVIDIA chips over the next four years.
More such statistics could be forthcoming: After speaking with Mr. Kress, a Goldman analyst wrote last month that the firm would release more ROI metrics this quarter “to instil confidence among investors.”
Blackwell Timing
Jensen Huang, co-founder and CEO of Nvidia Corp., showed off the new Blackwell GPU chip at the Nvidia GPU Technology Conference on March 18, 2024.
David Paul Morris/Bloomberg via Getty Images
Another big issue facing Nvidia is the schedule for its next-generation AI chip, called Blackwell. The Information reported earlier this month that the company was facing production issues and that large-scale shipments would likely be delayed until the first quarter of 2025. Nvidia said at the time that production was scheduled to ramp up later this year.
The report comes after Nvidia CEO Jensen Huang surprised investors and analysts in May by saying the company expected Blackwell's revenue to grow “significantly” this fiscal year.
Nvidia's current generation chip, “Hopper,” remains the top choice for deploying AI applications like ChatGPT, but competition from Advanced Micro Devices, Google, and a handful of startups is emerging, putting pressure on Nvidia to maintain its performance advantage through a smooth upgrade cycle.
Any potential Blackwell delays would likely push that revenue out to a future quarter, while current Hopper sales would grow, especially sales of the new H200 chip. The first Hopper chips are due to be in full production in September 2022.
“The timing change is less of an issue as supply and customer demand have pivoted rapidly to H200,” Morgan Stanley analysts wrote in a note this week.
Many of Nvidia's major customers have said they need the extra processing power of the Blackwell chips to train their next-generation, more advanced AI models, but they're willing to take what they can get.
“We expect Nvidia to reduce its Blackwell B100/B200 GPU allocation in the second half of the year and focus on ramping up Hopper H200,” HSBC analyst Frank Li wrote in an August note. He has a buy recommendation on the stock.
Correction: Colette Kress is Nvidia's CFO. An earlier version misspelled her name.
