Nvidia falls to extend post-earnings decline to 13% as questions linger over AI spending

• Shares of Nvidia fell 8% on Tuesday, extending their post-earnings decline to 13%.

• Despite Nvidia’s strong earnings, questions remain about the return on investment in AI-enabled GPUs.

• Investor patience is crucial as Nvidia’s data center revenue approaches all-time highs, JPMorgan said.

Shares of Nvidia fell as much as 8% on Monday, extending its subsequent earnings decline to 13%, as investor questions persist about the return on investment from the companies’ huge artificial intelligence spending.

According to JPMorgan’s President of Market and Investment Strategy, Michael Cembalest, Nvidia’s Q2 earnings results were extraordinary.

“Phenomenal and the antithesis of the dot-com era,” Cembalest said in a note Monday. “NVIDIA is not like dot-com leaders like Cisco, whose P/E multiple has also increased but without corresponding earnings.”

But while Nvidia has benefited enormously from technological advances in its GPUs, there are still lingering questions about how much its customers will benefit financially from the hundreds of billions of dollars they’re spending on GPU chips with AI.

If profits don’t come quickly for cloud hyperscalers, spending on Nvidia’s core product could eventually fall.

“Over the next two years, corporate AI adoption trends … must increase to avoid a ‘metaverse’ outcome for all capital deployed,” Cembalest said, referencing the chart below.

A recent report from The Information suggests that OpenAI could lose $5 billion this year, Anthropic’s CEO said it will cost $100 billion to train a single AI model in 2027, and the biggest companies tech companies could have $500 billion in “missing revenue” that they were supposed to pay for their massive AI investments, Cembalest pointed out.

“Barclays estimates that in 2024, enough GPUs have been built to generate ~$100 billion in revenue at peak usage rates. Actual payments in 2024 by GPU users: ~$10 billion. How long will it take for this gap to close?” asked Cembalest.

To be sure, each new computing platform required massive infrastructure investment before monetization strategies took off.

However, the ultimate question is how patient investors will be — and that patience, or lack thereof, will be directly reflected in Nvidia’s stock price.

“Every computing cycle involves first infrastructure, then platforms, and then applications. It is probably too early to worry that there is no killer generative AI application yet similar to the enterprise resource planning software of the 1990s or the search and e-commerce applications of the 2000s,” said Cembalest.

The story continues

He added: “But the clock is ticking: NVIDIA’s data center revenue as a share of market-wide capex will reach levels seen at the peak of the mainframe era in 1969 and the dot-com boom. As a result, the stakes for investors are high.”

For his part, Nvidia CEO Jensen Huang addressed the ROI question in his company’s earnings call last week.

But instead of pointing to a killer app that generates tons of revenue for his customers, Huang pointed to the huge cost savings his customers experience due to the immense computing power of his accelerated GPUs compared to traditional CPUs.

“The amount of computing demand continues to grow quite significantly. Maybe you could even estimate that it doubles every year. And so if we don’t have a new approach, computing inflation would drive up costs for every company and drive up the power consumption of data centers around the world,” Huang explained.

He added: “As they build the Hopper-based infrastructure and soon the Blackwell-based infrastructure, they start to save money. This is a tremendous return on investment. And the reason they’re starting to save money is because data processing saves money.”

Read the original article on Business Insider