Nvidia’s claims of huge market possibilities in robotics and self-driving cars could “comfort investors who are thinking about what might come next for the company.”
Shares of Nvidia Corp. went down on Wednesday, adding to the big drops they had the day before, even though analysts told investors to look at the big picture.
Wall Street wasn’t impressed by CEO Jensen Huang’s keynote speech on Monday night, but executives had another chance at them at an analyst meeting on Tuesday afternoon. The meeting covered a wide range of topics, from trying to get more information on the specifics of guidance for the January quarter to management’s views on opportunities in areas such as robotics and self-driving cars.
The price of Nvidia stock (NVDA) is down about 1% today, after going up as much as 2.7% earlier in the day. On Tuesday, it dropped 6.2%.
Analysts saw some good signs, especially when they used a wide lens. “While we understand the near-term concerns around constraints limiting upside into the April quarter, we believe this is minutia compared to the real question – ‘how big can this get?'” Ben Reitzes, an expert at Melius Research, wrote.
He said Huang’s words back up the idea that there is a huge chance ahead. People in charge at Nvidia, for example, have been talking about how $1 trillion worth of traditional data center infrastructure is being moved to accelerated computing, which is what the company’s technology powers. But Huang also showed some math on Tuesday about self-driving cars, which could open up another $2 trillion to $3 trillion in market opportunities.
“Then he went into how it was going to be the same for robots – and there would be more robots than humans,” he wrote. “So, before long—you get the point—each major application spent trillions of dollars building data centers that would run AI for training, inference, and taking in both fake and real data, with each application being bigger than the last.” It was clear from his answer that Nvidia wants to keep growing and that its CEO believes its “total addressable market” (TAM) is worth many trillions of dollars.
Bernstein’s Stacy Rasgon also talked about what “physical” AI could do.
“It is not hard to imagine applications like these as further extending Nvidia’s runway far into the future,” he wrote, pointing out that the company would have to spend an extra $2 billion to $3 billion on data centers for every million or so self-driving cars that are made, not to mention the potential for humanoid robots to grow.
That idea is “potentially comforting for investors thinking about what might come next for the company.”
Huang also had something to say about the argument between graphics processing units and application-specific integrated circuits, such as those made by Broadcom Inc. (AVGO). Broadcom made a positive long-term market forecast last year, which made experts and investors wonder if ASICs might be able to take market share from GPUs, in part because they are cheaper.
“Given the dynamism apparent in the space Jensen (while admittedly talking his book) clearly views the company’s GPU-based programmable platform approach as superior to ASICs given one doesn’t necessarily know what the next killer application might be, hence advantaging a flexible platform” Rasgon said.
“There was not much incremental on the data-center business,” which is Nvidia’s huge, AI-powered engine right now, Rasgon said in Monday’s address. A consumer electronics show, CES, is “not really the forum for that anyway,” says Rasgon. Also, Nvidia’s GTC event will happen in March.
