Unlock Editor’s Digest for free
FT editor Roula Khalaf selects her favorite stories in this weekly newsletter.
No company has gone from a $1 trillion stock value to $2 trillion as quickly as Nvidia. The chip designer doubled its market capitalization in less than nine months. The rise of generative artificial intelligence means that demand for Nvidia chips still outstrips supply. But reaching the $3 trillion milestone is proving more difficult.
For some investors, the AI hype has inflated valuations beyond reason. Michael Burry, whose short against the real estate market became famous in The big short, bet against semiconductor stocks last year. Cathie Wood’s Ark Invest exited Nvidia in early 2023, claiming the market was pricing in too much optimism.
However, if you compare Nvidia’s position with the company’s own market history, its valuation does not seem exaggerated. The stock trades at 35 times expected earnings, up from 55 times in early 2022. Given the size of its potential market and the pricing power it wields, Nvidia has a chance to become the most valuable stock in USA
Those concerned about Nvidia’s valuation point to advances made by Google’s TPU, AMD’s MI300X and other potential rivals. But Nvidia has two advantages: an advantage and a closed ecosystem. It created its graphics processing units (GPUs) as processors for computer graphics. They turned out to be perfect for the huge calculations needed to train large language models for generative AI. Rivals are limited by a lack of semiconductor manufacturing plants needed to assemble new high-power chips. Its construction requires years and tens of billions of dollars.
Nvidia also offers a one-stop shop for customers through its CUDA platform, which allows them to use programs that run on Nvidia chips. A comparison could be Apple, which creates software that runs on its own hardware.
Combined, these help Nvidia dominate the AI chip market and set its own prices. The company’s gross margins have increased from 65 percent to nearly 73 percent in the space of two years. Compare that to Intel, where margins are 41 percent. Or 51 percent of AMD.
Sales could continue to increase. LLM training requires a lot of GPUs. The same goes for the next stage of using generative AI: inference. The market for processing generative AI requests, such as videos, images, and text, could be much larger than the market for training the models. Nvidia’s latest chip, the Blackwell B200 GPU, processes them 30 times faster than its latest chip. That will keep the company ahead of rivals as the focus on generative AI shifts.