AI training is done mainly on Nvidia’s chips. The company temporarily breached the $1 trillion mark in the stock market.
Everyone seems to have a good idea of what AI is today and who are the big players—Google, Microsoft, OpenAI, Meta, and the random Elon Musk firm thrown into the mix with a bunch of startups all looking for a slice. What we missed to pay attention to was the hardware on which all of this training is happening. Nvidia makes graphics cards or GPUs which are used mainly for gaming, but it also designs GPUs for other workloads such as content creation and AI. The modern generation of Nvidia’s workstation GPUs has high-end AI accelerators that are being used by companies to train LLMs, which form the basis of the AI revolution.
Fueled by insane sales, Nvidia temporarily crossed the $1 trillion market cap on May 31 when trading opened at $405 per share. This is an exclusive club where Meta is an ex-member and current members include Apple, Microsoft, Google, and Amazon.
The market closed at $401.11 but peaked at about $419. At close, the Nvidia stock was valued at $992 billion. It might not be cemented there, but Nvidia has all it takes to join the club thanks to the current AI boom. You can try partnering with AMD for GPUs with AI acceleration cores but they’re not the best in the industry. Intel has entered the GPU market in some capacity but has a long way to go before it designs high-end AI accelerator chips.
With over $2 billion in profits over the last quarter, Nvidia’s market value also increased back in 2021-22 when the cryptocurrency mining boom introduced a GPU shortage and gamers were willing to pay 2-3x the price to get their hands on a GPU to play video games on. That, however, died down as Ethereum discontinued mining and secondhand GPUs were selling for 2-3x lower than MSRP throughout 2022’s last half.
Nvidia’s data center growth, the main hardware repository that’s utilized by big firms to rent chips to run their training workloads on, is central to its rising revenues.
CEO Jensen Huang strongly believes that AI has eliminated the “digital divide” and essentially made everyone into a programmer. In an interview with Wall Street Journal, Elon Musk said that “the GPUs are at this point considerably harder to get than drugs.”
The next biggest chipmaker in the world is TSMC, valued at $535 billion.
Nvidia faces a challenge too—One to deliver consistently high growth. And it can’t do that purely based on its data center capabilities. Nevertheless, the $1 trillion breach pushed many AI company stocks as well as chipmaker company stocks significantly.