Jim Keller, a well-known engineer who’s worked at AMD and Tesla and now leads the AI company Tenstorrent, recently said something interesting on the DemystifySci podcast. He thinks Nvidia is becoming like IBM back in the day but in the AI world. While this shows how powerful Nvidia is right now in AI hardware, it also raises some concerns.

Don't Miss:

Keller, an extremely accomplished engineer and entrepreneur, knows what he’s talking about. However, he might not realize just how unfortunate his comment that "Nvidia is slowly becoming the IBM of the AI era" is. 

In the early 1980s, IBM was huge in the personal computer market. Their PCs were everywhere, and saying "PC" often meant you were talking about an IBM machine. But, IBM’s reign didn’t last long because of a few key mistakes and growing competition.

Trending: Sun Belt's booming real estate market prepares for millions of new inhabitants — Here’s how to find the region's best deals!

IBM’s strategy of using an open architecture helped it dominate the market initially, but it also made it easy for other companies to make IBM-compatible PCs. These competitors sold similar products at lower prices, which ate into IBM's market share.

The partnership between IBM and Microsoft also backfired as Bill Gates’ company retained the rights to sell its operating system (MS-DOS) to other companies, paving the way for the rise of Microsoft Windows.

See Also: Many are surprised that they are eligible for a tax reduction on their home, with just a few seconds you can see if you can save in 3 minutes or less.

Nvidia could run into similar problems. As more companies start making AI hardware, Nvidia’s GPUs might get cheaper because of the competition. Technology moves fast, so today’s best chips can become outdated quickly. As Nvidia keeps growing, it could face more competition and rules to follow, just like IBM did.

In addition, there’s the $600 billion AI question. As Analyst David Chan from Sequoia writes, there’s a big gap between the money being spent on AI infrastructure and the revenue AI is actually bringing in. 

AI companies need to make about $600 billion a year to earn enough to cover these huge investments.

Chan said the gap was "only" $200 billion in September 2023. Even though Nvidia is making a lot of money, the rest of the AI market isn’t growing fast enough. 

Keller’s comparison to IBM is both a compliment and a warning because Nvidia's success now doesn’t guarantee its future. After all, the tech world is full of once market-leading companies that didn’t keep up with changes. 

Read Next: