AI gold rush: Nvidia joins the trillion-dollar club

The smartest insight and analysis, from all perspectives, rounded up from around the web:

The explosion of interest in AI turned the Silicon Valley chipmaker Nvidia into a trillion-dollar company this week, said Asa Fitch in The Wall Street Journal. Unlike other U.S. tech giants that have reached $1 trillion in market value — Apple, Microsoft, Google, and Amazon — Nvidia "isn't a household name." Behind the scenes, though, "its chips have become indispensable." Its graphics-processing units "made video games more crisp and less choppy," and were long the gold standard in gaming machines. It turned out that the computational power needed for graphics can also handle "the complex calculations that undergird modern AI systems" like ChatGPT. What we're seeing is the culmination of a decades-long plan first conceived during a Denny's lunch by Nvidia's leather-jacketed CEO, Jensen Huang. There is currently "no competitor that can match its breadth of chips and software for the computing-intensive demands of generative AI." 

Nvidia has the right product at the right time, said Tim Bradshaw and Richard Waters in the Financial Times. Its H100 chip, with three-times-better performance than its predecessor, is "one of the most powerful processors it ever built." When it launched last year, it "seemed badly timed, just as businesses sought to cut spending amid rampant inflation." But by an incredible stroke of luck, manufacturing of the chip at scale "began just a few weeks before ChatGPT debuted." The high-performance chip, which costs about $40,000, is now so in demand that it is "harder to get than drugs," according to Elon Musk.

Right now, AI is turning into a huge moneymaker for Nvidia but a big cost for many other companies, said Therese Poletti in MarketWatch. "Investors were excited to see any company put some actual numbers on this frightening and awesome technology," and there is a lot of hope here. But those investors need to remember that "spending on AI itself is going to mean a big investment cycle before some rewards are seen." Nvidia's success brings back bad memories of Cisco, said Jonathan Levin in Bloomberg. The dot-com era darling "surged to extraordinary levels on the promise that its routers offered exposure to the limitless upside of the booming internet." The internet's still around, but Cisco still hasn't regained the valuation of its 2000 peak. Even though Nvidia's market has essentially become anyone "investing in computationally intensive technology," it will still have to contend with the potential regulation from policymakers around the world. 

​​"Nothing short of a global ban" on the technology "is likely to stop the gold rush," said The Economist. While the companies designing AI software still need to figure out how to monetize their products, the sellers of specialist chips and computing infrastructure — right down to data-center floor space — are "already minting fortunes." Large language models are enormously computing intensive, and demand for AI computing power has surged an amazing 10-fold annually for each of the last six years. As long as everyone is rushing to mine AI gold, "the peddlers of picks and shovels will be cashing in."

This article was first published in the latest issue of The Week magazine. If you want to read more like it, you can try six risk-free issues of the magazine here.