Nvidia, reconsidered | Financial Times

Receive free U.S. and Canadian company updates

we will send you myFT Daily Digest Email summary of the latest information U.S. and Canadian companies Every morning there is news.

Good morning. The Atlanta Fed’s third-quarter GDP forecast may seem implausibly high, but it’s actually on the rise. It rose 5.9% yesterday.One of the contributors: Taylor Swift and Beyoncé sold out concerts, plus Barbenheimer brought in 0.5 percent, according to Bloomberg Economics boost estimate. We’re at Jackson Hole today awaiting Jay Powell’s comment on that. Email us: robert.armstrong@ft.com and ethan.wu@ft.com.

Rethinking Nvidia’s Valuation

A week or two ago, I wrote an article suggesting that Nvidia might be overvalued. After the company’s second-quarter earnings report Wednesday afternoon and yesterday’s market reaction, I feel half right.

The earnings report was an absolute monster. The results far exceeded expectations, and so did the outlook. The company expects third-quarter revenue of $16 billion; a year earlier, revenue was $5.9 billion. The company is growing like wildfire. But the stock reacted yesterday by gaining a tenth of a percent. This suggests that all the good news has already been priced in, or in other words, the stock may be nearing its valuation ceiling right now.

I think I’m only half right, because Nvidia’s growth is so strong that I have to reconsider my thesis that the stock is governed by irrational exuberance, as it was with Cisco stock in 1999-2000. Although Cisco’s stock price has experienced 23 years of outstanding performance, its stock price still has not returned to its high point at that time. But Cisco has never been as far as Nvidia is now, so Nvidia’s valuation may make more sense than I think.

Doing some rough math can bring clarity to the valuation picture. Let’s say the company reported GAAP net income per share of $3.25 for the quarter. That seems likely given the revenue guidance. Annualized, the company makes $13 a year. At current prices, the forward P/E ratio is 36. Apple’s P/E ratio is 27, and its revenue is growing in the low single digits. The S&P 500 Index ranks 20th. One could argue that relative to its growth, Nvidia is cheaper than Apple or even the market.What this week’s earnings report shows is that while Nvidia’s valuation is high, it’s definitely no The Internet is crazy. (Of course, you could argue that Apple is overvalued, or that Big Tech is overvalued, or that the market as a whole is overvalued; maybe it is; but what we’re talking about here is relatively valuation. )

The extent of Nvidia’s growth is undeniable. What about its stability?

It appears Nvidia’s competitive position in AI chips is unassailable, and not just because they’re better at the heavy parallel processing AI requires. Its “full-stack” offering combines chip design, hardware and software, and is the only out-of-the-box solution for companies looking to quickly add AI capabilities. As chipmakers like Intel and AMD catch up in parallel-processing chips, Nvidia’s installed base will become an insurmountable competitive moat, just as Intel did with its x86 processors in personal computing a few years ago.

A conversation with Gartner AI analyst Chirag Dekate convinced me it could be looked at in another way. In a sense, Nvidia’s main competitors are not other chipmakers but the big computing platforms: Amazon Web Services, Microsoft Azure and Google Cloud. The companies plan to use AI in their own services and sell AI capabilities to customers.

Crucially, platform companies compete on cost, and their approach to delivering low-cost AI capabilities and services is to use the computing infrastructure they already have. In other words, their motivation is to provide customers with a variety of choices, including Nvidia’s out-of-the-box, highly specialized products and their own in-house technology that can handle artificial intelligence work and their and their employees. All other types of computing work. Clients do this. These platforms’ own purpose-built AI infrastructure has the potential to bring additional computing benefits to the enterprise.

“If a platform uses Nvidia across the board, their cost advantage may disappear. Over time, the only way they will have a cost advantage is to start reducing costs at the infrastructure layer and explore alternative paths to improve efficiency,” DeKitter said. He believes that Nvidia’s alternative as soon as next year will offer businesses more flexibility and potentially excellent price/performance. The AI ​​chip market is still up for grabs. “You don’t announce a winner after the first inning,” he said.

With that in mind, it’s less of a question of whether Nvidia now looks grossly overvalued relative to its peers or the broader market. I don’t think so anymore. The question is what will happen in two years. The answer depends on how AI services are delivered in 2025. If you can predict this with confidence, please do email me.

good read

You really don’t want to be sitting in a “quantitative ejection seat”.

FT Unhedged Podcast

Isn’t Unhedged enough?listen our new podcastHosted by Ethan Wu and Katie Martin, this twice-weekly 15-minute in-depth look at the latest market news and financial headlines.Learn about past editions of the newsletter here.