Published in Svenska Dagbladet, 2024-08-17. Translated from Swedish.
After this spring’s stock market surge among tech companies, markets have started to doubt. Will AI deliver the economic explosion investors hoped for? Several influential voices are now expressing scepticism for the first time.
“I think it is insane.” Barton Briggs, strategist at investment bank Morgan Stanley, was speaking about Cisco in April 2000. The networking giant had been the star of the dotcom era, but some were beginning to wonder whether everything was as it seemed. “I cannot understand how you can make money on a stock trading at 150 times revenue, with a market cap of $500 billion,” Briggs continued, to the Wall Street Journal. He was right. A year later Cisco’s stock had fallen 80 percent. The dotcom bubble had burst.
After a turbulent summer for tech stocks, more experts are now asking whether we are facing a similar moment. Chipmaker Nvidia has fallen 15 percent from its summer peak. For the year as a whole the company is still up 145 percent, and looking back five years the gain is over 2,800 percent.
Attention is now turning to Nvidia’s next quarterly earnings, due within two weeks. Can the locomotive of the AI boom keep growing? One can sense some doubt, at least.
Much of the criticism centres on concern about how much is being invested in AI relative to the value companies are extracting from it. This summer Goldman Sachs published a report highlighting several critical voices. Daron Acemoglu, professor at MIT, is among the sceptics. He believes AI development is important, but that the major breakthroughs are further away — closer to ten years than one or two. While Goldman Sachs’ own economists believe AI could drive GDP growth of 6.1 percent, Acemoglu puts the equivalent figure closer to 1 percent. He is also sceptical that AI will deliver the cost savings many have hoped for.
Jim Covello, head of equity research at Goldman Sachs, argues that AI needs to be capable of solving very complex problems to justify today’s investments of around $1 trillion. That is something the technology currently cannot do. He also believes the savings being generated by AI today are too simple and too cheap to justify the enormous cost of building and maintaining these new services — and sees no reason to expect this to change in the near term.
There are naturally still enthusiasts in the market. The more optimistic camp sees a market that is still in its infancy. For them, AI development does not necessarily need to create wholly revolutionary products and services. Simply making existing things easier and faster may be entirely sufficient.
Looking back at the early dotcom period, the payback time on those investments was long too — particularly at the start. But as the technology matured, the pace accelerated. Eric Sheridan, a somewhat more optimistic equity analyst at Goldman Sachs, concedes that the paybacks need to be larger to justify the enormous investments now being made. But he does not believe that a somewhat extended build-out phase will deter the major companies from continuing to invest in AI.
For the large tech companies — Google, Amazon, Meta — the risk of investing too little and falling behind appears to feel considerably greater than the risk of wasting money on something that never materialises. Those with the most reason to be enthusiastic are those with large stakes in Nvidia, the company that has become the central node of AI development alongside OpenAI. Nvidia’s success is built largely on other tech companies’ investments in the sector — their capital expenditure becomes Nvidia’s revenue.
It comes down to patience, on several levels. How long will the tech giants be permitted to spend billions of dollars on bets that generate little in return? Will demands come from major shareholders wanting dividends or buybacks instead? Activist fund Elliott Management has taken a clear position, saying in a letter that Nvidia is in “bubble-land” — while adding that it would be “suicidal” to try to short the stock, given its trajectory.
As long as enthusiasm around AI remains high among the biggest tech companies, Nvidia should continue to do well. But for the first time in this cycle, at least some doubt is being heard. When Nvidia’s quarterly report is unveiled, we will see whether that doubt shows up in the numbers.
