Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
This tech earnings season has been a test of the bullish generative AI narrative that has propelled the valuations of the Magnificent Seven, a group of megacap stocks, and the entire US stock market, to new heights. Over the past two weeks, six of the seven companies published their second-quarter results. It has been a rollercoaster ride. Daily market capitalisations have gyrated by hundreds of billions of dollars, and the S&P 500 index has oscillated before falling around 6 per cent from the record it reached in mid-July. AI optimism survives, but the euphoria has been taken down a peg.
The likes of Google, Microsoft, Amazon and Meta have been investing heavily in AI technology to develop their own large language models, and to integrate it into their business. That has buoyed companies across the tech supply chain. Those involved in making chips, building data centres, model maintenance and training have seen revenues rise. Between the start of 2023 and mid-June, shares of semiconductor firm Nvidia — which is due to report results at the end of August — rose over eight-fold.
But with the AI infrastructure buildout cost projected by some to be around $1tn in the next few years alone, investors have been growing impatient. They were eager to see stronger earnings from big tech companies and progress on new apps as evidence that the hefty capital expenditure can be returned.
The second-quarter results have, however, been mixed. Results from Google, Microsoft and Amazon all underwhelmed investors. Apple reported a quarterly rise in revenues, though the market response was muted. On Thursday, solid revenue growth at Meta, coupled with a commitment from founder Mark Zuckerberg to continue spending heavily on AI, offered some uplift. Out of the seven companies — which make up about 30 per cent of the S&P 500 — only Apple and Tesla saw a rise in their market capitalisation in July, according to LSEG data.
The easing in AI hype offers a healthy pause for breath. Many investors piled into tech stocks amid a “fear of missing out”. Tech valuations have become stretched, and earnings have been highly concentrated — over half of the S&P 500’s total gains last year came from the Magnificent Seven. In a recent letter sent to clients, hedge fund Elliott Management said that AI is “overhyped” and that megacap stocks, particularly Nvidia, were in “bubble land”. If so, the removal of some air from the bubble is a good thing.
Although shaken, investors remain fairly confident that AI will bring about economy-wide gains. They are now, however, adjusting their expectations of when and how it will arrive. AI is still in the “picks-and-shovels” phase where upfront capital expenditure takes place before productivity gains can be reaped. Hardware takes time to build. Data centres need planning permission and ample electrical connectivity. Software glitches need ironing out, and as models develop, they will compete with one another too. Businesses also need to work out how to best use AI to drive profits.
Still, the turbulence highlights how a shift in confidence in a few highly valued and dominant tech stocks can infect the wider market. That means gradually closing the gap between perception and reality on generative AI’s capabilities takes on even greater importance, particularly as questions over the US Federal Reserve’s rates policy adds to volatility.
Excitement and “fomo” can indeed be a quick way to get investment into a potentially transformative though not yet fully developed technology. But now, a bit more patience and closer scrutiny can hopefully prompt truer pricing and more targeted investments in the AI sector.
Read the full article here