Get your free copy of Editor's Digest
FT editor Roula Khalaf picks her favourite stories in this weekly newsletter.
There are several truisms about the tech bubble, one of which is that when you're in a bubble, it's hard to recognize it: Individual spending and investment decisions can seem rational, even if the ultimate impact seems extreme.
When there's general agreement that a bubble is forming, it often continues to inflate, causing losses for investors who exited early.And after the bubble bursts, it can take years to determine whether it was just a product of hype or a harbinger of an even bigger tech boom to come.
As US tech companies enter their latest earnings season, Wall Street is starting to feel a distinctly euphoric vibe in tech valuations. Mid-2024 was always expected to be an “air pocket” period for generative artificial intelligence. The investment boom generated by this technology is visible, but it will take time for all of its new capabilities to be productively utilized by end customers in the tech industry. Wall Street’s patience is about to be put to the test.
Software companies that are best positioned to capture value from new technologies by building them into existing products are struggling under the stock market dark clouds.
And despite Apple's recent promise to generously infuse its devices with AI technology, AI has yet to produce compelling new services for consumers. Many who discovered ChatGPT were intrigued, but unlike the first time they picked up an iPhone, used Google's search box, or found friends on Facebook, AI has not transformed their digital lives.
At best, this signals a slow start to widespread adoption of generative AI. At worst, it means the technology is not as transformative as tech companies claim. The longer the stagnation continues, the more pronounced the gap between booming investment and sluggish final demand will become.
Underscoring the point is the narrow base of this investment boom: Nvidia said late last year that cloud companies—a market dominated by a few big players—now account for more than half of its data-center chip sales. Any signs that big tech companies are pulling back on spending during earnings season would be a severe blow.
But as tech companies prepare to release their latest earnings, there are all signs that the boom is still in full swing. Many enterprise customers are only just starting their first pilot projects with the technology, and will likely ramp up testing of the technology in the coming months, even as the ultimate use remains unclear. Investing in large-scale language models and the infrastructure to support them has also become a strategic imperative for the big tech companies themselves. If, as many in the tech industry expect, machines that can “understand” language and images will become an entirely new computing platform, all the big companies will need to get a firmer foothold in the technology.
It's also worth noting that these companies have deep pockets to sustain and even intensify their fight: The combined operating cash flow of Apple, Microsoft, Alphabet, Amazon and Meta has grown 99% over the past five years, reaching $456 billion in 2023. That's more than enough to cover capital expenditures, which have ballooned 96% to $151 billion.
Meanwhile, for chipmaker Nvidia, the next big product cycle based on its new Blackwell chip architecture isn't even scheduled to start until later this year, which is expected to bring down the costs of training and running large-scale AI models, ensuring a customer stampede even as demand for previous-generation chips remains strong.
Herein lies the paradox common to all new technology platforms: When the cost of using a new technology plummets, customers can, in theory, buy less. But when costs fall too quickly, new uses are usually found and demand surges. As with everything with generative AI, this is happening at lightning speed, and it's hard to say how similar this is to other disruptive technology cycles.
Of course, eventually all of this investment has to pay off, or the CEOs who have been demanding their companies find ways to use generative AI, driven by fears of the board and competition, will eventually lose patience and move on. But all the signs are that we're not there yet.
RichardWaters@ft.com