![[Kenya] Digital credit provider Tala disbursed Sh240 billion in loans in eight years](https://theafricanbusiness.com/wp-content/uploads/2023/02/TALA-APP-360x180.jpg "[Kenya] Digital credit provider Tala disbursed Sh240 billion in loans in eight years 19")
In the two years since ChatGPT burst onto the scene, artificial intelligence has come to dominate investor consciousness more than any other technological breakthrough in the past two decades. Tech giants are spending tens of billions of dollars per quarter to beef up the computing power needed to develop and run AI systems.
To its biggest boosters, the impact will be astonishing: AI will replace legions of workers, help researchers discover lifesaving drugs, enable companies to push into new markets and unleash vast efficiencies that will juice corporate profits for years to come. As a result, AI-related stocks have been responsible for much of the bull market that began in October 2022.
While AI is surely promising, it isn’t generating much revenue relative to the cost. A recent Gallup poll found that only 4% of US workers use AI every day. More than two-thirds said they never do. Daron Acemoglu, a Nobel Prize-winning economist and Massachusetts Institute of Technology professor, has argued that common expectations around AI advances are overly optimistic. “The models we have right now are pretty impressive in some respects,” he says. “They’re still not usable broadly.”
Some investors are feeling vibes from the 1990s, when the nascent internet was generating a similar euphoria. That online transformation took a lot longer than expected, creating a big dislocation between stock prices and fundamentals that ended in an epic crash. Some Big Tech stocks are once again trading at prices, relative to their earnings, that are well above historical averages.
Today, whether they mean to or not, most investors are making some kind of a bet on AI.
Own an S&P 500 index fund? A third of your money is in eight companies, including Nvidia, Microsoft and Apple, that have staked some of their future on AI. That’s not to mention tangential sectors, such as utilities, that are benefiting from AI’s electricity-thirsty data centres.
For AI bulls and bears alike, it’s worthwhile to weigh the risks and opportunities the technology presents. Start by understanding the difference between companies creating AI services, such as Microsoft Corp and Alphabet Inc, and ones providing the infrastructure — chips, servers and power — that makes computing possible.
The tech giants are vastly profitable and responsible for the bulk of AI spending. In the third quarter of 2024, Alphabet, Amazon.com, Apple, Meta Platforms and Microsoft spent a combined $62 billion on capital expenditures. That record sum amounted to a more than 50% jump from the same period the previous year. For the first time since 2018, the group put more money into capex than stock buybacks — $5 billion more, in fact. These companies can clearly afford the spending binge. They still generated $76 billion in free cash flow. Investors generally support Big Tech capital spending because the sector has succeeded in the past.
Dave Mazza, chief executive officer of Roundhill Financial Inc, a New York investment firm, believes AI will generate enough revenue but thinks investors may lose their patience. “If companies are coming out and saying, ‘We’re going to plow this much into spending, but I can’t tell you when you’re going to see payback on this investment,’ investors are probably going to start voting with their feet,” he says. “At some point reality can set in, and the clock is ticking.”
Arvind Narayanan, co-author of AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference, says big AI payoffs will take more than a year or two. “Technology adoption requires changes to people’s workflows and sometimes organizational structures and legal guardrails,” says Narayanan, a Princeton University computer science professor. “These changes don’t happen at the speed of technical innovation. So we should expect that it will take a decade or two for businesses to fully reap the benefits of even today’s generative AI.”
So far, AI-related revenue for the tech giants is most evident at cloud computing businesses such as Amazon Web Services and Google Cloud. But Microsoft—an early AI winner, thanks to its partnership with ChatGPT’s owner, OpenAI — is already feeling some pressure. As of the end of November, its stock had been lagging the S&P 500 for four months. The company suffered its biggest stock drop in two years in October after its forecast for revenue growth at its Azure cloud computing business fell short of expectations. Still, Microsoft is projected to spend $62 billion on capex in its current fiscal year, more than twice its outlay two years ago. Alphabet, Amazon and Meta have also said they’ll boost investment.
That commitment has reassured those using a popular approach toward betting on AI: the so-called pick-and-shovel trade — a reference to merchants who made money by selling gear to prospectors, who rarely hit it big, in the 19th-century California gold rush. Nvidia Corp is a classic example. It dominates the market for the computer chips best suited to handle the computing required by AI. Shares have continued to hit records, gaining more than 700% since ChatGPT was released in November 2022. Nvidia’s soaring profits and sales have made it one of the most valuable companies in the world, with a market capitalisation of $3.4 trillion at the end of November, jockeying with Apple Inc. for the top spot.
The usually staid stocks of power producers are also surging. Vistra Corp and Constellation Energy Corp are among the best performers in the S&P 500 in 2024, helping put an index of utility stocks on track for their best year since 2019.
With tech giants pledging to continue spending, even bears such as Jim Covello see no reason to abandon the pick-and-shovel trade while the money is still flowing. Covello, head of equity research at Goldman Sachs Group Inc, says AI isn’t capable of doing many of the things its proponents claim. “Most technology transitions in history, particularly the ones that have been transformational, have seen us replace very expensive solutions with very cheap solutions,” he said in July. “So far, AI — and its tremendous costs — is the polar opposite.” In his view, when that becomes widely realized, the tech giants will start to rethink their massive spending, and the infrastructure stocks will be in for a tumble.
That scenario would also be a blow for Big Tech companies, denting their prospects for growth. But it probably wouldn’t do much damage to their appeal over the long term, because growth from core businesses would remain intact.
Consider Meta Platforms in 2022, when the Facebook parent pumped money into the development of the so-called metaverse. Once revenue slowed, the stock collapsed. But Meta recovered, and it’s now trading at record levels after many mea culpas from CEO Mark Zuckerberg and a change of course.
Picking this round’s long-term winners might be a fool’s errand, says Michael O’Rourke, chief market strategist at Chicago-based Jonestrading. In 2000, hardware makers such as Cisco Systems Inc and Sun Microsystems — the pick-and-shovel crowd during the dot-com boom—were some of the era’s best-performing stocks. But ultimately, such upstarts as Amazon and Google built massive businesses around the technology.
“It’s usually someone else who refines these ideas who comes along and dominates,” O’Rourke says. “We’re still so early in the process for seeing what the environment is going to look like in the future. I don’t want to say it’s too early to tell, because that would imply, well, ‘if you’re early you should buy.’ The late ’90s was early in the internet, and that wasn’t the right time to buy.”
