Stay up to date with free updates
Just sign up artificial intelligence myFT Digest – delivered straight to your inbox.
Some predictions about the future seem absurd but at the same time reasonable, such as star tech investor James Anderson’s estimate that US semiconductor company Nvidia could be worth $49 trillion in 10 years’ time.
At first glance, it seems absurd for Anderson (whose Lingot Investment Management fund holds a large stake in Nvidia) to suggest that the U.S. chipmaker could one day be worth more than the current S&P 500 index combined.
Meanwhile, someone would have to make astronomical profits to justify the extraordinary surge in capital spending by big U.S. tech companies, and the most likely candidate is Nvidia, the maker of the graphics processing units that are powering the artificial intelligence revolution and described by one analyst as the “nerve center” of AI systems.
The scale of capital spending by the big US technology companies (Microsoft, Alphabet, Amazon, Apple, Meta) is staggering, as they put their poker chips forward in the belief that AI will change the world. Together, these companies are investing in what is likely the largest, and certainly the fastest, infrastructure rollout in history.
Arete Research projects that these companies will spend about $480 billion in capital expenditures over the next two years, much of it for 100 data centers currently under construction. Many of those data centers will run on Nvidia’s GPUs. For now, that gives the company a near monopoly and pricing power. This market power helps explain Nvidia’s stock price’s 2,700% rise over the past five years. But the company’s sharp decline over the past week suggests the first signs of a nervous breakdown.
But the bigger questions for both the stock market and the economy are who will benefit most from AI and when the economic fruits of this technology will be fully realized.
Analysts say tech booms go through cycles, with infrastructure companies such as Nvidia benefiting first, followed by platform companies such as cloud providers Microsoft, Amazon and Alphabet, followed by application companies such as Uber and Airbnb in the last internet revolution.To date, no company has developed a “killer app” for generative AI, but a number of startups have successfully sold the dream to venture capital investors.
There are differing views on how quickly these applications will take off and what their economic impact will be. In a controversial Goldman Sachs paper, MIT economist Daron Acemoglu made a strong case that the benefits of AI will be much smaller than investors assume and will take much longer to realize. On the other hand, there are asymmetric risks that the drawbacks of AI technologies, such as deepfakes, will emerge sooner than the benefits.
Acemoglu predicted that AI will boost U.S. productivity by only about 0.5% and GDP by about 1% over the next decade, well below Goldman’s projections of 9% and 6.1%, respectively. If Acemoglu’s analysis is correct, the U.S. stock market, including Nvidia’s, is headed for some messy earnings reports.
But Joseph Briggs, a senior economist at Goldman, countered that AI will automate far more business processes than Acemoglu envisions. Briggs argued that, as in previous economic cycles, workers who lose their jobs will find new roles opened up by the potential of modern technology, making them even more productive. Economist David Autor has calculated that 60% of today’s workers hold jobs that didn’t exist in 1940.
But as one Goldman analyst noted in the same report, big investments in technology in the past haven’t necessarily delivered big returns, with investors still waiting for the fruits of virtual reality, the metaverse and blockchain to materialize.
Moreover, other powerful headwinds are likely to hit the corporate sector in the coming years. U.S. corporate profits as a percentage of GDP are already near their highest levels since World War II, while labor’s share is near its lowest on record. Further increases in corporate profits, which are needed to justify these huge capital investments, may only cause further social unrest.
Even Anderson, a bullish analyst, thinks there’s only a 10% to 15% chance that the data-center investment boom will continue and Nvidia will reach a $49 trillion valuation, so investors should be prepared for other possibilities.
john.thornhill@ft.com
