Over the past few months, the prices of Artificial Intelligence (AI)-linked equities have boomed, largely driven by private deals. OpenAI’s valuation surged more than three times to $500 billion within a year, and Anthropic’s trebled in value. According to a leading computer scientist, Stuart Russell, the value of what AI can achieve in terms of artificial superintelligence, do everything a human can, and improve quality of life is $14 quadrillion. Investors are possibly right to buy into this future value today. Or are they?
A recent note by the Bank of England warns of a possible “correction” in the existing staggering valuations. It adds that the AI firms’ prices may be “stretched.” The fact remains that despite OpenAI’s worth, the company lost almost $8 billion in the first half of the calendar year, and may need 10 times more revenues to turn around. Like the pre-dotcom stock con at the turn of the century, this time too, the investors are betting on futures, even though AI firms incur losses.
Sam Altman, the chief executive of OpenAI, has spoken about the possibility of an AI bubble, or that investor expectations may have outpaced what the technology can realistically deliver in the short term. Yet, he leads one of the fastest-growing firms of the decade, and whose name is synonymous with the promises and perils of AI. His reflections represent the current paradox: A belief that AI will reshape everything, which is tempered by a recognition that the former inflates the stock prices and valuations beyond reason.
Altman argues that most tech-led revolutions begin with speculation, irrational exuberances, and overvaluations. The ups and downs in the stock market fuels the engines, via massive inflows of capital, to trigger sustained and genuine innovations. The stock frenzy, as was proved by several tech booms in the previous century, is the price that investors, firms, and nations pay for progress. His argument places OpenAI and other firms’ valuations not as the evidence of a bubble about to burst but as proof of the long-term convictions.
Yet, the problem today is about circular investments and vendor financing deals. This is what creates an impression of a bubble, which may be pricked anytime, anywhere. For instance, the deal between OpenAI and Nvidia, a renowned chip maker, envisages the latter’s investment in the former. In exchange, OpenAI will buy Nvidia’s chips. Morgan Stanley, an investment bank, is worried about the proliferation of circular financing which, according to another expert, “keeps everything afloat for now, but at some point investors will need to see returns” from their investments.
There is disquiet on Wall Street. The worry is akin to the pre-dotcom days, when every angel investor, venture capitalist, institutional investor, and high net worth individual was waiting for a crash, or a fall, in valuations. It was not a matter of if, but when. Today, there is a growing concern that borrowing costs remain high, capital is tight, and few firms can confidently project when their massive outlays will begin to pay off. The market is being pulled between expectations and estimations, and predictions and probabilities.
However, as usual, not everyone agrees with the demonic future. A note by Bank of America states that the existing “AI vendor financing concern” comes with “debatable optics.” Its semiconductor analyst, Vivek Arya, argues that “recent concerns (reference) AI financing are highly overstated.” He estimates that circular or vendor financing deals account for 5-10 per cent of the projected $5 trillion AI spend by 2030. Hence, their impact will be minimal. Most of the money is going into productive areas that will yield returns in the future.
UBS, a Swiss investment bank, in its report, ‘Should recent AI financing deals be a cause for concern?’ answers in a firm no. Although interconnected vendor pacts echo the telecom and dotcom eras, AI firms have stronger balance sheets, robust operating cash flows, and more prudence in capital deployment. According to UBS, AI giants generate billions in free cash flow even after capex. It is obviously a case of looking at the same financials from different perspectives. Is the capital glass half-empty or half-full?
A study by CEPR, ‘Unpacking US Tech Valuations,’ breaks down the price-to-earnings ratios across major tech names. It concludes that while the numbers are elevated, they are below the peaks seen during the dotcom era. The growth rates implied by stock prices are ambitious, but not impossible. The problem is that this leaves little room to accommodate investors’ expectations. If estimates go slightly awry, like a mild blip in earnings, or an unexpected regulatory delay, the reaction can be aggressive and intense.
What experts are concerned about is the centrality of AI in the western economies. A recent analysis in The Atlantic magazine warns that any sharp downturn can have far-reaching consequences. It cites experiments that show that professional coders using AI tools sometimes perform worse than those working unaided. The technology’s productivity benefits may be overstated. It points to surveys that indicate that many firms have yet to see meaningful profit gains from AI integration. If investors’ confidence were to crack, the dip in valuations may be swift, global, and brutal. Given how intertwined AI is, the fallout will spread beyond Silicon Valley, and run deep and wide into the real economy.
History provides both warnings and comfort. The dotcom crash was driven by weak business models and speculations. But it left behind a digital base that powers the next generation of the Internet. The telecom bubble wiped out the excesses, and preserved the networks that were essential for global communication. The same may unfold this time. The bubble may burst, if it does, there will be painful memories, followed by panicky resets in business models. What will remain will be the infrastructure for genuine long-term value.
Only, this time, the valuations are crazier than ever. Hence, a prick may devastate investors across the board, cripple stock prices, and detonate corrections that are more volatile than before. As a counter, one can contend that if the AI optimism aligns with productivity gains, if the tech fulfills its promise, the valuations may seem reasonable, even in hindsight. If not, another crash will enter the lexicon as a cautionary tale of ambition outpacing execution. At present, the boom is suspended between belief and proof.
As a recent article concludes, “However, the big four, Meta, Alphabet, Microsoft, and Amazon are this year spending the equivalent of the GDP of Portugal on AI infrastructure. This is not investment in new targeted ads, it is investment in an AI future. The bubble will burst if and when this future is in doubt.”
