The stock market has been captivated by artificial intelligence (AI) for the past year and a half. Pundits have heralded it as a revolutionary force poised to transform corporate profits and human productivity. However, a wave of skepticism is starting to wash over Wall Street, with analysts taking a closer look at the numbers underpinning the technology.
AI Investments: A Reality Check
Analysts are concerned that the massive capital expenditures being poured into AI by tech giants like Microsoft (MSFT), Amazon (AMZN), and Alphabet (GOOGL) may not deliver the promised returns. These “hyperscalers,” known for their vast cloud-service businesses, are investing heavily in AI infrastructure. However, some analysts doubt whether this spending spree will translate into significant profits or transformative change.
Goldman Sachs analyst Jim Covello, the bank’s global head of equity research, is a prominent voice of dissent. He questions the fundamental premise of the “AI revolution,” arguing that its impact on profitability and daily lives may be overstated compared to past technological advancements like the internet and smartphones.
Covello estimates that companies will shell out a staggering $1 trillion on AI-related capital expenditures in the coming years. However, he argues that AI, in its current form, may not be equipped to tackle complex problems, making it difficult to justify these immense costs.
The High Cost of Powering AI
Covello further challenges the notion that the expenses associated with powering sophisticated AI products will decrease over time. He points to the dominance of Nvidia (NVDA) in the graphics processing unit (GPU) market, a key component for training AI models. Despite years of competition, no chip designer has come close to dethroning Nvidia, suggesting limited potential for cost reduction in this crucial area.
Building for a Demand that Might Not Exist
The substantial costs associated with building and maintaining the infrastructure required for AI functionality are another concern. Analyst teams at Barclays worry that the current rush to build data centers by cloud-computing giants is driven more by “FOMO” (fear of missing out) than a sound business strategy.
They highlight a discrepancy between Wall Street’s projections for AI-related capital expenditures and the expected revenue generated by these investments. While cloud service providers are projected to spend an additional $60 billion on data centers and chips annually, the anticipated revenue increase from these investments is a mere $20 billion by 2026.
Barclays draws a parallel to the dot-com bubble, where companies overinvested in fiber-optic cable infrastructure. Their analysis suggests that the data center capacity currently under construction could power the entire existing internet, with room to spare for applications with user demands similar to ChatGPT. This implies that the hyperscalers’ data center expansion plans might be outpacing actual demand.
Investor Euphoria and Overvalued Stocks
Adding to the concerns, strategists at Citigroup warn of investor euphoria surrounding leading AI stocks. They advise clients to consider taking profits on high-flying semiconductor companies like Nvidia and Advanced Micro Devices (AMD).
Citigroup also highlights the issue of overvaluation in many AI-related stocks, even compared to Wall Street’s already optimistic projections for future cash flow growth. Tesla (TSLA), Nvidia, and Microsoft are mentioned as examples of companies whose stock prices appear to be priced for growth exceeding analysts’ expectations.
Skepticism Beyond Wall Street
The doubts surrounding AI’s potential extend beyond Wall Street. Academics like MIT’s Daron Acemoglu have expressed reservations about the transformative impact of AI on global productivity growth. Acemoglu’s research suggests that AI will only contribute to a 0.5% increase in US productivity over the next decade, with a corresponding 0.9% boost to GDP.
This stands in stark contrast to a study by the McKinsey Global Institute, which predicts that developed economies could experience an additional 3.4 percentage points of growth added to their GDP by 2030 due to AI and automation.
Investor Belief Persists
Despite the growing skepticism, investor enthusiasm for AI seems to be holding strong. Many AI-related stocks have already recovered from a brief pullback earlier this year. For instance, highflying names like Nvidia and Broadcom (AVGO) experienced a short-lived decline in June, but have since recouped most of those losses.
Similarly, AI software-focused funds, as reflected by the Invesco AI and Next Generation Software ETF (IGPT), are near their highest levels since late 2021, according to FactSet data.
Conclusion: A Sober Look at AI’s Future
While AI undoubtedly holds immense potential, a wave of skepticism is washing over Wall Street. Analysts are scrutinizing the high
