Investors’ concerns are growing over artificial intelligence spending by the US tech giants. A Goldman Sachs report provides a glimpse of investors’ worries regarding profitability from heavy AI spending.
Also read: US proposes new rule to curb AI investments in China
Hyperscale tech companies have spent nearly $357 billion on capital expenditures and research and development, especially in AI, during the past year, according to the strategist’s team led by Ryan Hammond. According to a Bloomberg report, Meta Platforms Inc., Alphabet Inc., Amazon.com Inc., and Microsoft Crop are considered hyperscale tech firms by strategists.
Big tech firms will soon cross the $1 trillion mark in AI spending
A significant portion of the multibillion-dollar spending is in the AI sector. This spending is so big that it equals almost a quarter of the S&P 500 total for research and development and Capital expenditures. Goldman Sachs analysts are skeptical of the time frame that most AI advocates claim, as they think it might take longer than expected for the cloud service providers’ investments to pay off.
Some even question the global expectations from the AI revolution, such as Jim Cavello, global head of equity at Goldman Sachs. He expressed his doubts about the technology, raising questions about the scale of its influence and profitability. Cavello said,
“Many people seem to believe that AI will be the most important technological invention of their lifetime, but I don’t agree given the extent to which the internet, cellphones, and laptops have fundamentally transformed our daily lives,”
Cavello continued, saying that all these technologies have already enabled us to do things that were not possible previously, like making calls, online shopping, and computing. He also estimated that AI expenditures by big tech will reach the $1 trillion mark over the next few years.
The bank’s equity head also opposes the idea that massive costs of powering AI products will decrease over the next year or two. Other chip makers are trying to overtake Nvidia’s dominance but haven’t succeeded much. This situation also contradicts the idea that increased competition would bring down the prices of the hardware needed for inference and training AI.
Companies will have to prove the profitability of their AI spending
A team of researchers at Barclays noted that big US tech firms’ heavy spending in the data center sector appears to be a “FOMO” (fear of missing out). The researchers also pointed out the difference between the amount of money Wall Street expects companies to invest in AI capex and the revenue these investments are supposed to bring in.
Wall Street expects cloud service providers to spend around $60 billion on AI-related infrastructure every year. However, the firms are expected to generate only $20 billion in additional revenue from these investments by 2026
Hammond said hyper scalers would eventually have to prove revenue generation and profits from the investments. He noted if there are early signs that these investments might not be profitable, it could lead to “valuation de-rating.”
Also read: SoftBank dumps Alibaba shares to pursue AI investments
Facebook-owner Meta and Google-parent Alphabet will spend record amounts on AI this year according to Bloomberg. Amazon will spend $63 billion in 2024, which is more than $53 billion last year. Nvidia is the biggest beneficiary of the artificial intelligence trade, though the AI boom has propelled some other US stocks to record highs as well.
Investors consider the AI rush to continue for the remainder of 2024 as it will continue boosting the rally. However, some analysts are bullish on infrastructure and utilities providers to lead the market in the second half of 2024.
The Goldman Sachs strategists said that AI spending was still less than capex levels during the dot-com crash. However, the Barclays team says the big firms are overspending on infrastructure. According to them, big firms’ investments in data centers would exceed the expected demand, as the projects already underway would be sufficient to power the internet. Hammond said sales revisions will be a key factor in assessing the profitability of AI investments.