The AI bubble is likely to get bigger before it deflates, analysts at Jefferies said in a recent note.
Since ChatGPT’s introduction, the market cap of a select group of 27 large-cap AI stocks has surged by nearly $10 trillion, a 127% increase. However, this growth stands in stark contrast to the modest 29% increase in their projected 2025 net earnings, which translates to a 73x price-to-earnings ratio.
Analysts noted that while AI capital expenditures are expected to remain strong in 2024 and 2025, the stock market has heavily rewarded major players like NVIDIA (NASDAQ:) and its major customers.
NVIDIA alone has seen its market cap soar 656%, while its core customers, including six cloud service providers (CSPs) and Tesla (NASDAQ:), have seen their market caps double, “meaning the stock market has also rewarded its customers for investing in AI.”
“This is a huge incentive for these clients to continue investing, at least until 2025,” the analysts said.
Moreover, these six CSP clients are financially well-positioned to invest, with total net cash flow of $108 billion at the end of 2023 and generating combined free cash flow (after capital expenditures) of $223 billion. However, analysts expect that by mid-2025, investors will begin to ask these companies harder questions regarding their monetization and ROI plans, leading to a significant reduction in visibility into AI capital expenditures in 2026.
Jefferies also points to significant challenges in monetizing AI. It estimates that between 2023 and 2025, global investment in AI servers, particularly those using GPUs in Nvidia data centers, will range from $400 billion to $500 billion.
But there is still a lack of promising business models or clear profit strategies from major AI players, the investment bank highlighted. Moreover, the annual energy cost of running these GPUs in massive data centers is expected to be around $27 billion at current energy prices.
“ASICs and non-NVDA GPUs are not counted. We would likely need to see additional AI revenues of over $100 billion in order to generate a 10% nominal after-tax ROI,” the analysts wrote.
When it comes to valuation, the Jefferies team believes it’s not “crazy,” “especially in the face of the dot-com bubble,” as analysts claim, comparing it to past market bubbles. They point out that while AI stocks are trading at high valuations today, they are backed by strong cash flows from their core businesses, unlike some of the speculative companies of the dot-com era.
However, analysts expect the bubble to deflate “if monetization does not occur in 2025/26.”