History doesn't repeat itself, but it certainly it helps you see patterns -- especially in financial markets, which are characterized by booms and busts driven by macroeconomic factors like the business cycle and interest rates. After gaining a whopping 79% during the past five years, the S&P 500 (^GSPC +0.55%) might be too pricey based on some valuation metrics. Let's dig deeper to decide how this story may play out during the coming months.
AI-related spending looks unsustainable
The launch of OpenAI's ChatGPT in late 2022 set off a technology boom on a level not seen since the internet took off in the late '90s and early 2000s. Companies are scrambling to stay relevant in the market for training and running large language models (LLMs), which means spending huge sums on artificial intelligence (AI) chips and other compute hardware.
Analysts at Goldman Sachs estimate that cloud computing giants may spend an eye-popping $500 billion on AI hardware in 2026 alone. And the data center buildout will likely be responsible for a significant percentage of GDP growth this year.
There are several alarm bells about this trend. For starters, these companies are spending hundreds of billions on depreciating hardware, such as graphics processing units (GPUs), which will eventually fail or become obsolete as technology improves. Depreciation could be a long-term drag on corporate earnings if the AI investments don't meet expectations. And so far, the technology isn't translating to huge gains for many of the companies involved.
Where are the profits?
So far, the generative AI boom has played out similarly to the California gold rush, where the people making most of the money were selling picks and shovels instead of mining gold. The biggest winner has been AI chipmaker Nvidia, whose third-quarter earnings rose 62% year over year to $57 billion, while profit soared 65% to $31.9 billion.
However, not every company has the same experience. The Economist believes OpenAI could burn through as much as $17 billion in cash in 2026. And the company's planned initial public offering (IPO) in the second half of the year could bring investors face to face with the alarming economics behind the AI hype. They might not react well to what they see.
Oracle could be a harbinger of what comes next. Shares of the data specialist are down roughly 52% from their all-time high as the market contrasts its debt-fueled data center spending with lackluster operating performance.
To put these numbers in perspective, Oracle's top line only grew by 14% year over year to $16.1 billion. And it plans to make $50 billion in capital expenditures this year.
In 2026, expect the market to become less patient with AI spending that doesn't show results. And that could lead to a widespread retreat in technology stock valuations.
One valuation metric flashes a warning not seen in decades
Image source: Getty Images.
Investors face some serious challenges this year as the economy grapples with potentially unsustainable AI data center spending. And if history is anything to go by, the S&P 500 may be due for a correction.
The cyclically adjusted price-to-earnings (CAPE) ratio is a market valuation metric that divides the current price of the S&P 500 by its average inflation-adjusted earnings during the past 10 years. It's designed to smooth out factors like inflation and the business cycle to give investors an idea of whether stocks are cheap, fairly valued, or expensive. Right now, it stands at about 40 -- a level not seen since the dot-com bubble and far above its long-running average of just 17.33.
Historically, CAPE ratios don't stay this high for long. And investors should consider shifting away from richly valued technology stocks that might be highly exposed to potential downside.
