IMPORTANT: The Revaluation of AI Stocks Is Coming and Most Will Not Survive

I’ll be straight to the point: you’re late in investing in AI stocks.
We are not in the early stages of a new technology cycle, but already deep into it. Gartner officially classified generative AI (GenAI) as in the recession phase last year.
On average, each company will spend $1.9 million on GenAI by 2025, and less than 30% of CEOs say they are satisfied with their ROI (Return on Investment). That is a BIG warning.
However, the market still prices these companies as if they will succeed in the long term.
Let’s do some math. The total market capitalization of publicly listed AI-related companies is around $21 to $23 trillion.
To justify that with an annual profit margin of 10%, they need about $2.2 trillion in annual profits.
Their current net income is nearly $420 billion, and most of that isn’t even from AI.
Investors are paying five times the future profits that don’t even exist, over a time frame that no one can model, in a field where unit economic efficiency has been broken.
OpenAI, perhaps the most important AI company today, spends about $1.69 for every dollar earned.
The company expects a $14 billion loss this year, with total accumulated losses of $115 billion before turning a profit in 2029.
It is raising $100 billion with a valuation close to $830 billion.
This figure is even larger than Argentina’s GDP for a business that is still losing money at a rate similar to WeWork.
Meanwhile, large cloud service providers are planning to pour between $650 and $690 billion into AI investments this year.
Amazon alone has spent $200 billion.
The simple problem: data centers coming online in 2025 have depreciation costs of $40 billion annually but only generate revenue of $15 to $20 billion at current utilization.
That calculation is not feasible.
In Deutsche Bank’s global market survey, 57% of investors believe the AI valuation crash is the biggest risk in 2026.
One of their strategists frankly said: “The AI and tech bubble risk far exceeds anything else.”
This is like the dot-com bubble era repeating itself, only with extra letters added.
In 1999, adding “.com” to a company’s name boosted market cap by billions overnight.
Today, just mentioning “AI” in earnings reports can do the same.
The psychology is exactly the same.
Morgan Stanley estimates retail investors have poured about $700 billion into stocks since January, five times the amount during the 2000 bubble.
The dot-com bubble burst doesn’t prove the internet was wrong.
It shows that valuation is crucial, and picking the winners is nearly impossible until reality adjusts expectations.
Cisco peaked at $555 billion in 2000 and took two decades to recover.
Amazon, trading at just a few cents in 2001, quietly became a $2 trillion company.
That’s what I will be watching closely.
When the revaluation happens, it will be brutal.
Companies solely focused on AI without competitive advantages or revenue will be crushed.
Companies with P/E ratios of 70 based on non-existent projected revenues will plummet.
But what comes after that is where the real potential lies.
The survivors will be those with genuine ecosystems, highly integrated products, revenue streams outside of AI, and strong balance sheets to survive.
Think of the Amazon and Google of this cycle—players in infrastructure providing energy for the entire system.
When things settle down and real money starts to be made, those surviving companies will not just be worth hundreds of billions—they will be measured in trillions.
This technology is transformative, but not as quickly or as widely as the market assumes.
I’m not pessimistic about AI. I’m pessimistic about how some people think about something that is still uncertain.
Be patient. Let the cycle run its course.
The real move is knowing which stocks to own when everyone else has given up.
When that moment comes, I will tell you where I plan to invest.
Many will wish they had followed me earlier.