AI Hype, Yes. AI stock market bubble? No, not yet.
Forecast: mostly clear with heavy hype
The Wet Bias
Have you ever suspected that weather forecasts are overly pessimistic? You’re right! Research shows that local TV stations and The Weather Channel overestimate the chances of rain by as much as 30%.
This isn’t some sinister weather cabal - it’s human behavior known as the “Wet Bias”. No one blames the weatherman if it doesn’t rain, but they catch grief if they predict sun and picnics get rained out. Viewers are much more likely to tune in to a dismal forecast than sunny skies.
Which brings us to the stock market. Today’s newsfeeds are flooded with headlines like “Is the AI stock bubble about to explode?” It’s safer to wrongly predict a bubble than to miss one, especially when everyone else is calling it. And these headlines get a lot more clicks than ones that read “nothing to worry about”.
Bubbles aren’t a joke. The dot-com crash wiped out 80% of the Nasdaq’s value, and it took 13 years for markets to reclaim their 2000 peaks. So grab a poncho and let’s see what the forecast has in store for AI.
Bubble Watching
So what’s got the forecasters bubbling? Four main concerns:
Valuations are approaching dot-com levels: The S&P 500’s price-to-earnings ratio (which tracks stock prices against profits) is nearly as high as it was during the dot-com bubble. This means investors are betting on massive growth.
Big Tech dominates the market: The nine biggest tech companies make up 41% of the value of the S&P 500. During the dot-com peak, the top companies controlled only 16%, and retail investors were betting on startups like Pets.com and Webvan. The market is so dependent on these giants that the Financial Times proclaimed that “America is one big bet on AI”.
AI infrastructure investments are massive: Amazon, Google, Microsoft and Meta will spend almost $400 billion this year on infrastructure, that’s before counting OpenAI’s buildouts. The scale is staggering: Harvard economist John Furman estimated that 92% of US GDP growth in the first half of 2025 came from datacenter spending.
Revenue hasn’t caught up to investment: Morgan Stanley predicts total AI revenue of $153 billion in 2025, but companies are spending $400 billion on infrastructure. $3 spent for every dollar earned is an expensive bet on future growth.
Checking the Radar
Historically high valuations, massive investments that dwarf today’s revenues…that sounds pretty bubbly! But let’s look a little closer:
Revenue is Growing: Unlike dot-com darlings like Pets.com and Webvan, today’s leading AI companies are making real money. Anthropic has grown revenue 900% this year, while OpenAI grew 400%. Big Tech is seeing similar acceleration, with cloud revenue growth rates picking up after years of decline.
Prices aren’t unreasonable: Remember that 41% concentration for Big Tech that sounds scary? It shouldn’t be. These are established companies trading at reasonable historical levels. Excluding outliers like Tesla, today’s big tech PE ratios are around 30, compared to the dotcom boom when Microsoft traded at 69, Cisco at 101 and Oracle at 90.
Cash over Credit: While infrastructure spending is unprecedented, the big tech firms can afford it. They generate $350 billion of net income per year and hold $300 billion in cash reserves. This is a sharp contrast to the dotcom bubble, when telecom companies loaded up on debt they couldn’t repay when their stock values crashed.
Whew. Thanks for hanging in there while we checked the financials. But none of this matters if AI isn’t actually useful.
AI is the Real Deal
Is AI actually useful? The answer is a resounding yes. Enterprise usage doubled this year to 44%. Developers using AI coding assistants report 55% productivity gains and AI customer service agents are handling millions of routine inquiries. Companies are deploying AI across the enterprise, from hiring to marketing to contract reviews.
There is also a ton of room to grow. More than half of companies haven’t deployed AI yet. Customer service is only at 22% penetration, though Gartner predicts AI will address 80% of all issues by 2028. And leaders are betting big: 7 in 10 CEOs called AI a top investment priority in KPMG’s 2025 annual survey.
Not Yet - But Keep Watch
We may not be in a bubble today, but there are three factors that could put us there in the next couple years.
1. Perishable GPUs. Nvidia refreshes its AI chips yearly and most companies replace them every two years. This creates a catch-22: Nvidia needs accelerating sales to hit growth targets, while tech companies will need to reduce chip spending to return to restore profitability.
2. Saving the world? While AI companies are chasing moonshots like curing cancer and nuclear fusion, companies need tools that reliably solve business problems. Scientific breakthroughs are great for humanity, but pursuing them could delay profitability.
3. Business Adoption Timing. Remember Pets.com and Webvan? These were the same ideas as Chewy and Instacart, just 12 years too early. If businesses adopt AI too slowly we could end up in another dot-com scenario.
An Emotional Bubble
There is one bubble we can all agree on, and that’s the emotional one. Each week brings a new revolutionary tool, round of AI-based layoffs, or doomsday prediction. After three years of dizzying headlines, a market correction would almost be a relief, a chance to catch our collective breath. And here’s the thing: this mental exhaustion could be driving the bubble fears as much as the financial data.
It’s hard not to worry, especially for those who remember the dot-com crash. But when you look at the numbers, the financial signals are solid: growing revenue, reasonable valuations, real business impact. So yes, keep an umbrella (and maybe a poncho) handy. But also let yourself take a break from the AI headlines for a bit. AI will still be here next week.
Dad Joke: Why does the AI weatherman always predict rain? Because it’s thinking in the cloud! 🤣
Thanks for reading!
If you enjoyed this edition, share it with someone still carrying an umbrella indoors.