Reading · Swift Start Go | By Caroline Swift Holden

This may be a hot take, but I have a theory: you can tell if a bubble is about to pop when you tell people you work in a hot industry and they have a strong, negative, visceral reaction.

In the last month, I’ve had roughly a dozen people scowl, roll their eyes, flinch, or even burst out laughing when I mention that I write content about AI.

Nearly every time, the conversation quickly shifts to:

“Aren’t you worried AI is destroying education?”
“Isn’t it causing people to spiral into delusions?”
“What about the climate crisis from all those data centers?”

These are well-informed people I deeply admire. But they’re also puzzled when I explain that I’m not an AI evangelist or a doomsday critic. I see myself as a realist: someone who wants the facts, who refuses to get locked into black-and-white thinking.

And yet, lately I can’t shake the feeling that the walls are closing in, just like during the crypto boom. Everyone’s picked a side, armed with talking points, and convinced the other camp is clueless. But in my experience, it’s rarely the loud debates that pop a bubble.

Crypto didn’t implode because people were arguing about decentralization. It collapsed because Sam Bankman-Fried wasn’t doing basic accounting.

My prediction for an AI bubble popping? OpenAI’s strange nonprofit/for-profit hybrid and its brutal burn rate may eventually become its undoing.

Because at the end of the day, it’s usually math that topples the hype.

The last few weeks have been buzzing with speculation: is AI officially in a bubble?

Even Sam Altman himself, who has every incentive to keep the hype train running, recently admitted that, yes, we’re in bubble territory. And the numbers around AI are simply nuts:

Valuations are off the charts. OpenAI ran a $5B tender offer (making it the most valuable private company in the world) and is valued at 30x its revenue, and Thinking Machines raised a $2B seed round at a $12B valuation.
Big Tech is spending at historic levels. Over $100B on data centers this year alone, surpassing office construction in the U.S. and AI spending is now so large that it has officially surpassed US consumer spending.
Capital is desperate. Investors are forcing their way into deals through SPV’s, even when companies are burning cash with negative gross margins.

Most of these companies are losing money and the unit economics isn’t, as the kids say, “mathing.”

We’re living in two economies: the roaring, AI-driven, B2B oriented stock market, and the consumer economy where job creation has frozen, and consumer spending is dropping.

Hell, last week the New York Times reported that only 956 private sector jobs were created in NYC and almost 78k tech jobs lost in the first six months of 2025.

Meanwhile, research groups predict 85 million to 800 million jobs could be lost by 2030 with 40% of workers needing upskilling.

We are officially in a place where the stock market is no longer a reliable indicator for how well the entire US economy is actually performing.

So what happens next? I see three broad possibilities:

Cooling, not crashing. AI hype deflates while a shiny new distraction grabs investor attention.
Startup dominoes. A major AI startup collapses or IPOs poorly, sparking a chain reaction across venture portfolios.
Broader Crash. AI’s correction coincides with tariffs or political instability, tipping into recession.

And where does OpenAI fit into all this? For years I’ve worried that OpenAI’s unusual business structure (a nonprofit with a “for profit” branch) and market dominance could destabilize the entire industry if they collapsed.

Now, I’m less concerned. Competition is heating up:

Google’s Gemini and NotebookLM keep winning praise.
Meta has invested billions in talent, even if hiring is frozen.
Anthropic’s Claude models are beloved by engineers.
Perplexity is emerging fast with its new Comet browser.

In short: unlike a year or so ago the AI ecosystem probably no longer lives or dies with OpenAI (yay competition!).

But it does lead me to speculate that if OpenAI does collapse, that only the AI startup ecosystem will suffer, while Big Tech companies gobble up what’s left of AI startups.

If you’re feeling overwhelmed, you’re not alone. But history tells us that financial downturns often result in the creation of some of the most impactful companies.

Here’s what I recommend:

Invest in yourself. Take courses, learn new skills, network and be adaptable.
Watch where the puck is going. Don’t go on autopilot. Pay attention to market signals and what’s happening in your industry.
Remember: downturns breed startups. Many iconic companies were born in recessions.
Keep your portfolio balanced. My mom scolded me for suggesting AI stock picks a few weeks ago and she’s right. If you’re over-leveraged in tech, consider balancing your portfolio.

When I tell people I work in AI and they react negatively, I don’t take it personally. They’re sensing the tension that comes when hype outruns reality. It’s the same tension we saw in the run-up to past tech bubbles like crypto and the internet bubble.

Right now, we’re seeing a a mix of fear, FOMO, and magical thinking is pushing money into companies that may never succeed.

All I can do, like any of you, is do my best to watch where the puck is going, invest in myself, and make hard decisions in challenging environments.

What do YOU think? Are we in a bubble or is all this spending actually realistic? Let me know in the comments!

Bubble Bubble Toil and Trouble | What to Do if the Bubble Pops

Comments