The bubble isn't where you think it is

Everyone's asking me the same question right now. "Seb, is AI a bubble?"

My answer fits in one sentence. Yes, there's a bubble. No, it's not on the technology. It's on valuations.

These two things have nothing to do with each other. And confusing them is the best way to make a bad decision this year.

History repeats itself

In 1840, England went crazy for railways. Investors rushed into railway stocks. The bubble burst, fortunes vanished. But the rails stayed. They transformed the country for a century.

In 2000, same thing with the Internet. The Nasdaq collapsed. And yet, the Internet really did change the world. The financial bubble and the technological revolution were both true at the same time.

AI is exactly that. The revolution is real. The financial over-investment too. The question isn't to pick a side, it's to understand that both coexist.

What's real

The revolution is here. I see it every day, in the field, with our clients. RAG pipelines in production with real uptime. Agents automating business workflows that used to cost tens of thousands of euros per year. Automated code review in banking. It works. This isn't a Twitter demo, it's engineering in production.

When a tech gives you 20% savings on support costs or cuts your dev timelines by three, that's not hype. That's competitive advantage. And it's being captured now.

What's crazy

Now, look at the market.

Big Tech will spend about $660 billion on AI this year. Alphabet announced up to $185 billion in capex, almost double last year. Immediate result: sell-off, because nobody knows when all that money will turn into profits.

Nvidia hit $4.5 trillion market cap. The top five S&P 500 stocks represent 30% of the index, the highest concentration in half a century. And most troubling are the circular deals: Nvidia invests in OpenAI, which orders Nvidia chips, financed by Microsoft cloud, which owns 27% of OpenAI. Money goes in circles and inflates demand that looks more solid than it is.

Meanwhile, a February NBER study shows that 90% of companies still see no measurable impact of AI on their productivity.

The disconnect is there. The tech creates real but modest value, today. The markets have already priced in the complete revolution, the one from 2035.

My conviction as a leader

Here's my point. The right question isn't "should we believe in AI?". It's "where to put your money?".

My answer: invest in technology, not in speculation.

Investing in technology means building durable capability in your company. Equipping your teams, industrializing your POCs, putting real AI into production on your processes. That return, you control. It's measurable. It doesn't depend on Nasdaq's mood.

And above all, do it agnostically. Don't marry any model, any vendor, any platform. Build your stack to be able to swap engines tomorrow without redoing everything. Those who bet everything on a single player in 2023 are paying for it now.

Because there's good news nobody's watching: model prices are collapsing. For equal intelligence, inference today costs a fraction of what it did eighteen months ago, and it keeps dropping. This is the best time in history to build. You capture the power while the price crashes. Take advantage, stay portable, and let vendors fight each other on pricing.

Investing in stocks at the top of the cycle is the opposite. Betting that already extreme multiples will climb even higher. That, you don't control.

What to remember

The revolution is obvious. The financial investment is excessive. These two truths coexist, like railways in 1840 and the Internet in 2000.

The trap, for a leader, is mixing everything up. Either you get euphoric and overpay for anything with "AI" in its pitch. Or you get cynical, cry bubble, and miss the only technological wave of your generation.

The way out is discipline. Adopt the tech fast, agnostically, taking advantage of low prices. Pay the valuations slowly. Put your energy where ROI is measured in weeks on your own operations, not in promises on a ticker.

The bubble will eventually deflate. The revolution will continue. Like the rails after the crash. Position yourself to survive the first and profit from the second.