Worried About an AI Bubble? You're Asking the Wrong Question.

What the Bubble Debate Misses

There is growing nervousness about an AI bubble. Valuations are stretched. Investment has flooded in at rates reminiscent of the dot-com era. Sceptics point to companies with little revenue commanding massive market caps. The parallels are uncomfortable.

But focusing on valuations misses what matters. The question isn’t whether some AI companies are overvalued, they almost certainly are. The question is whether AI delivers real, measurable value to the organisations that adopt it thoughtfully.

The Capabilities Are Real

Unlike speculative bubbles built on promises, today’s AI tools already work. Large language models summarise documents, draft content, and answer complex questions across enterprise knowledge bases. Vision models inspect products, read documents, and navigate physical spaces. Agents are beginning to coordinate multi-step workflows with limited human supervision.

These aren’t prototypes. They’re in production at companies across sectors, delivering cost savings, productivity gains, and new capabilities that weren’t feasible two years ago.

Where the Risk Actually Lies

The risk isn’t that AI turns out to be useless. The risk is in how organisations approach adoption. Companies that chase hype, launch unfocused AI projects, or bolt models onto broken data infrastructure will waste money. Those that take a disciplined approach, starting with clear use cases and solid data foundations, will generate returns.

In the dot-com era, plenty of companies failed, but the companies that understood how to use the internet, Amazon, Google, Booking.com, became some of the most valuable in history. The lesson wasn’t “the internet was a bubble.” It was “the winners knew how to build real businesses on new technology.”

What This Means for You

If you’re waiting for the bubble to pop before engaging with AI, you’re making a mistake. The capabilities are here. The tools are accessible. Your competitors are already experimenting.

The right approach is to start small, prove value, and scale deliberately. Use AI where it solves real problems: reducing manual work, improving decision speed, or enabling services that weren’t previously economic.

Stop asking whether AI is overvalued. Start asking whether your organisation is positioned to capture the value AI can deliver. That’s the question that matters.