Let’s be honest, the amount of money being funnelled into artificial intelligence right now is staggering. It seems every company, from the behemoths of Silicon Valley to the smallest start-ups in a Shoreditch basement, is either “AI-powered” or raising a king’s ransom to become one. We’re told it’s the next industrial revolution. But as the valuations soar and the hype machine works overtime, a nagging question keeps cropping up: Are we witnessing a genuine technological breakthrough, or are we inflating the next great AI investment bubble?
It’s a question that brings on a serious case of déjà vu. The giddy excitement, the promises of a remade world, the fear of being left behind—it all smells a bit like 1999. Back then, dot-com was the magic word. Today, it’s AI. And as we’re reminded by recent market analysis, just seven AI-related companies now account for over a third of the entire stock market’s value. When so much capital is concentrated on so few, you have to start asking the hard questions.
Déjà Vu All Over Again? Spotting the Bubble
So, what exactly is an investment bubble? Think of it like this: it’s when the price of something—be it Dutch tulips in the 1600s or a dot-com stock in 1999—gets detached from its actual, fundamental value. Instead, prices are driven by speculation and the belief that you can always sell it to someone else for more money. It’s a game of musical chairs with billions of pounds at stake.
The parallels between now and the late ’90s are difficult to ignore. The narrative is intoxicatingly similar. Back then, the internet was going to change everything. Companies with little more than a snazzy website and a “.com” in their name were achieving fantastical valuations. Today, the promise is that AI will automate work, create unimaginable efficiencies, and solve humanity’s biggest problems.
The key ingredients for these boom-bust cycles in technology economics are always the same: a genuinely exciting new technology, a flood of investment capital looking for the next big thing, and a media ecosystem that amplifies the hype to a fever pitch. We are currently three for three. The real test, however, is whether the underlying value can ever catch up to the sky-high expectations.
The Two Faces of AI: Co-pilot or Redundancy?
The central confusion in the AI debate comes down to one simple question: What is this technology for? Is it here to make humans better, or is it here to replace them? The answer you choose dramatically changes how you approach innovation valuation.
As Cal Newport points out in his recent piece for The New Yorker, AI tools like ChatGPT and Claude are proving to be phenomenal accelerators for human learning. He describes teaching his son to code at a pace that would have been unthinkable just a few years ago. In this model, AI isn’t the worker; it’s the ultimate tutor, a force multiplier for human capability. This creates a fascinating problem for traditional companies: what do you do when your employees can upskill themselves out of their current roles in a matter of months?
Yet, the boardroom narrative, and certainly the one driving much of the investment, leans heavily towards replacement. The dream being sold isn’t about creating a more skilled workforce; it’s about cutting headcount. Look at the case studies OpenAI proudly presents, boasting of companies saving “hundreds of millions of dollars annually” by replacing human call-centre staff with AI agents. Consider Microsoft, which is asking large companies to fork out “many millions of dollars each year” for its Copilot system. The pitch is pure cost reduction. This is AI as an efficiency tool, not an empowerment engine.
Valuing Vapourware: The Great AI Unknown
This brings us to the trickiest part of the whole equation. How do you put a price on a technology whose fundamental nature is still a subject of fierce debate among the scientists creating it?
During the dot-com bubble, for all the nonsensical business plans, everyone at least understood what the internet was. It was a global network of computers. The applications built on top of it—e-commerce, search, social media—were up for grabs, but the foundation was solid and understood.
AI is different. As Newport notes, there is profound scientific uncertainty about its trajectory. Will models continue to improve indefinitely with more data and computing power, or will they hit a wall? Can they ever achieve true reasoning, or are they just breathtakingly sophisticated mimics? Nobody knows for sure. Investing in AI today isn’t like investing in a railway when you know what tracks and trains are. It’s like investing in a transport company when the inventors haven’t decided if they’re building a better horse, a car, or a teleportation device. This makes any genuine attempt at innovation valuation feel more like astrology than finance.
The Cassandra of Code: Are We Listening?
Against this backdrop of hype and uncertainty, some voices are urging caution. The writer and activist Cory Doctorow puts it bluntly: ‘A.I. is a bubble and it will burst. Most of the companies will fail. Most of the data-centers will be shuttered or sold for parts.’
Is this just needless pessimism? Perhaps not. The current investment frenzy seems to be pricing in the most optimistic, world-changing outcomes—AI as a universal replacement for human labour. Yet, the most compelling, real-world evidence we have today points to a different, albeit still powerful, conclusion: AI’s greatest strength is in augmenting human intelligence, not supplanting it.
For businesses navigating this terrain, the risk isn’t necessarily in using AI. The risk is in making enormous, singular bets on the “replacement” theory. The smart play seems to be focusing on AI sustainability by investing in tools and training that make your people better, faster, and smarter. This approach builds real, lasting value, regardless of whether the grander prophecies of AGI (Artificial General Intelligence) ever come to pass. It’s a strategy grounded in the present reality of the technology, not the speculative fantasy fuelling the AI investment bubble.
The future of technology is notoriously difficult to predict. Experts at IBM famously forecasted a world market for “maybe five computers” back in 1943. But we can act on the evidence we have. Right now, the evidence shows that AI is a powerful tool for enhancing what people can do. The market, however, is betting billions that it’s a tool for replacing them. That a gap exists between reality and valuation is clear. The only question is how, and when, it will close.
What’s your take? Is your organisation using AI to empower its people or to shrink its payroll? And which strategy do you believe will truly create value in the long run?
