AI bubble, bubble, toil and trouble: What does it mean for tech scaleups?

Last night we were treated to an evening of fine wine and art in the heart of London. It was our wine tasting event with tech clients, journalists and friends. The setting? Tick. Great company? Tick. Topical conversations? Tick. Too much wine? …. We’ll leave it there.

One topic not far from everyone’s lips (aside from the rich, vibrant notes of a vintage Pinot Noir) was AI. And specifically, the AI bubble. While mutterings about its emergence began over a year ago, the idea took hold when OpenAI’s CEO Sam Altman said we could be in an AI bubble. Essentially, this means heaps of unsustainable investment into AI. 

The bubble is being fuelled by vast swathes of money pouring into the physical infrastructure that supports AI – think datacentres, semiconductor chips, and so forth. The World Economic Forum (WEF) has written a great read on the subject, with comparisons to the dot-com craze at the millennium and ‘tulip mania’ in 17th century Netherlands. Yet it makes clear that no bubble is the same. 

Presently, the messaging on one side is that AI is propelling a radical revolution and the bubble discourse is misguided. The other side says that AI’s value varies and the pace of improvement is grinding to a halt – and that’s when the bubble could pop. 

So, which is it?

The value divide

Major infrastructure projects always create a bit of a bubble; people need to discover how much of the new infrastructure they actually need to use. Yet Big Tech has portrayed an idea that AI must be adopted at all costs, even if this adoption ends up being wayward and then requires a remediation of processes to decipher its benefits. 

One notable difference in AI’s value is between creative tasks and more objective tasks. In fact, it can become a hindrance to creative endeavours. From a PR perspective, one journalist last night was at pains to point out that they just want to hear insights from the expert and not a boggy swamp of AI slop. In this respect, we couldn’t agree more. Use it to support your opinions if necessary, not to make them for you. 

A compelling example shared in another article was about the perceived value of AI for software developers. Many believe that AI is making them more productive, but is that the reality? One study showed developers thought they completed tasks faster with AI support. Yet the “programmers who used AI took 19% longer, on average, than programmers who didn’t”. Another study, however, revealed it did increase productivity by a noteworthy 26-39%. 

The article’s probably on the money when it says AI tools speed up some tasks and slow down others, the key is learning which. We’re also still in the primitive stages of the technology; some tools will naturally become more sophisticated and advanced as time goes on. 

Startups and scaleups proving their value 

Whether the bubble bursts or not is out of startups’ and scaleups’ control. So, discussion around how to quell its expansion is perhaps irrelevant for them. What’s undeniable, however, is that we work with companies whose AI and technology does provide real, tangible value. 

Take Peak. Its AI and agentic AI platforms are able to analyse streams of customer and contextual data to optimise retailers’ inventory levels and the pricing of their products. Then, Agreena uses AI to analyse data and help farms prove how much carbon their soil sequesters. And in a tough funding climate, ICS.AI provides AI that helps councils triage requests and offer human support where it matters most.

In the same way, my finance software can comb through all of my transactions to provide me with top-level insights in a digestible format, saving me heaps of time. That’s helpful, and valuable. But many of the bold ideas and ambitions of Big Tech are far less tangible. It’s why the bubble is becoming a go-to topic of conversation. 

Time to show your value

Our potion of choice last night was a good glass of vino. But echoing the words of the Macbeth witches, if the AI bubble does double and double, there could well be toil and trouble. For AI startups and scaleups, this reemphasises the need to cast a light on the real value their technology brings to the market, to investors and to society at large. 

“Some bubbles leave behind more real value than others,” the WEF article states. A large part of that value could well be from the work of startups and scaleups, just as it was in the dot-com era when a startup called Google pulled through. 

Tell your story, show your value. (And then have a glass of wine if you like.)
Want to find out more? Get in touch with the team here.