Is AI overvalued? The question of whether the sector is heading toward a bubble is one of the biggest questions in tech today. While there is growing concern that the U.S. market may be overvalued, Europe could prove a more resilient structure.
According to the Stanford Institute for Human-Centered AI (HAI), the U.S. market collects the bulk of private AI investment, raising $109.1 billion USD in 2024. Meanwhile, key European countries including the UK, Germany, France, Sweden, and Switzerland raised just $63.71 billion combined.
Lower investment across Europe, however, could mean less of a chance for startups becoming overvalued. Writing in the Financial Times, AI researcher Marietje Schaake warned that the U.S. market has pursued a “hyper scale model,” creating a resource-intensive AI platform bubble which “cannot last.”
Europe, by contrast, has opted for a more regulated approach toward development, producing legislation such as the landmark EU AI Act, which promises to reduce risk. At the same time, the EU plans to invest €1 billion per year in AI, which signals long-term controlled growth rather than speculative and explosive expansion.
Is there an AI bubble?
Although the existence of an AI bubble is up for debate, increasingly more tech leaders have acknowledged that AI might be overvalued, including Sam Altman, co-founder and CEO of OpenAI, Demis Hassabis, CEO and co-founder of Google DeepMind, and Bill Gates, co-founder of Microsoft.
Ultimately, the presence of a bubble comes down to whether or not investment in AI exceeds its true worth. With OpenAI valued at $500 billion USD while failing to provide a profitable business model- particularly around large language models (LLMs)- the disconnect with investor hype and true results is difficult to ignore.
“There is a fair bit of evidence that some of the valuations of the companies in the AI space have grown exponentially high, and there are concerns that these ‘round-trip deals’, where investments flow back to original investors, are over inflating the market,” noted Nate MacLeitch, CEO of Quickblox, a provider of chatbot and video consultation APIs and SDKs, while in conversation with 150sec.
“However, this was also common with internet deals back in 2001 and here we are in a massive tech boom 20 years later,” he added.
MacLeitch pointed to a large-scale UK National Health Service (NHS) AI physiotherapy pilot that reduced waiting lists by 44% and back-pain wait times by 55%, demonstrating where AI delivers measurable value.
“There might be corrections with financial valuations and talks with 95% [of] products failing, but there are cases where AI is transforming workflow and efficiency. The large-scale NHS AI-enabled physiotherapy pilot delivered in February is one example of the continent’s success.” he said.
Here, while there are areas in which AI demonstrates measurable value- and others where utility is less clear- providers that demonstrate it will be able to stave off concerns about overvaluation.
Europe’s resilience to the AI bubble
When considering the bubble, it’s important to note market differences. Although there are similarities between the U.S. and EU markets, they are each subject to different economic pressures and incentives. For one thing, the U.S. has access to much more private investment than its European counterparts.
Regardless, European startups like DeepL and Mistral AI have still managed to achieve unicorn status; there still is opportunity for top performers to attract high valuations, just not quite at the same scale seen with Silicon Valley firms like OpenAI and Anthropic.
“There are strong indicators in parts of the global AI market, particularly in the U.S., that valuations and investment flows outpace fundamentals. This looks very much like bubble dynamics: large sums chasing a limited set of proven business models. In Europe’s case, however, the picture is quite different,” Ali Yilmaz, co-founder and CEO of mental health platform Aitherapy told 150Sec.
“The overall size of the European AI investment ecosystem is relatively modest compared with the U.S. and China. That means Europe simply does not have the capital depth of speculative fervour required to inflate an AI bubble on its own. We may see pockets of overvaluation in specific startups or verticals, but Europe as a whole is too small and too fragmented to sustain a continent scale bubble that could burst with systemic impact,” Yilmaz added.
Embracing responsible AI and developing solutions in compliance with European regulation has the potential to pay off in the long-term, as providers opt for slow and sustainable AI development over an innovate at all costs mindset.
Why a bubble could affect Europe
Although there are many factors that suggest Europe could be resistant to a bubble, it’s important to note that many European companies are reliant not only on U.S. tech, but also on external investment- a bubble bursting in the U.S. would likely see access to funding declining across Europe.
At the same time, European AI has some core weaknesses that could be a problem, as signaled by Maria Sukhareva, principal key expert in AI at Siemens and author of the AI Realist Substack:
“Europe currently has no competitive AI and is unlikely to achieve it in the near future. It lacks the energy resources to power large data centres, EU restrictions are unwelcoming to training general-purpose AI models, and bureaucracy hampers innovation,” she told 150sec.
“In practice, Europe will rely on Chinese open-weight models and on purchasing American cloud services. If something were to happen to American AI such as major defunding, bankruptcies, or a political decision that cuts Europe off, the region would be affected immediately,” Sukhareva said.
The European AI ecosystem remains strong and has many strengths that make it resistant to overvaluation, but the potential for the AI bubble to burst can’t be ignored. Whether the market can avoid such a scenario remains to be seen.
Featured image: Getty Images via Unsplash+
Disclosure: This article mentions a client of an Espacio portfolio company.