At the end of 2025, a deal made between Nvidia, producers of the advanced chips needed to train and run AI models, and OpenAI, the creators of ChatGPT, stirred some serious concerns of a potential AI bubble. Certainly, the deal did look suspiciously circular: Nvidia agreed to advance $100 billion of credit to OpenAI, effectively financing its own customers’ purchase of more chips. Observers quickly began speculating about the extraordinary concentration of capital in AI, illustrated by the fact that by 2025 nearly 50% of all venture capital funding went to AI-related companies. This concerned many onlookers in the general public, especially those who remembered the dot com boom and bust of the early 2000s, which experienced similar levels of hype (although nowhere near as much wealth). In 2025, a handful of AI-driven stocks represented about 80% S&P 500 gains (although that may be explained by turbulence among more traditional companies following tariffs). Furthermore, the practical hurdle of the high energy consumption necessary for training these models compounded concerns. Thankfully, concerns of any imminent correction settled down after a stellar earnings report from Nvidia, but some speculative fears remain.
The Nordics, ever so innovative, have invested heavily in AI. Due to the cool climate, cheap energy and forward-thinking entrepreneurs, Sweden in particular has become a globally-renowned AI hub, both because of companies like Lovable, Legora, Tandem and many other AI unicorns, as well as infrastructure like data centres. It goes without saying then, that the Nordics could be severely affected by the sharp deflation if fears of a market correction are borne out. However, there are reasons to believe that AI isn’t a “bubble” at all in a traditional sense, and the hype is based on real metrics which support these incredible investment levels. Indeed, not all bubbles are wholly bad, although they may hurt in the short term. First, let’s understand the evidence.
To Bubble, or Not to Bubble?
The case for a “bubble” that may one day pop has been outlined above, but to add to that there are also concerns that hype is running ahead of reality which is somewhat echoed in data on enterprise adoption. A late-2025 MIT Media Lab study found that despite $30 to 40 billion invested in enterprise “GenAI” projects, 95% of pilot projects failed to deliver any impact on bottom lines, based on 150 interviews with corporate leaders, a survey of 350 employees, and an analysis of 300 AI deployments. However, the pitch to investors is not necessarily that AI will make immediate returns for companies, but rather that they need less people to do the same thing, thus improving margins. Aditya Challapally, the lead author of the report, pointed out that startups led by 19- or 20-year-olds, for example, “have seen revenues jump from zero to $20 million in a year.” But for established businesses, revenue gains from AI implementations would require layoffs, which are unpopular, and as the technology still produces the best results in partnership with humans, organisations have adopted AI as an additional tool rather than a replacement. There are indications that the number of entry-level jobs on the market have decreased, and so joblessness among recent graduates have conversely increased, however it’s hard to know if AI is the only factor driving these numbers, meaning more time will be needed before that conclusion can be drawn.
However, there’s another way to look at it. Indeed, there are a number of concerning factors, but not all bubbles are bad. To use a previous example, while the dot com crash was devastating to some, it’s not like the internet faded away. The opposite, it is today way more valuable than before or during the crash, the hype had just run ahead of where technology actually was. Even if there is a change in investor sentiment that gets labeled a correction, or worse, a “bubble,” the technology itself will remain and the market will bounce back. There are also factors that are completely different from the dot com bubble, which do suggest this is not an overheated market. Unlike the dot com boom, earnings growth have nearly kept pace with stock prices, many AI companies (although not all) have proven business models and cashflow, the infrastructure built such as data centres will endure as valuable assets, and although early adopters of the technology haven’t proved increased earnings as of yet, the practical adoption is still rapidly expanding. Of course, there are some companies with absurdly high valuations and very little to show for it, but these are private companies backed by very well-capitalised venture and growth investors. In the event of any correction, larger players would have to rethink their model training and development budgets, while some smaller startups who haven’t yet secured necessary funding or found meaningful traction may shutter. The high concentration of investment into one industry remains a point of concern, although that is a different problem.
What It Means for Us
As illustrated, it is a bit more complicated than simply asking “is it a bubble?” but for the sake of argument and to more effectively illustrate possible outcomes, this section will treat the issue as such. In a hypothetical scenario where the sector experiences sharp valuation deflation, one of the outcomes would be a surge of layoffs within the Nordic tech sectors and a squeeze on startups that could hamper growth for a few years. These skilled engineers, which the Nordics already have a shortage of, may then be sucked up by bigger, richer countries like the US. Furthermore, there have been significant investments in data centres in the Nordics, which may then become stranded infrastructure if fewer customers are paying for the latest servers and chips. Indeed, some analysts already warn of potential overbuild, which in a downturn could become even more pronounced. Of course, as with any correction, the wider economy will eventually recover, but any immediate pain could be particularly acute in the Nordics.
On the other hand, if the boom continues (and hopefully with healthy levels of investment), the Nordics are poised to benefit massively. Their natural advantages, in terms of natural cooling and abundant, cheap energy, will likely attract more investors as time goes on. The Nordics already punch well above their weight in digitalization, number of startup unicorns and their generally high quality of life. All of these factors attract new talent who then go on to start their own companies in the region, as well as keep its very entrepreneurial populations in tune with technological advances. A Deloitte survey found Nordic executives actually expect slower short-term ROI from generative AI than many of their European peers, suggesting they are in it for the long-haul rather than chasing quick returns.
Whether you love or hate AI, whether it is a bubble or not, the technology is here to stay, and Nordic countries are counting on it. The interesting position the Nordics seem to find themselves in is not found in the balance between promise and peril, but rather as an example to the rest of the world how AI can be integrated into society. The high rates of digitalization and digital literacy combined with high levels of trust in institutions, such as their generous welfare systems, perfectly positions them to absorb the hit of any future market correction, but also to lead Europe and succeed on a global level if AI delivers the automation it promises. If AI were to succeed, the mass layoffs predicted by its creators would have serious consequences in the US for example, where your healthcare is more often than not tied to your job. Pitching their utopian vision, some forward-thinking AI founders hope that Universal Basic Income could solve the problem of AI directly taking jobs, but that transition is much likelier to occur smoothly in the Nordics than practically any other place on earth. While the discussion around worrying levels of capital concentration is warranted, it may be more important to consider long-term planning for a world where AI has succeeded in automating away more jobs than can be replaced.