“Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes” declares Sam Altman, CEO of Open AI, during a recent interview with The Verge and other reporters in San Francesco.
The tech leader is not the first high-profile figure to issue a bubble warning. During the last few months, financial institutions, venture capitalists and investors have begun rising concerns about the excessive expectations related to AI growth, questioning a potential over-hype.
But what exactly is an economic bubble? Should we be worried about what is going on?
A bubble occurs when investors, expecting to make profits from an asset resale in the future, increase the demand for a good by inevitably causing its price to rise. However, the cost increase becomes so extreme that the price no longer reflects the intrinsic value of the asset. For a bubble to pop, a collective realization is required that valuations far exceed their worth: this recognition reduces demand, pushing prices back down to more reasonable levels.
More and more, the perception of AI as a bubble is entering mainstream discussion. Long presented as an unstoppable growth sector, AI is now increasingly perceived as an unfulfilled promise. As an example, the recent launch of GPT-5 has been immediately followed by a social media boom of content showing limitations and mistakes committed by this new model, despite widespread expectations that it would deliver groundbreaking improvements similar to previous versions of ChatGPT.
As explained in the article “Is AI hitting a wall?” published in the Financial Times, companies such as OpenAI have managed to show consistent gains in the performance of their systems by increasing data and computing power to build more advanced models. However, this approach is starting to reach the limits of the available resources and, as Altman himself noted, chatbots like ChatGPT are “not going to get much better”.
In this context, a useful reference is the 2024 Goldman Sachs research paper “Gen AI: too much spend, too little benefit?” which investigates whether the massive spending on Generative AI infrastructure will ever translate into meaningful benefits and returns. Interestingly, the GS analysis points out how corporations perceive artificial intelligence as a groundbreaking tool, even though the technology often proves unreliable and fails to deliver the expected boost. “AI technology is exceptionally expensive, and to justify those costs, the technology must be able to solve complex problems, which it isn’t designed to do” – affirms Jim Covello, Goldman Sachs Head of Global Equity Research.
Inaccurate outputs, privacy concerns and intellectual property violations do not seem to align with the expectations that we – whether investors or ordinary users – had for these new technologies. Moreover, AI has contributed to rising emissions in many industries, pushing several carbon-neutrality goals further into the future. In addition, the energy-intensive process of training and running AI models keeps raising serious concerns over environmental impact and overexploitation of water resources.
Beyond its effects on nature, it is important not to overlook the impact that AI has on labour: an increasing fear spreads in work places where people are more aware of how easily AI can replace their job, while others face wage stagnation or even decline. The creative industries represent a useful example to illustrate this trend: editors are increasingly favoring AI illustrations, which are far cheaper than an artist’s creative output.
But apparently, not every cloud has a silver lining. Indeed, the Goldman Sachs report asserts that even though AI doesn’t “deliver on its promises”, it may still create value. “Someone is going to lose a phenomenal amount of money. We don’t know who, and a lot of people are going to make a phenomenal amount of money” – Altman explains according to the American online magazine The Verge – My personal belief, although I may turn out to be wrong, is that, on the whole, this would be a huge net win for the economy.”
Does it mean that an economic bubble can be beneficial?
Well, some experts see this financial phenomenon as positive since bubbles could be the reflection of massive opportunities opening up. As a matter of fact, a growing body of research suggests that bubbles may have advantages – even if they result in massive losses for some market participants.
In “Boom: Bubbles and the End of Stagnation”, investors and writers Byrne Hobart and Tobias Huber investigate how bubbles can create the ideal conditions for transformative innovation. In particular, they argue that a societal aversion to risk, shaped by demographic changes, has led to economic stasis. This problem, they suggest, can be overcome through a degree of financial exuberance and irrational investments which may help to break free from this stagnant trap and ultimately benefit society as a whole.
Remarkably, the two authors are not alone in acknowledging that the social benefits of bubbles can outweigh the costs borne by investors. Similarly, the American venture capitalist Bill Janeway supports the idea that “productive bubbles” exist. From his point of view, unsustainable speculation could generate social benefits by fostering the exploration of new technological applications – even though most of the ventures involved eventually collapse.
In conclusion, the only certainty is that the future remains unclear. Whether we are heading toward financial shakeouts or long-term social benefits, all we can do is keep a close eye on the tech world and stay informed.
Curious to learn more? Check out these articles:
- “Is AI Hitting a Wall?” – Financial Times (https://www.ft.com/content/d01290c9-cc92-4c1f-bd70-ac332cd40f94)
- “Would an Artificial-Intelligence Bubble Be So Bad?” – The Economist (https://www.economist.com/finance-and-economics/2025/01/02/would-an-artificial-intelligence-bubble-be-so-bad)
- “Watching the Generative AI Hype Bubble Deflate” – Cornell University research paper (https://arxiv.org/pdf/2408.08778)
