AI investment has reached record levels, yet financial analysts increasingly warn of a forming investment bubble in the industry. Tech companies are pouring hundreds of billions of dollars into building AI data centres, using debt financing and complex circular-investment schemes.
Nvidia and OpenAI: a record $100B deal
AI chip maker Nvidia announced its intention to invest up to $100 billion in OpenAI for the construction of data centres. However, the structure of this deal is raising questions among experts. OpenAI will use the funds received to purchase Nvidia chips and fill the data centres with equipment from the same manufacturer.
Paul Kedrosky, venture capitalist and economist, commented on the scheme: Nvidia is providing funds to OpenAI so that OpenAI can buy Nvidia’s own products. Similar practices were widespread at the tens-and-hundreds-of-billions scale during the dot-com bubble.
Circular investment in the AI industry
Beyond Nvidia and OpenAI, less prominent companies are participating in complex investment schemes. CoreWeave, formerly a cryptocurrency miner, has pivoted to building AI data centres. OpenAI signed multibillion-dollar agreements with CoreWeave to rent compute capacity in exchange for CoreWeave shares, then uses those same shares to pay the rent.
Financial analysts note that this circular investment structure artificially inflates the real demand for AI technology. OpenAI plans to invest $1.4 trillion in data centres over the next eight years against annual revenue of $20 billion.
Warnings from Nobel laureates
Daron Acemoglu, MIT economist and 2024 Nobel Prize laureate in economics, expressed scepticism about current AI investment levels. In his view, AI models are overvalued, and the industry is putting in more than it should.
Research shows that most companies do not see a chatbot impact on financial results. According to the analysis, only 3% of users are willing to pay for AI services. OpenAI CEO Sam Altman acknowledged the market overheating in August, noting excessive investor enthusiasm.
Comparison with the dot-com bubble
Twenty-five years ago the original dot-com bubble burst after debt financing was used to lay fibre-optic cables for a future that had not yet arrived. Analyst Daniel Luria warns of history repeating: if it turns out a few years from now that hundreds of billions of dollars were spent on data centres that aren’t needed, this could lead to a new financial crisis.
Hedge-fund investor Michael Burry, famous for his bets against the 2008 housing market, is now betting against Nvidia. He accuses the AI industry of accounting tricks and highlights the problematic nature of circular deals between companies.
Admissions from tech giants’ executives
Leaders of the largest tech companies openly admit irrational elements in the AI market. Google CEO Sundar Pichai told the BBC that there are irrational elements in the AI market, adding that should a bubble burst, no company — Google included — would be sheltered.
Nvidia CEO Jensen Huang attempted to dispel bubble concerns at a recent investor meeting. Nvidia’s share price has risen 300% over the past two years, making the company one of the main beneficiaries of the AI frenzy.
Tech companies continue to invest billions in AI infrastructure despite growing analyst warnings about industry overvaluation and possible consequences for the global economy.
You may also be interested in:
- Huawei and DeepSeek-R1-Safe: a tech view on safe AI
- Construction-robot news in the 2025 building industry
- The OpenAI Sora 2 project: a revolution in short AI videos

Comments on this article