Recently, OpenAI proposed to the US government to acquire a stake of about 5% of the company, opening a discussion that goes far beyond the boundaries of just another typical business deal and leads to the adoption of a new business-state model in the US that resembles that of China.
For the first time, one of the leading artificial intelligence companies is reportedly considering a model of direct government participation in a technology that is evolving at a pace without historical precedent since the Industrial Revolution.
The proposal raises a deeper question: will governments gradually be called upon to take on the role not only of regulator, but also of strategic investor, in a bailout program like those for banks in 2008-2009.
It is worth noting that China has formed a hybrid technological ecosystem of private and state-owned companies that develop competing language models, with the cost being paid for by the monstrous trade surplus of $1.2 trillion that it achieved in 2025.
The exponential increase in the demand for new resources
The creation of modern large language models (LLMs) no longer resembles software development as we knew it a decade ago.
Artificial intelligence has become one of the most “capital-intensive” sectors of the global economy. Training an advanced model requires hundreds of thousands of specialized processors (GPUs), state-of-the-art data centers, huge amounts of electricity, specialized fiber optic networks and teams of highly skilled engineers. The cost does not stop at training.
Every query submitted to a large model requires computing power, which drastically increases operating costs. AI models consume ever greater computing resources.
The Data Behind the AI Revolution
- Microsoft has committed tens of billions of dollars to OpenAI and the development of Azure AI infrastructure.
- Meta invests tens of billions annually in AI infrastructure, aiming to develop open models like Llama.
- Google is constantly expanding its data centers to support Gemini.
- Anthropic has secured huge funds from Amazon and Google.
- Elon Musk’s xAI is also investing in large-scale infrastructure, while also seeking additional funding to expand its computing capabilities.
The common denominator is one: all companies need more and more capital.
The Geometrically Expanding AI Industry
The industry’s biggest companies have entered a capital spending race that is more reminiscent of building national energy or rail networks than developing software – on a scale reminiscent of the 19th century Industrial Revolution.
- Microsoft said its investment in AI infrastructure is expected to reach about $80 billion in a fiscal year, mostly for new data centers and artificial intelligence equipment.
- Meta has increased its capital spending budget for 2025 to $64–72 billion, with much of it going toward building supercomputers and AI facilities.
- Amazon, through AWS, is also investing tens of billions of dollars annually in cloud and AI infrastructure, and has invested about $8 billion in Anthropic.
- Alphabet (Google) has announced investments exceeding $75 billion for data centers, TPU processors and the development of Gemini models.
- OpenAI, although not listed, has a valuation approaching $300 billion, while it continues to seek new capital to expand its computing infrastructure.
- Anthropic is valued at over $60 billion, while
- Elon Musk’s xAI has also reached a valuation of tens of billions of dollars in a short period of time.
In total, the investments announced by large technology groups for artificial intelligence infrastructure already exceed $300 billion annually, not counting private investments in start-ups, government subsidies and the energy infrastructure required to power the new data centers.
US State Involvement
The potential state involvement in OpenAI cannot be considered independently of the geopolitical competition with China. Washington now treats AI in a manner similar to that it has previously treated nuclear technology, the internet, the development of nuclear weapons or semiconductors.
Artificial intelligence is considered critical for economic development as well as for defense, cybersecurity, intelligence gathering, the development of autonomous weapons systems and scientific research.
A potential US State involvement in cutting-edge companies could ensure greater influence over their strategic decisions, without requiring immediate nationalization.
This is not a bailout in the traditional sense – but what could it mean?
So far, there is no evidence that OpenAI or other major AI companies are on the verge of financial collapse, but there is widespread speculation that there is a huge capitalization bubble that is not supported by the founding giants, but also by private credit.
On the contrary, they continue to attract significant private capital and are valued at particularly high levels.
However, their business model is characterized by a paradox: as the use of their services increases, so does the need for new investments in computing power and energy infrastructure. This creates a perpetual cycle of financing needs.
A new model for industrial policy?
The history of the United States shows that the state has repeatedly intervened in strategic sectors. The creation of the Internet relied on public funding through DARPA.
The development of the American aerospace industry relied on government contracts. The CHIPS Act mobilized tens of billions of dollars to boost domestic semiconductor production.
During the 2008 banking crisis, the TARP program prevented the collapse of the financial system through government intervention.
These cases differ significantly from each other, but they demonstrate that when an industry is considered critical to national security or economic stability, government involvement is not an exception.
Increased government involvement is not without risks. On the one hand, it can provide stability and access to long-term capital.
On the other hand, it raises issues of competition, concentration of power, and state influence in the development of the technology. In addition, the close interconnection of AI companies with government agencies may reinforce concerns about the use of artificial intelligence for military or surveillance purposes.
The big question
The proposal for US government involvement in the AI industry signals that the conversation has shifted: from how AI will be developed, to who will finance the ever-growing infrastructure it requires.
If costs continue to rise at the current pace, private investment may not be enough to support the next phase of the technological race on its own.
In such an environment, it is not excluded that we will see more forms of cooperation between the state and the private sector, whether through investments, tax incentives, or shared infrastructure.
What is certain is that artificial intelligence is no longer the exclusive domain of Silicon Valley. It is gradually becoming a field of national strategy, industrial policy, and geopolitical competition. The discussion about OpenAI may be the first sign of this new era.
Future Costs and Risks
If demand turns out to be significantly lower than expected after a period of overinvestment, large write-downs in data centers, GPUs, power purchase agreements, and software would be possible.
Assuming that 30–40% of the investments already made in AI infrastructure were to be written off, the total accounting losses could approach $1.0–1.5 trillion over a few years.
This is comparable to the losses suffered by US banks during the 2008 crisis, although it would mainly concern fixed assets rather than financial products.
If a TARP-like approach were implemented, where the government would acquire shares or convertible bonds instead of directly subsidizing losses, then the required financing could range between $200 billion and $400 billion.
Of this amount, a significant part could be recovered in the future through sales of participations, dividends or repayments, with the result that the final net fiscal burden would be significantly less than the initial disbursements.
The key difference with respect to the 2008 crisis is that systemic risk today does not come from banks’ balance sheets but from the concentration of huge investments in digital infrastructure and computing power.
Therefore, even if an AI “bubble” were to lead to losses comparable to those of the financial crisis, government intervention would likely be smaller as a percentage of the total losses, but it would still be a terrible blow to the US capital market.
Millions of small investors would lose their savings and jobs, while large insurance funds would be destroyed in an economic turmoil with multiple consequences from the financial crisis of 2008-2009.
It is no coincidence that the “tech bros”, the leaders of the largest companies, wanted to join Donald Trump’s closed system of power after his return to the White House in 2025, nor did Paladir’s (no longer) strange manifesto on the connection of national security with the development of AI.
