What sovereign AI is
It means models, data and computing capacity that a country controls itself and will not lose because of a decision by a foreign company or government. Our section on sovereign AI covers the concept, the models and the requirements in detail.
This page is about money: who pays for such independence. In some countries the budget pays; in others the state sets a target and gathers private capital for it.
Countries start from different positions. Per Stanford HAI, corporate AI investment in 2025 was $285.88 billion in the US, $20.92 billion in Europe and $12.41 billion in China.
Countries that lag in corporate money add their own tools: funds, tenders for computing centers and national programs. Below are the numbers for each region with a source.
The US and Europe
1. The US. On 21 January 2025, OpenAI, SoftBank and Oracle announced the Stargate project: $100 billion at once and up to $500 billion over four years for AI infrastructure in the US. This is private company money, not the federal budget.
2. The European Union. On 11 February 2025, the European Commission launched the InvestAI program with the goal of attracting 200 billion euros of AI investment. It includes a new European fund of 20 billion euros for AI gigafactories, large computing centers.
On 30 July 2026, EuroHPC opened a call to build up to seven such gigafactories in the EU. Expected private investment is more than 20 billion euros, and applications are accepted until 12 November 2026.
3. France. In February 2025, Emmanuel Macron announced 109 billion euros of private AI investment in France. The largest commitments came from the UAE, 30–50 billion euros, and from Brookfield of Canada, 20 billion euros.
4. The United Kingdom. The government set aside more than 2 billion pounds for AI computing infrastructure: over 1 billion pounds to expand the national research computing resource at least 20 times by 2030, and up to 750 million pounds for a supercomputer in Edinburgh.
In the US, France and the EU, private investors give or are expected to give most of the money, and the state sets the goal and the site. The United Kingdom pays from the budget. EU programs are covered on the page on AI investment in Europe.
China
In China, a state fund of 60 billion yuan for investment in the AI industry has started work. The Xinhua news agency reported the launch on 21 January 2026.
Stanford HAI puts corporate AI investment in China at $12.41 billion for 2025. That is less than in Europe, $20.92 billion, and many times less than in the US, $285.88 billion.
China’s model differs from Europe’s. The fund puts state money directly into companies of the industry instead of gathering promises from private investors.
This matters for a company in Russia that buys Chinese hardware or models: the state invests in its own producers, and that is worth counting when you choose a supplier. Deals and rounds of Chinese AI companies are covered on the page on AI investment in China.
Russia: the national project and the industry’s request
In Russia, the main channel of state money for digital technology is the national project Data Economy and Digital Transformation of the State. Per the explanatory note to the draft budget, it gets $1.9 billion in 2026, more than $2 billion in 2027 and more than $2.2 billion in 2028.
This is money for all digital transformation, not only for AI. So you cannot compare it directly with the AI programs of the EU or China.
Large Russian AI developers, per RBC, asked the state for $4.8 billion to $5.4 billion a year for AI development. They also proposed giving 5% of computing capacity to education and state bodies. There is no decision on the request.
The national project, the AI law and support measures for business are covered on the page on state support for AI in Russia.
What this means for companies
State programs change the conditions for companies even if a company gets no money from them at all. Three practical consequences:
1. Capacity moves inside national borders. Europe is building up to seven gigafactories, and Russia keeps a national project of $1.9 billion in 2026. If your AI works with data that must stay in the country, find out in advance where the capacity will be and on what terms.
2. Programs run on deadlines. Applications for the European gigafactories are accepted until 12 November 2026. In Russia, the industry’s request for $4.8 billion to $5.4 billion a year has no decision yet, so it is too early to put this money in your plan.
3. Independence costs money. Countries spend hundreds of billions on their own capacity for one reason: not to depend on another country’s decision. For a company, the same logic works on a small scale.
A practical step: list the company’s AI tasks and mark which of them run on foreign services. Start the move with tasks that process personal data or trade secrets.
Corporate money flows into AI without the state too: $581.69 billion worldwide in 2025, per Stanford HAI. Model and capacity suppliers also invest, so compare them by price and by the terms for leaving the contract, and do not take the first one.