MMOOZZGG.com

AI investment

AI investment in Europe: deals and startups of 2026

5 min read3 sources

AI investment in Europe was $20.92 billion in 2025, according to Stanford HAI. In Q3 2026, two European startups, Mistral AI and Nscale, raised more than $3 billion each.

On this page · 5 sections

Europe in numbers

In its AI Index 2026, Stanford HAI counted $581.69 billion of corporate AI investment worldwide in 2025. Europe took $20.92 billion, the US $285.88 billion and China $12.41 billion.

By this figure, Europe is ahead of China but far behind the US. By our count, Europe has less than one twentieth of the world total.

The Stanford HAI count includes private investment, $344.66 billion worldwide in 2025, and mergers and acquisitions, $214.44 billion. State programs that support AI are covered on our page on sovereign AI of states.

For a company that follows the European market, the main point of these figures is scale. By our count, Europe’s annual sum is less than one tenth of the US one, so European startups often take money from US funds, as in the CuspAI example below.

Large rounds of 2026

In Q3 2026, Crunchbase noted two European rounds of more than $3 billion each:

1. Mistral AI, a French developer of language models.

2. Nscale, a UK company that builds computing capacity for AI.

They are on the same list as Databricks and Safe Superintelligence, which got $5 billion each, and Crusoe, Moonshot AI and Kling AI, which got more than $3 billion.

Nscale invests not in models but in computing capacity: data centers to train and run AI. Money goes to this segment because models do not work without capacity.

There are smaller but notable rounds too. CuspAI, a UK company that uses AI in materials science, raised $450 million in a Series B in July 2026 at a valuation of $2.6 billion.

CuspAI’s investors are Kleiner Perkins, NEA and Bezos Expeditions, all from the US. This is a typical picture: a European company grows on US money.

All rounds with amounts and sources are in our deal tracker.

Acquisitions

Stanford HAI valued worldwide mergers and acquisitions of AI companies in 2025 at $214.44 billion. Our sources have no separate sum for Europe, so here we show an example, not a total.

Example: Bending Spoons from Italy builds its business on buying other companies’ apps and rebuilding them. In Q3 2026 the company went public and raised $1.6 billion at a valuation of $18.4 billion.

The listing of a company that grows through acquisitions shows that the European market has a path to money other than venture rounds. Investors pay for the skill of buying and rebuilding products.

For a Russian owner, this model holds a practical idea: buying a ready team or product is sometimes faster than building your own. Before such a purchase, check what in the product actually works and whose data it uses.

AI company purchases worldwide in 2026, with amounts and buyers, are on our page about mergers and acquisitions.

The gap to the US

The money gap is large: $285.88 billion of corporate AI investment in the US in 2025 against $20.92 billion in Europe, according to Stanford HAI.

The deals show the reasons. Of the seven companies on Crunchbase’s list of the largest rounds of Q3 2026, two are European: Mistral AI and Nscale.

The second reason is the source of the money. CuspAI got $450 million from US funds, and many large European rounds work this way: the decision about the money is made in the US.

Europe catches up not with sums but with its choice of niches. Mistral AI makes models, Nscale builds computing capacity and CuspAI works in materials science, and each niche gives Europe its own place in the global AI chain.

Hence a rule for reading the news: a $3 billion European round means more for Europe than the same round means for the US, because Europe’s annual base is smaller.

What matters for Russia

Europe shows how a market with less money holds its position: $20.92 billion in 2025 against $285.88 billion in the US, and it still has companies with rounds of more than $3 billion.

For a Russian company, this leads to three conclusions:

1. Countries build their own models and computing capacity even where US ones are available. Mistral AI and Nscale raise money for exactly this.

2. A narrow niche gives more than an attempt to copy everything at once. CuspAI, valued at $2.6 billion, works in one industry.

3. Buying a ready product, as Bending Spoons does, can be faster than building from scratch. But a product is bought after a check: what in it works, and who owns the data and the code.

A practical step: before you invest in AI, decide which one task in your industry it should solve, and count the effect on that task. How Europe compares with China is covered on our page on the Chinese market.

Sources and fact-checking

Checked by the editors on 14 October 2026.

  1. Stanford HAI, AI Index 2026, chapter “Economy”, 13 April 2026 (hai.stanford.edu)
  2. Crunchbase News, 5 October 2026 (news.crunchbase.com)
  3. vc.ru, 2026 (vc.ru)

Found a mistake? Tell us

Publications

Mike Mozg

Agentic AI use cases at AT&T: nearly 500 processes now run on agents

AT&T no longer uses AI only to help staff work faster: it is rebuilding whole business processes around agents. SiliconANGLE reported on 14 November 2026 that nearly 500 agentic AI use cases at AT&T are already in production. Here is how the company rebuilds them, where a human still checks the agent, and what other large companies can take from it.

News4 min
Contact us