Anthropic’s IPO could tell us something important about the future of AI investments: how much are investors willing to pay for AI growth, and how long are they prepared to wait for returns?
If Anthropic’s share price falls because public-market investors decide that the economics of frontier AI do not justify its valuation, the consequences could extend beyond the stock market. Venture capital investors could become more cautious, valuations could fall and funding rounds could become harder to close. Startups that rely on successive rounds of investment to develop their technology and grow, could find themselves under pressure.
A fall in the share price would not automatically trigger an AI investment winter. But it could expose a weakness in Europe’s AI ambitions that already exists: Europe has far less capital available to finance its AI companies than its main competitors.
And in AI, that is not just a funding problem. It is a technology problem.
Europe can build AI startups. But can it keep up?
Europe does not lack AI talent, research or promising companies. Its challenge is financing those companies as they move from early-stage development to global scale.
The European Central Bank estimates that US venture capital funds hold around €930 billion in assets, compared with approximately €150 billion for EU-based funds. The difference is particularly consequential at later stages. According to the European Commission’s EU Startup and Scaleup Strategy analysis, European late-stage venture capital investment amounted to around $21.3 billion in 2024, compared with approximately $133 billion in the US.
These are not just different levels of investment. They represent different capacities to finance companies through the expensive stages of growth.
AI companies need capital to acquire customers, expand internationally, recruit specialised talent, secure computing capacity and improve their models. Frontier AI developers face especially high costs as they train and deploy increasingly capable systems. Even companies building applications on top of existing models need funding to establish distribution, integrate into customer workflows and develop defensible products.
The technology race makes this more urgent. Stanford’s 2026 AI Index found that leading US and Chinese models had converged in performance, with the leading US model ahead of its Chinese counterpart by just 2.7% as of March 2026.
Europe has yet to produce frontier models at the same scale and frequency as the US and China. Its funding constraints make closing that gap harder.
When capital is scarce, companies may have to delay model training, limit computing expenditure, slow research or abandon international expansion. These decisions can compound over time. Less capital means less capacity to improve the technology; weaker capabilities can make it harder to attract customers and investors; and that, in turn, makes raising the next round more difficult.
Europe risks falling behind twice: first in its ability to finance AI businesses, and then in the technology those businesses can develop.
AI sovereignty needs more than regulation
Europe says it wants greater control over its AI capabilities, infrastructure and strategic technologies. That ambition is understandable. AI will influence productivity, industrial competitiveness and the future of many critical services.
But sovereignty cannot be achieved through regulation alone.
Research, computing infrastructure, access to data and clear rules all matter. So does the ability to finance companies that turn those assets into products used at scale. Without that final step, Europe risks developing strong research and regulatory frameworks while the most valuable companies and capabilities emerge elsewhere.
Foreign investment is not inherently a problem. International capital can help European companies grow, and European founders should be able to access it. The risk is having too few financing options at home, particularly when companies need large amounts of capital to compete globally.
When a company depends on foreign investors for successive funding rounds, those investors may ultimately influence where it expands, which markets it prioritises and where strategic decisions are made. That is not an inevitable outcome of foreign investment, but it is a risk when domestic alternatives are limited.
If Europe wants to retain a meaningful role in the AI economy, it needs more than the ability to regulate technology developed elsewhere. It needs companies capable of building and owning important parts of that economy.
What needs to change?
Europe needs to mobilise more capital for AI companies, particularly those approaching the scale-up stage.
First, institutional investors need a greater role. Pension funds, insurers and other long-term investors can provide capital at a scale that traditional early-stage venture funds often cannot. Regulatory barriers, risk perceptions and the structure of investment vehicles all influence how much of that capital reaches venture-backed companies. Addressing these constraints should be part of Europe’s growth agenda.
Second, public policy needs to treat late-stage AI financing as a strategic investment area. Public capital can help mobilise private investment through co-investment vehicles, guarantees and incentives designed to bring institutional investors into the market. The objective should not be to fund every AI company or protect weak businesses from failure. It should be to ensure that promising companies have a credible path to scale.
China offers one example of a different approach. According to Stanford’s 2026 AI Index, Chinese government guidance funds deployed an estimated $184 billion into AI firms between 2000 and 2023. Europe does not need to replicate China’s model. But if AI is genuinely strategic, the financing system must reflect that ambition.
Third, private investors need to distinguish between different parts of the AI market. A frontier model developer with enormous infrastructure costs has a different risk profile from an application company using existing models to solve a specific business problem. A correction in valuations for one category does not necessarily mean the entire sector has lost its potential.
Finally, founders need to build businesses that can withstand more selective funding conditions. Technical capability matters, but so do commercial evidence, improving unit economics, customer adoption and defensibility beyond access to an underlying model. Companies that can demonstrate these qualities will be better positioned if capital becomes more expensive.
A more selective market would not necessarily be bad for AI. It could reduce speculative valuations and force companies to demonstrate real value. The danger is that promising European businesses are starved of capital not because their opportunities are weak, but because the financing ecosystem cannot support them at the scale required.
Europe needs to walk the talk
Anthropic’s IPO will not determine the future of AI investment on its own. But the market’s response could test how willing investors are to finance the next stage of AI development, particularly if expectations around growth, costs and profitability change.
For Europe, that test comes with an existing disadvantage. Its AI ambitions depend on companies being able to compete globally, yet its funding ecosystem remains much smaller than that of the US.
Europe cannot guarantee that its AI companies will win. Nor should it try to eliminate the risks inherent in building new technology. But it can decide whether its financing system gives those companies a realistic chance to compete.
If Europe wants AI to be a source of economic growth and strategic independence, it needs to mobilise capital accordingly. Otherwise, Europe may end up regulating an AI economy whose most important capabilities, companies and investment decisions are controlled elsewhere.
Member discussion