Europe’s €30 Billion AI Gamble

On July 30, Europe formally opened bidding for one of the largest technology infrastructure projects in its history.

Over the next few years, roughly €30 billion could be invested to build seven AI gigafactories across the continent.

The obvious question is whether Europe can build its own AI champions.

That is not the question I have been asking.

Every major infrastructure program creates winners long before anyone knows whether the project itself will ultimately be considered a success. That is the lens through which I think investors should look at Europe’s AI gigafactories.

The plan is to build seven facilities, each containing at least 100,000 advanced AI processors alongside the data centers, networking, cloud infrastructure and software required to train frontier AI models.

The European Union and participating governments are committing approximately €10 billion in public funding, with the objective of attracting at least another €20 billion from private investors.

The objective is straightforward.

Europe wants to rebuild its position in the AI race.

“Access to the raw scale of computing power within AI gigafactories is a strategic necessity for Europe as AI development accelerates,” EU technology chief Henna Virkkunen said when the tender was announced.

She is right about the problem.

Europe has world-class engineering talent, leading industrial companies and strong positions in areas such as semiconductor equipment, automation, power infrastructure and telecommunications.

What it lacks is enough frontier-scale computing capacity.

The United States dominates hyperscale cloud infrastructure and hosts most of the leading AI model developers. China has spent years building massive domestic compute capacity supported by abundant power infrastructure. Europe’s existing network of AI factories is valuable, but it remains too small for companies training the largest foundation models.

The new gigafactories are designed to change that. Successful bidders are expected to be selected in early 2027, with construction beginning shortly afterwards.

But the project contains an interesting contradiction.

Europe is investing in sovereign AI infrastructure using technology it largely does not produce itself.

AMD, Nvidia and Qualcomm have already signed letters of intent to supply processors to participating projects. Many of the largest cloud providers involved are American, while electricity prices across much of Europe remain significantly higher than in the United States or China.

Europe may own the facilities, regulate the data and subsidize the computing capacity.

A meaningful share of the economic value could still flow elsewhere.

That does not necessarily make the investment a failure.

Building seven AI gigafactories will require hundreds of thousands of processors, substations, transformers, electrical equipment, optical networking, cooling systems and years of construction. Even if Europe never produces its own OpenAI, someone still has to build the infrastructure.

And that is where I think investors should focus.

Most commentary is centered on whether Europe can create globally competitive AI companies.

I am more interested in identifying the businesses that will generate revenue long before we know the answer.

History shows that infrastructure booms often create the biggest winners upstream, not necessarily among the companies expected to dominate the end market.

There are still obvious risks.

Europe has repeatedly struggled to transform ambitious industrial policy into globally competitive technology businesses. Projects can be delayed by permitting, fragmented procurement, national politics and changing budgets. Around €4 billion of the proposed EU contribution still depends on the bloc’s next long-term budget, which has yet to be approved.

At the same time, interest has been stronger than expected.

Eighteen countries have joined the initiative, and the original plan was expanded from five gigafactories to seven following demand from member states.

The deadline for proposals is November 12.

Between now and then, governments, utilities, data-center operators, semiconductor companies and infrastructure suppliers will begin forming the consortia that could build Europe’s next generation of AI infrastructure.

Whether Europe ultimately succeeds is still uncertain.

The companies supplying that infrastructure, however, will start generating revenue long before we know the answer.

Most investors spend their time debating which AI company will dominate tomorrow.

That is not how I approach investing.

At Tech Economics, I follow where capital is being deployed, identify which businesses capture the economics of each technology cycle, and look for opportunities before they become consensus.

Europe’s AI gigafactory program is a perfect example.

Rather than asking whether Europe will build its own OpenAI, I traced where the €30 billion is most likely to flow.

In this edition, I reveal:

• The countries and consortia most likely to secure the seven projects.

• How much computing, electricity, networking and cooling infrastructure each site could require.

• The European and American companies positioned to supply the buildout.

• Why power availability—not AI expertise—may determine the winning locations.

• The businesses most exposed to political delays, budget risk and uneconomic capacity.

• My exact public-market investment strategy: positions, entry zones, portfolio sizing and the catalysts I will be watching.

By the time these companies become the obvious AI infrastructure winners, a large part of the opportunity may already be reflected in their valuations.

If you want to consistently identify the technology businesses capturing the most value—not simply the ones making headlines—Tech Economics is exactly what this newsletter is built for.

It costs just $100 per year (the price will increase soon). Subscribe before the price goes up — if you’re already subscribed, your price will be locked in and won’t change.