EU Funding AI: Private Sector to Boost 'AI Gigafactories'

EU Seeks Private Investment to Boost AI Compute Capacity

The European Union is actively engaging with the private sector in an effort to significantly increase its computational resources for the development of large AI models.

Von der Leyen Highlights Need for Scalable Infrastructure

Speaking at the AI Action Summit in Paris on Monday, EU President Ursula von der Leyen acknowledged the potential of European AI startups. However, she emphasized that these developers require access to sufficiently powerful infrastructure to effectively scale their innovations and realize the full benefits of the technology.

Consequently, the EU is re-evaluating its current strategy for supporting AI infrastructure.

From AI Factories to AI Gigafactories

The existing initiative to establish talent and computing hubs utilizing the bloc’s supercomputing infrastructure – currently referred to as “AI factories” – is deemed insufficient. Von der Leyen proposed the need for “AI Gigafactories” specifically designed for training “very large models.”

The construction of this “next level” of AI infrastructure will necessitate substantial private sector investment.

Expanding Cooperation for Frontier Innovation

“Our startups require resources to expand, and widespread AI adoption within our economy and society remains a goal,” stated von der Leyen. “Therefore, we are progressing to a new phase, aiming to broaden our model of open collaboration to accommodate cutting-edge AI innovation.”

She continued, “Given the immense computational demands of AI, our next step involves launching AI Gigafactories – large-scale data and computing infrastructures for training very large models.”

EU Responds to Global Competition

Similar initiatives have been announced in the United States by prominent AI companies. However, von der Leyen asserted that European Gigafactories will ensure computational power is not concentrated in the hands of a few, but rather accessible as a service to all.

France unveiled a private sector AI investment package of approximately $112 billion within the EU just before the summit. However, the U.S. Stargate project previously committed up to $500 billion over four years to develop data center infrastructure, aiming to solidify U.S. leadership in AI. This demonstrates the pressure on the EU to respond to the escalating global competition in AI compute.

Private Sector Involvement is Crucial

“For AI, we need full engagement from the private sector in our gigafactories, and we require increased capital to achieve this,” von der Leyen added, noting that the topic would be a focus of a closed-door meeting at the summit.

Leveraging a Collaborative Approach

Von der Leyen highlighted Europe’s collaborative approach to intellectual property development as a potential advantage in attracting investment for the next phase of AI infrastructure. This approach prioritizes pooling and sharing knowledge across member states for the public benefit.

Benefits of Open Access to Infrastructure

“I want to emphasize our commitment to providing researchers and startups with a unique opportunity to access state-of-the-art computing infrastructure,” she explained. “This will facilitate cooperation and data federation among industries.”

She suggested this would allow hospitals to securely train models using their image and genomic data, leading to significant advancements in basic science and climate modeling.

Attracting Top Talent

The EU’s planned AI gigafactories, she concluded, “will be open to the best talents.”

€200 Billion Investment to Propel AI Implementation

Ursula von der Leyen, during subsequent statements at the summit on Tuesday, revealed the EU’s commitment of €50 billion towards AI investment. This contribution is intended to supplement a €150 billion pledge from a coalition of private sector companies, collectively known as EU AI Champions, focused on fostering regional AI adoption.

“The European AI Champions Initiative, with its €150 billion commitment from providers, investors, and industry, is highly encouraging,” she stated to attendees. “Furthermore, through our InvestAI initiative, we will add an additional €50 billion, aiming to mobilize a total of €200 billion for AI investment within Europe.”

Focus on Critical Applications

Von der Leyen emphasized a strategic focus on industrial and mission-critical applications. This initiative is positioned as the world’s largest public-private partnership dedicated to the development of trustworthy AI.

The EU’s InvestAI initiative, as detailed in a press release, will allocate €20 billion to establish AI Gigafactories. The objective is to facilitate open and collaborative development of advanced AI models, solidifying Europe’s position as a leading AI continent.

These gigafactories are designed to provide substantial compute capacity, with the EU Commission anticipating breakthroughs in specialized fields like medicine and scientific research.

Gigafactory Specifications and Funding Structure

The planned gigafactories will be equipped with approximately 100,000 cutting-edge AI chips. This represents a fourfold increase in capacity compared to currently established AI facilities.

InvestAI will utilize a “layered fund” structure, offering diverse risk and return profiles. The EU budget will play a crucial role in mitigating risks for private sector investment partners.

Initial funding for InvestAI will be sourced from existing EU programs with a digital focus, including the Digital Europe Programme and Horizon Europe, alongside InvestEU. Member states will also have the opportunity to contribute financially.

This article has been updated to reflect additional announcements made by the EU on Tuesday.

For comprehensive coverage, see our full report on the Artificial Intelligence Action Summit in Paris.

Stay informed with TechCrunch’s AI newsletter! Subscribe here to receive it weekly on Wednesdays.