TL;DR:
- EU accelerates support for AI startups by granting access to supercomputing power.
- France’s Mistral AI joins early pilot phase.
- Dedicated support for AI startups is crucial for maximizing benefits.
- “Centers of excellence” to develop optimized AI algorithms on EU supercomputers.
- EU focuses on providing skills, knowledge, and experience alongside access.
- Strategic priority: The EU aims to foster secure, trustworthy, and ethical AI algorithms.
- Plans for an “AI support center” with a dedicated SME and startup track.
- EU’s significant investment in supercomputers and plans for exascale machines.
- Integration of quantum simulators with supercomputers to create a hybrid resource.
- EU’s support extends to various applications beyond AI.
- France’s Mistral AI was an early beneficiary of the program.
- EU aims to compete globally with a strong local AI ecosystem.
Main AI News:
The European Union (EU) is intensifying its efforts to bolster homegrown AI startups by granting them access to the computational muscle of the bloc’s supercomputers. This initiative, first unveiled in September and launched recently, has witnessed participation from France’s Mistral AI during its early pilot phase, according to a recent EU update. However, a key takeaway from this endeavor is the imperative need to provide dedicated support to AI startups, equipping them with the skills, knowledge, and experience necessary to maximize the potential of the EU’s high-performance computing infrastructure.
In a press briefing, an EU official emphasized the importance of not only offering access to supercomputers but also facilitating their utilization. This includes the development of training algorithms that harness the full capabilities of the architecture and computing power available in each supercomputing center. The plan entails establishing “centers of excellence” dedicated to fostering the creation of AI algorithms optimized for deployment on EU supercomputers.
AI startups traditionally rely on dedicated computing hardware provided by American hyperscalers for their model training. Therefore, the EU’s program to provide high-performance computing access for AI training is complemented with a comprehensive support system. This strategic move aims to bridge the gap between AI startups’ existing expertise and the potential offered by supercomputers, preparing them for the future landscape.
The EU is committed to making AI a strategic priority, aligning with the AI Act. To achieve this, the EU intends to create an extensive innovation window for small and medium-sized enterprises (SMEs) and startups, enabling them to harness the capabilities of the EU’s supercomputing infrastructure for the development of secure, trustworthy, and ethical AI algorithms.
In addition to setting up an “AI support center,” which will have a dedicated track for SMEs and startups, the EU acknowledges that the AI community has limited experience with supercomputers. The AI community’s expertise primarily lies in GPU utilization, while supercomputers require additional skills and assistance for optimal utilization.
Over the past five years, the EU has significantly increased its investment in supercomputers, culminating in a network of eight interconnected machines across the region. These interconnected supercomputers are slated to form a federated resource accessible in the cloud, serving users throughout Europe. The EU is also preparing to launch its first exascale supercomputers, with Germany and France at the forefront. Moreover, the EU plans to integrate quantum simulators with supercomputers to create a hybrid resource that combines both types of hardware.
This investment in high-performance computing is not limited to AI startups alone but extends to various applications, including projects aimed at simulating Earth’s ecosystems for climate change modeling and creating a digital twin of the human body to advance medical science.
The EU’s commitment to supporting commercial AI is evident through programs and resources dedicated to nurturing ‘Made in Europe’ AI on a global scale. France’s Mistral AI, a startup vying with US foundation model giants, has been an early beneficiary of the EU’s supercomputer access program.
As the EU forges ahead with its ‘supercompute for AI’ program, its ultimate goal is to transform its supercomputers into a powerful resource for the growth of the local AI ecosystem. By providing comprehensive support to AI startups and aligning supercomputer hardware with AI model training, the EU aims to gain a competitive edge against American hyperscaler-proximate AI giants.
With plans to acquire more dedicated AI supercomputing machines focused on accelerators, the EU is poised to further bolster its position in the global AI landscape. Whether this strategy aligns with Member States’ aspirations to foster national AI champions remains to be seen, but Mistral AI’s involvement in the EU’s supercomputing access program signals potential alignment in their objectives.
Conclusion:
fostering a competitive local AI ecosystem. By bridging the gap between AI startups’ existing expertise and the potential offered by supercomputers, the EU aims to position itself as a formidable player in the global AI market, challenging the dominance of American hyperscaler-proximate AI giants. This initiative signals an exciting development for the AI market, as it indicates a concerted effort to nurture ‘Made in Europe’ AI innovation.
