Table of Contents
The European Union (EU) has launched a plan to support homegrown AI startups by providing them with access and support to use its supercomputers for training AI models. This initiative is part of the EU’s strategy to foster a local AI ecosystem that can compete with the U.S. AI giants.
In this article, we will explore how the EU is implementing this plan and what are its implications for the future of AI in Europe.
The EU’s Plan to Provide Supercomputing Access to AI Startups
The EU has invested heavily in supercomputers over the past few years, creating a cluster of eight machines located across the region, with plans to interconnect them via terabit networks. This will enable users from all over Europe to access the computing power in the cloud. The EU is also working towards deploying its first exascale supercomputers in the coming years, as well as investing in quantum computing.
The EU recognizes the potential of supercomputers for advancing AI research and development, especially for training large-scale AI models that require massive amounts of data and processing power. However, the EU also acknowledges the need for dedicated support to train AI startups on effectively utilizing the high-performance computing resources provided by the EU.
To address this need, the EU has initiated a pilot phase of its plan to provide supercomputing access to AI startups. One of the participants in this pilot phase is Mistral AI, a startup based in France that specializes in natural language processing and computer vision. Mistral AI has collaborated with the EU’s current-gen supercomputers to train its AI models, but it has also emphasized that it primarily relies on its cluster for its training needs.
How the EU is Supporting AI Startups with Training and Skills Development
The EU’s plan to provide supercomputing access to AI startups is not limited to just providing access, but also to providing the necessary skills, knowledge, and experience to AI startups. The EU understands that AI startups are more accustomed to using dedicated compute hardware from U.S. hyperscalers for their training needs, rather than tapping into the processing power offered by the supercomputers.
To bridge this gap, the EU has set up “centers of excellence” to support the development of dedicated AI algorithms that can run on the EU’s supercomputers. These centers will help AI startups understand how to best utilize the architecture and computing power of the supercomputing centers for training their algorithms. This includes training on parallelizing algorithms for AI models and maximizing the potential of the supercomputing resources.
The EU is also planning to launch an “AI support center” specifically geared towards SMEs and startups. This center will have a dedicated track to assist these startups in obtaining the maximum benefit from the EU’s supercomputing resources. The center will provide facilities and resources to help these startups access and effectively use the EU’s supercomputers.
The EU’s Vision for a Competitive AI Ecosystem Powered by Supercomputers
The EU’s focus on supporting AI startups to tap into its investments in high-performance computing aligns with its strategy to foster a local AI ecosystem that can compete with the U.S. AI giants. The EU hopes that by providing support to AI startups and leveraging its supercomputers, it can nurture a competitive AI ecosystem in Europe.
The EU has also introduced a competition called the “Large AI grand challenge,” targeting European AI startups experienced in large-scale AI models. The competition aims to select up to four promising startups that will receive a total of 4 million hours of supercomputing access to support the development of foundational AI models. The winners will also be eligible for a €1 million prize and are expected to share their research findings or release their models under open-source licenses.
The EU’s supercomputers are also being used for various applications that can benefit society and the environment. For example, one of the projects is “Destination Earth,” which focuses on simulating Earth’s ecosystems to better model climate change and weather systems. Another project aims to create a digital twin of the human body, offering potential advancements in medical science and personalized medicine.
The EU’s “supercomputer for AI” program is still in its early stages, and its impact on model training is yet to be fully assessed. However, the EU is taking proactive steps to support AI startups and leverage its high-performance computing investments. The EU’s plan to provide supercomputing access to AI startups is a promising initiative that could boost AI innovation and competitiveness in Europe.