Key Points
- AxeleraAI has received up to €61.6 million in funding from EuroHPC as part of the EU’s initiative to strengthen domestic semiconductor production.
- The grant aims to develop a Titania chip, which differs from training-focused chips for inference computing in data centers.
- The chip will be built on the open-source RISC-V standard, providing a cost-effective alternative to Intel and Arm.
- AxeleraAI’s existing chip, Metis, supports edge AI applications, while Titania aims for high-performance data center use.
AxeleraAI, one of Europe’s few companies producing advanced chips for artificial intelligence, has been awarded up to €61.6 million ($66 million) by EuroHPC to develop its new Titania chip for data centers. This funding is part of a broader EU initiative to bridge the AI competitiveness gap with the United States and China by boosting domestic semiconductor capabilities and building publicly funded “AI factories” that benefit European scientists, companies, and startups.
Based in Eindhoven, the Netherlands, AxeleraAI has been making strides in the AI chip market with its existing product, Metis, which is used for edge AI applications such as monitoring CCTV footage in factories. Now, the company is venturing into the data center space by developing Titania—a chip optimized for “inference” computing. Unlike the intensive computing required for training large AI models, inference involves the “thinking” phase, where the model processes data to make decisions.
This process can benefit from chips designed for efficiency rather than sheer raw power. “We are not here to challenge Nvidia in the data center space for training,” explained Axelera CEO Fabrizio Del Maffeo. “But when the network is ready and you want to run it, we are developing a solution that can deliver extremely high performance.”
The Titania chip will be built on the open-source RISC-V architecture, a standard gaining traction as a cost-effective alternative to the dominant Intel and Arm platforms, particularly in the auto industry and increasingly in China. The emergence of Chinese AI models like DeepSeek, which deliver cutting-edge performance at lower costs, may further drive demand for efficient inference computing solutions.
Since its founding in 2021, AxeleraAI has raised approximately $200 million from investors, including Samsung. With this new grant, the company is poised to expand its business and contribute to the EU’s strategy of enhancing technological self-reliance in semiconductor manufacturing while offering competitive solutions in the rapidly growing AI market.
