Scaleway and Fujitsu have joined forces to explore CPU-based AI inference as a sustainable option for European organizations. Combining Fujitsu’s advanced processor research with Scaleway’s cloud expertise, the collaboration will evaluate the MONAKA processor, designed for power efficiency, cost efficiency, and reliable inference performance. Early results from Fujitsu indicate gains of up to 1.9 times in power efficiency and up to 1.7 times in cost efficiency for selected inference workloads, with the ability to support models of up to 70 billion parameters. This aims to broaden efficient compute options available on Scaleway’s platform and support the development of high-performing, responsible AI systems.
Scaleway and Fujitsu Collaboration on AI Inference
Scaleway and Fujitsu are collaborating to explore CPU-based AI inference, aiming for a sustainable and sovereign option for European organizations. This partnership combines Fujitsu’s processor research with Scaleway’s European cloud expertise to evaluate efficient AI deployment at scale. The collaboration addresses Europe’s growing demand for infrastructure balancing performance, energy efficiency, and data sovereignty, particularly as AI models are increasingly used continuously in production environments.
Fujitsu’s MONAKA processor is central to this effort, designed for power and cost efficiency alongside reliable inference performance. Early results indicate gains of up to 1.9 times in power efficiency and 1.7 times in cost efficiency for select AI workloads. MONAKA also supports models of up to 70 billion parameters, providing a high-performance option for organizations needing efficient and scalable AI deployment.
The collaboration will involve testing MONAKA across real inference scenarios within Scaleway’s cloud platform. This aims to expand the range of efficient compute options available to European builders and support the development of responsible AI systems. Scaleway intends to offer the appropriate architecture for varying AI workloads while maintaining transparency regarding performance, cost, and environmental impact.
MONAKA: CPU-Based Efficiency and Performance
Scaleway and Fujitsu are collaborating to evaluate MONAKA, a CPU-based system designed for efficient AI inference. MONAKA aims to address the growing demand for infrastructures that balance performance, energy use, and data sovereignty within Europe. The system is intended to complement existing AI deployments and offers a viable option for organizations prioritizing stable performance and reduced environmental impact. This collaboration seeks to broaden efficient compute options available on Scaleway’s cloud platform.
Fujitsu designed MONAKA with a focus on power and cost efficiency, as well as reliable inference performance. Early results indicate gains of up to 1.9 times in power efficiency and 1.7 times in cost efficiency for certain AI workloads. Importantly, MONAKA is built to support AI models containing up to 70 billion parameters, providing a high-performance solution for organizations scaling AI deployments.
This partnership aligns with Scaleway’s goal to provide infrastructure tailored to diverse operational and strategic needs. By testing MONAKA in real-world scenarios, both companies aim to expand the range of efficient compute options available and support the development of high-performing, responsible AI systems. This ultimately contributes to a more sustainable digital future and Europe’s green transformation.
Early results shared by Fujitsu indicate gains of up to 1.9 times in power efficiency and up to 1.7 times in cost efficiency for selected inference workloads.
Fujitsu
Commitment to Sustainable and Scalable AI Infrastructure
Scaleway and Fujitsu are collaborating to offer more sustainable and scalable AI infrastructure, focusing on CPU-based AI inference. This approach aims to address the increasing demand for efficient, responsible, and secure AI deployment within Europe. Recognizing that AI workloads vary, both companies intend to evaluate how Fujitsu’s MONAKA processor can complement existing AI systems and offer appropriate architectures for different use cases while maintaining transparency regarding performance and environmental impact.
Fujitsu’s MONAKA processor is designed with power and cost efficiency as key features. Early results indicate gains of up to 1.9 times in power efficiency and 1.7 times in cost efficiency for select inference workloads. Importantly, MONAKA supports AI models of up to 70 billion parameters, offering a high-performance option for organizations needing to deploy AI at scale. This technology is intended to be integrated into Scaleway’s platform to broaden efficient compute options.
This partnership reflects a commitment to long-term sustainability and a desire to meet the evolving demands of the AI market. Scaleway aims to provide infrastructure matching customer requirements, while Fujitsu seeks to address the critical global challenge of AI energy consumption. Both companies believe this collaboration will not only optimize customer costs but also accelerate Europe’s green transformation and build a sustainable future.
