Paris, October 22, 2026. French cloud computing pioneer Shadow has officially launched Blade, a GPU platform designed to help developers, startups, and enterprises run AI models and move applications into production without managing complex infrastructure. The announcement marks a strategic expansion beyond Shadow's historic cloud gaming and cloud PC business. Blade combines competitive pricing, simplified access to inference, and a sovereign European option.
A Decade of GPU Expertise
Shadow first gained attention as Blade, a company that allowed gamers to access powerful cloud PCs instead of buying high-end hardware. Over time the service attracted more than 1.1 million users since 2017 and expanded into professional workstations. Following financial turbulence and a 2021 acquisition by Jezby Ventures, the brand now operates under Shadow and is using its decade of GPU management experience to enter the AI infrastructure market.
Leveraging Existing GPU Capacity
Shadow currently operates 12,000 GPUs across three data centers in Europe and three in the United States. Those resources primarily power Shadow PC users during evening and weekend gaming hours, leaving significant daytime capacity available for other computing tasks. Blade monetizes that idle capacity by directing GPUs toward professional AI workloads, an approach that improves utilization and enables rates between 25% and 40% below major market players.
Two Offerings for AI Development
The platform launches with Blade Models, available immediately, and AI Fabric, expected in November. Blade Models offers access by API to roughly 20 open-weight models covering text, code, vision, image generation, and audio transcription. AI Fabric goes further by letting developers deploy their own models, code, or applications directly from Python or Hugging Face without creating or administering GPU instances.
Pricing and Developer Experience
Blade uses an API compatible with the OpenAI standard, so developers can integrate models into existing tools and change providers more easily. The company does not charge for data egress or additional services, and it provides scheduling tools that allow batch operations to run during lower-cost periods. Developers can track consumption in real time and set budget caps, while billing is calculated per minute for GPU and CPU resources.
Web Grounding and Sovereign Infrastructure
Blade also connects to Staan, the web search API developed by Qwant and Ecosia, enabling models to pull recent, structured, and sourced information from the internet. For organizations concerned about data residency, Blade can operate on infrastructure located in Europe and hosted in OVHcloud data centers. This setup aims to reduce dependence on non-European cloud providers while maintaining straightforward access to competitive GPU capacity.
Strategic Ambition for European AI
Shadow chief executive Boris Lecoeur said Europe does not lack AI talent or models but struggles with production because compute remains costly, complex, and concentrated outside the region. He positioned Blade as a European alternative that lets teams focus on products rather than infrastructure management. The service is part of Synfonium, the group created by Octave and Miroslaw Klaba, and is intended to become a meaningful B2B growth engine alongside Shadow's cloud PC activity.
With Blade, Shadow is moving beyond its consumer cloud gaming roots to address a fast-growing AI inference market. The platform bundles compute, search grounding, and sovereign infrastructure in a simpler package aimed at startups and developers. Whether demand meets that ambition remains to be seen, but the launch gives European AI builders a credible option outside incumbent hyperscalers.