Shadow launches Blade: cloud and GPU for sovereign AI in Europe

The optimization of infrastructure usage rates allows you to reduce calculation costs by a percentage of between twenty-five and forty percent compared to the main market suppliers, eliminating the fees for outgoing data transfer and allowing the planning of non-urgent loads in the lowest consumption time slots.

For organizations requiring high standards of data protection, the platform offers the option to run on European sovereign servers managed in collaboration with OVHcloud. On an operational level, the suite is made up of the Blade Models service for immediate access via API to approximately twenty algorithmic architectures intended for text, code, vision and audio transcription, which will be joined by the AI ​​Fabric functionality for automatic orchestration and monitoring of resources starting from Python environments or repositories such as Hugging Face. The offer also integrates the Staan interface, developed by Qwant and Ecosia, for retrieving structured information from the web and supporting generation systems augmented by recovery.

Shadow CEO Boris Lecoeur underlined the innovative scope of the operation: “IIn Europe there is no shortage of talent or AI models. The challenge for startups is to bring their applications into production. Computing capacity is still expensive, difficult to manage and largely controlled by non-European providers. We have been running large-scale GPUs for over ten years. With Blade, we make this infrastructure and our expertise available to companies developing AI solutions in Europe, so they can focus on building their products and bringing them to market faster. From search indices to computing infrastructures, Europe is building a credible technological alternative. With Blade you can now control a key component of this technology stack”.