Blackfuel, a Paris-based AI-native neocloud, has emerged from stealth with more than US$250 million in contracted revenue under multi-year customer agreements. The company is developing architecture-specific accelerator clusters connected through a common software and commercial layer, with the goal of turning dedicated compute into measurable, reservable AI output. Its initial European-operated inference capacity is being developed with AMD and NTT DATA, and the first deployment is scheduled to enter service in 2026.
European Deployment and Infrastructure
Blackfuel is building an integrated compute platform that combines dedicated accelerator infrastructure with the software and services needed to make production AI predictable. Its first inference-only deployment in Barcelona, Spain, will be hosted by Digital Realty and built on liquid-cooled Dell PowerRack systems featuring Dell PowerEdge XE9785L servers with AMD Instinct MI355X accelerators. The Dell AI Data Platform, using Dell PowerScale and Dell PowerStore, will provide the data foundation, with deployment and integration support from Dell and NTT DATA.
Dedicated and Managed Capacity Offerings
The company is also advancing additional cluster deployments in Spain, Finland, and France. Blackfuel's offering spans dedicated GPU clusters, managed inference, and reserved serving capacity, giving customers the option to retain control of their compute environment or engage Blackfuel to manage the serving layer against agreed workloads. These deployments are designed to support large customer commitments while creating a broader inference offering for model developers, enterprises, and AI-native businesses.
Fractional Capacity and Logical GPUs
Blackfuel is introducing fractional GPU capacity through Logical GPUs, which are service units that let customers reserve a share of a model's serving capacity for a fixed monthly fee. Commitments specify the selected model, input and output token quotas, and applicable service levels, all sized around the customer's workload. Customers can reserve less than a full model replica, while Blackfuel can evolve the physical hardware and serving configuration without changing the agreed capacity.
Token Platform and Supported Models
Alongside dedicated customer deployments, Blackfuel is developing a separate inference pool for its Token Platform, with day-one support at launch for selected models from Meta, Mistral, H Company, DeepSeek, Qwen, GLM, and Moonshot AI's Kimi. Coding assistants and agentic applications are priority workloads. Through a single enterprise interface, customers can select models and reserve inference capacity against defined throughput, latency, and data-location requirements without managing the underlying infrastructure.
Software Optimization
Blackfuel's inference stack combines open serving technologies, including vLLM and SGLang, with proprietary routing, allocation, metering, and commercial-control software. For AMD-based infrastructure, the stack builds on AMD ROCm Software, and the company is developing an agent-driven process to build, benchmark, and continuously optimize inference runtimes for each model and accelerator architecture. The approach is designed to lower serving costs and increase useful tokens per megawatt while remaining within defined quality and latency requirements.
Sovereign AI and Financing Model
The company's sovereign AI model pairs European-operated infrastructure with a service structure focused on customer control over model selection, deployment location, and capacity commitments. Blackfuel aims to make contracted token output as financeable as contracted power by connecting customer demand, data-center capacity, accelerator systems, operating partners, and project financing into a coordinated deployment model. Starting with European deployments, the company intends to replicate this model worldwide.
Partner Perspectives
AMD, Dell Technologies, Digital Realty, and NTT DATA all signaled support for Blackfuel's approach. AMD's senior vice president for EMEA, Stephanie Dismore, said the company looks forward to working with Blackfuel to bring AMD technology to its initial deployment and expand access to high-performance AI infrastructure. Dell Technologies said organizations are moving from AI ambition to AI impact, while Digital Realty highlighted the importance of power, cooling, connectivity, and room to grow.
Blackfuel, co-founded by Charles Kantor, Xavier Fischer, and Dali Kilani, is positioning itself as an AI-native neocloud built for inference at scale. The company combines dedicated accelerator infrastructure, proprietary software, and long-term commercial commitments to serve dedicated capacity and managed inference needs. With more than US$250 million in contracted revenue already secured and European deployments forming the foundation, Blackfuel is preparing to expand its AI Token Grid internationally.