Gcore AI

Gcore AI is the AI-infrastructure arm of Gcore, delivering NVIDIA H100 and A100 GPU compute, inference endpoints and Everywhere Inference edge serving for training and deploying large AI models. Headquartered in Luxembourg and built on Gcore's 180+ point-of-presence network, it brings model inference physically close to users for sub-100 ms latency worldwide.

The platform combines bare-metal and virtual GPU instances with managed inference services: one-click deployment of open-source models (Llama, Mistral, Stable Diffusion, Whisper), custom model uploads and Everywhere Inference that replicates models across Gcore's global edge footprint. Integrations with Gcore CDN, storage and security give teams a vertically integrated AI-delivery stack.

Key benefits:

• NVIDIA H100 and A100 GPUs available hourly, monthly and on multi-year reservations
• One-click inference endpoints for Llama, Mistral, Stable Diffusion, Whisper and custom models
• Everywhere Inference replicating models across 180+ edge PoPs for low-latency serving
• Bare-metal and virtual GPU instances with high-bandwidth networking
• Native integration with Gcore CDN, object storage, DDoS protection and WAF
• Usage-based billing with transparent per-second pricing and reservation discounts
• ISO 27001, ISO 27018, PCI DSS and GDPR coverage across regions

Gcore AI is headquartered in Luxembourg, with data centres across Europe, the Americas, Asia, the Middle East and Africa and EU-residency guarantees under GDPR. Customers can pin workloads to EU-only regions and receive a full DPA with SCCs for any sub-processor.

Ideal for AI product teams, foundation-model builders and enterprises that need GPU training capacity plus globally distributed inference from a Luxembourg-headquartered European cloud provider.