OVHcloud is partnering with SambaNova to deliver ultra-low latency AI inferencing, targeting applications where speed and reliability are critical.
At the annual OVHcloud Summit in Paris, the European cloud provider announces support for SambaNova’s SambaStack infrastructure for AI endpoints. SambaNova uses Reconfigurable Dataflow Units (RDUs), specialized chips developed for AI inferencing. The technology should support organizations in tasks such as financial trading, cybersecurity, industrial automation, and logistics optimization.
The collaboration aims to improve performance in terms of ‘time to first token’ and ‘time per output token’, two important parameters when performing large-scale AI workloads. The new service is intended for both real-time applications with guaranteed performance and batch APIs that can process large numbers of requests when immediate response is not required. The availability of a 99.8% uptime SLA indicates that OVHcloud is aiming for production-ready environments.
Complement to OVHcloud’s GPU Offering
The SambaNova infrastructure complements OVHcloud’s existing GPU-based AI endpoints. The technology is said to use energy and data center capacity more efficiently and deliver more computing power per kilowatt hour. This makes it suitable for AI agents, live translation, agent-to-agent communication, and large-scale batch processing such as web crawling or dataset renewal.
read also
OVHcloud Implements AI for Sustainable Cooling Technology in Data Centers
For OVHcloud, the collaboration with SambaNova is a strategic move to broaden its AI offering and respond to the growing demand for high-performance inferencing. The service will be available in regions within France before the end of the year. Later expansion to other European countries is planned. The service is offered via a pay-as-you-go model with a mandatory commitment.
Berlin and Quantum
Leading up to the OVHcloud Summit in Paris, the European cloud player announced the opening of a cloud region in Germany. It is the company’s first such region in Germany, and the third in Europe after previous launches in Paris and Milan.
With the new region in Berlin, OVHcloud aims to respond to the growing demand for digital sovereignty, security and resilience within the European market.
Other news on the sidelines of OVHcloud Summit is the launch of Quantum Platform. This gives you cloud access to European quantum systems. The first available processor is the Pasqal Orion Beta QPU from the French Pasqal. OVH wants to quickly offer the eight most advanced quantum computers of the moment via the service. The platform should help organizations test use cases without their own specialized infrastructure.