Cerebras

AMD and Cerebras announced a technical partnership on July 23, 2026 to deliver a disaggregated AI inference solution that pairs AMD Helios rack-scale infrastructure with the Cerebras Wafer-Scale Engine. The joint architecture is designed so AMD Helios handles high-throughput prompt processing and large context windows, Cerebras plans to deploy AMD Helios in its data centers, with the combined solution expected to first become available through Cerebras Cloud in the second half of 2026. The partnership was unveiled at AMD's Advancing AI 2026 event and reflects a broader industry shift toward heterogeneous inference infrastructure that

Cerebras is the go-to platform for fast and effortless AI training. Learn more at cerebras.ai.

Cerebras and Compute Nordic Finland announced a new AI data centre in Mikkeli, Finland, developed under a phased plan that will scale to 165 MW of contracted IT capacity, with construction on the initial 50 MW phase already under way. The agreement is formalized through a series of service orders each carrying seven-ye The facility is purpose-designed for efficiency, using closed-loop cooling systems that recirculate water rather than drawing continuously from municipal supply, and the campus is designed to support waste-heat recovery that converts thermal energy from AI compute into a resource for the surrounding community. Cerebras

Cerebras announced a strategic partnership with London-based frontier AI software company Callosum on August 20, 2026 to deliver ultra-low-latency, heterogeneous agentic inference. The collaboration integrates Cerebras' Wafer-Scale Engine silicon directly into Callosum's software platform, which decomposes and orchestr Per the announcement, customers will be able to deploy sophisticated agentic AI systems across diverse hardware through API-based access mediated by Callosum, combining Cerebras' wafer-scale silicon with Callosum's orchestration software for multi-agent systems. The partnership expands Cerebras' customer footprint in E

Cerebras Systems announced the CS-4, a next-generation rack-scale AI accelerator built on three new WSE-3 Turbo wafers delivering 750 PFLOPS of AI compute. According to the announcement, CS-4 is designed to be up to 2x faster than CS-3 and provide up to 30x more tokens-per-second-per-user than leading GPU-based solutio This hardware unveil is significant for developers and customers building latency-sensitive AI applications, as the CS-4 platform raises per-user inference throughput by an order of magnitude compared to GPU-based alternatives while reducing I/O latency from 5 microseconds (CS-3) to as low as 2 microseconds. The new Ne

Lovable and Cerebras Systems announced a partnership, reported on August 5, 2026, to integrate Cerebras' inference platform powered by its Wafer-Scale Engine into Lovable's AI software creation product. By keeping model weights on a single wafer, the Cerebras engine delivers higher memory bandwidth than traditional GPU The collaboration is aimed at latency-sensitive, agentic development workflows, and Lovable highlighted that its platform has already facilitated more than 50 million projects since its 2024 launch. Both companies frame speed as a core product differentiator and say the partnership will open up new interactive, AI-nati

On July 9, 2026, Cerebras announced a major expansion of its European infrastructure, with the first European data center coming online by the end of 2026 and total capacity targeted at 200 MW by the end of 2027. The initial sites are planned in France and the Nordic countries, designed to deliver low-latency AI comput The expansion is framed around meeting regional demand for local AI infrastructure and reducing latency for European customers, with dedicated capacity complementing Cerebras's public inference endpoints and Dedicated Endpoints product. Forward-looking statements in the announcement carry typical execution risk languag

AWS and Cerebras will deploy CS-3 systems in AWS data centers for the fastest AI inference on Amazon Bedrock, launching in coming months14.