General Compute signs multi-year deal for Cerebras ultra-fast inference

Neocloud General Compute signed a multi-year agreement to deploy Cerebras inference at scale, starting with agentic coding. Wafer-scale inference will be available through General Compute from Q1 2027. The release gives no capacity figures, and the financing comes from lenders funding the hardware, not from Cerebras's balance sheet.

• Multi-year agreement; capacity not disclosed
• Available through General Compute starting Q1 2027
• First use case is agentic coding
• Lenders fund purchases, so Cerebras does not carry the financing directly