Cerebras Systems expanded both its AI systems manufacturing capacity and international infrastructure footprint, announcing plans to increase production capacity for its CS-3 systems by approximately 7x through 2026 while building up to 200 MW of AI compute capacity across Europe by the end of 2027. The company expanded its manufacturing partnership with Flex to add dedicated CS-3 production lines at Flex facilities in Milpitas, California, and plans to bring its first European data center capacity online by the end of 2026.
The Milpitas expansion adds assembly and integration lines, floor space, automated test infrastructure, burn-in and validation areas, and warehouse and logistics capacity. Flex and Cerebras developed specialized assembly flows, tooling, automated test stations, and manufacturing processes for the CS-3, which uses Cerebras’ wafer-scale processor architecture. Production operations include precision mechanical assembly, high-power electrical integration, liquid cooling installation, optical networking validation, and full-rack system qualification.
In Europe, Cerebras plans to deploy infrastructure in France and the Nordic region, including data centers planned for Norway and Finland. The company expects to scale its European capacity to 200 MW by the end of 2027, with part of that capacity expected to support OpenAI workloads under the companies’ existing partnership. Cerebras said the regional deployment strategy will place inference infrastructure closer to European enterprises, research institutions, governments, and other customers seeking locally deployed AI compute capacity.
• CS-3 production capacity is expected to increase approximately 7x through 2026.
• Flex is expanding dedicated Cerebras manufacturing operations in Milpitas, California.
• New manufacturing infrastructure includes parallel integration lines, automated test stations, expanded burn-in and validation areas, and additional logistics capacity.
• CS-3 manufacturing includes liquid cooling installation, high-power electrical integration, optical networking validation, and full-rack system qualification.
• Cerebras plans to activate its first European data center capacity by the end of 2026.
• European infrastructure capacity is targeted to reach 200 MW by the end of 2027.
• Planned European locations include France, Norway, and Finland.
• Part of the European capacity is expected to support OpenAI workloads under the companies’ existing partnership.
“The CS-3 is unlike any computer system ever built, and scaling its production requires an extraordinary manufacturing partner. Flex brings the technical depth, operational rigor, and manufacturing expertise needed to support that scale,” said Dhiraj Mallick, COO of Cerebras.
🌐 Analysis: The two announcements connect Cerebras’ system-level manufacturing strategy with its expansion as an AI infrastructure operator. Increasing CS-3 production capacity addresses the hardware supply side of the company’s growth plans, while the 200 MW European deployment target expands the infrastructure available to operate wafer-scale systems closer to regional AI workloads.
The strategy also illustrates the growing importance of manufacturing, power availability, cooling, networking, and geographic capacity deployment as competitive factors in AI infrastructure. Cerebras is scaling a vertically integrated architecture built around wafer-scale processors and complete systems while GPU-based providers and alternative accelerator companies expand their own cloud capacity, regional deployments, and system-level platforms.
🌐 We’re tracking the latest developments in AI infrastructure, accelerators, data centers, and large-scale compute deployments. Follow our ongoing coverage at: https://convergedigest.com/category/ai-infrastructure/
Cerebras Systems Wafer-scale AI computing systems and cloud infrastructure for large-model training and high-speed inference. | |
| Overview | Cerebras Systems develops AI processors, complete computing systems, software and cloud services based on a wafer-scale architecture. Its platform integrates compute cores, memory and a high-bandwidth communication fabric across a processor built from an entire silicon wafer. Cerebras systems support generative AI, large-language-model training, inference and high-performance computing. |
| Why It Matters | Cerebras represents an alternative to conventional multi-GPU AI clusters. Its wafer-scale design is intended to reduce memory, interconnect and distributed-computing bottlenecks for large AI and scientific workloads. |
| Founded | 2015 |
| Headquarters | Sunnyvale, California |
| CEO / Leadership | Andrew Feldman, co-founder and Chief Executive Officer |
| Core Technologies | Wafer-Scale IntegrationAI AcceleratorsOn-Chip MemoryDataflow ComputingAI InferenceModel Training |
| Key Products / Platforms | Wafer-Scale Engine 3 Third-generation wafer-scale AI processor built on a 5 nm process. CS-3 AI system based on the WSE-3 for training, inference and HPC workloads. Cerebras Inference Cloud-based infrastructure for running generative-AI models. Cerebras AI Supercomputers Multi-system deployments combining CS-series systems into larger computing environments. |
| Major Milestone | Cerebras completed its initial public offering in May 2026 and began trading on Nasdaq under the ticker CBRS. |
| Target Markets | AI InfrastructureCloud AIGenerative AIHPCEnterprise AI |
| Editorial Coverage | Cerebras Knowledge Hub on Converge Digest |
| Industry Context | Cerebras competes in an AI accelerator market dominated by GPU-based platforms. Its architecture differentiates through wafer-scale integration, distributed on-chip memory and a high-bandwidth internal communication fabric. |
| Related Knowledge Hubs | AI Infrastructure Semiconductors Data Centers |
| Profile Updated: July 2026 | |