Cerebras Systems

Cerebras Systems is an AI company that designs processors for AI training and inference. The company builds AI systems to power, cool, and supply data to those processors. Cerebras Systems also develops software that links these systems together into industry-leading supercomputers that are simple to use, even for the most complex AI workloads, using familiar machine learning frameworks such as PyTorch.

Customers use Cerebras Systems’ supercomputers to train industry-leading models. The company also uses these supercomputers to run inference at speeds that are not achievable with alternative commercial technologies. Cerebras Systems delivers these AI capabilities to customers both on premise and through the cloud.

Cerebras Systems solved a problem that had remained open throughout the 75-year history of the computer industry: building chips the size of full silicon wafers. The third-generation Cerebras Wafer-Scale Engine, or WSE-3, is the largest chip ever sold. It is 57 times larger than the leading commercially available GPU, with 52 times more compute cores, 880 times more on-chip memory, totaling 44 gigabytes, and 7,000 times more memory bandwidth, reaching 21 petabytes per second.

The sheer size of the wafer-scale chip allows Cerebras Systems to keep more work on silicon and minimize the time-consuming, power-hungry movement of data. This enables the company’s customers to solve problems faster and with lower power consumption.

Cerebras Systems’ AI compute platform combines processors, systems, software, and AI expert services to deliver massive acceleration for even the largest and most capable AI models. The platform substantially reduces training times and inference latencies while also lowering programming complexity.