Cerebras Unveils CS-4 AI System, Challenges Nvidia's Dominance
Cerebras Systems has launched its latest AI system, CS-4, which it claims can run large models dramatically faster than traditional GPU infrastructure. This move positions Cerebras as a bold challenger to Nvidia's dominance in the AI market.
The CS-4 combines three WSE-3 Turbo processors and uses Nexus rack-scale architecture, delivering up to twice the performance of its predecessor. According to Cerebras, it can produce more than 4,400 tokens per second per user on the GPT-OSS-120B model, which is as much as 30 times faster than GPU-based alternatives.
The company also reports a significant increase in cloud and services revenue, with a 281% year-over-year jump to $126 million in the second quarter. Cerebras ended June with $25.4 billion in remaining performance obligations and more than 600 megawatts of data-center capacity live or contracted for delivery through 2027.
Cerebras CEO Andrew Feldman stated, 'Historically, fast inference meant using smaller and less capable models. Cerebras CS-4 delivers industry-leading speeds on the largest frontier models, fundamentally changing the paradigm.'