# Cerebras unveils CS-4 with double the performance on the same chip > Source: > Published: 2026-08-24 10:16:06+00:00 # Cerebras unveils CS-4 with double the performance on the same chip **Cerebras has introduced its CS-4 AI accelerator, which CEO Andrew Feldman calls the fastest system in the industry.** The CS-4 is a rack-scale product, meaning a full server cabinet that goes into data centers as a single unit, complete with compute units, power, and cooling. The CS-4 still runs on the 5nm [WSE-3](https://the-decoder.com/cerebras-wse-3-enables-ai-models-to-be-ten-times-larger-than-gpt-4-and-gemini/) chip, but it doubles the performance of the earlier [CS-3](https://the-decoder.com/nvidia-rival-cerebras-launches-six-data-centers-for-rapid-ai-inference/) by boosting clock speed through more power and better cooling. A single rack now holds three wafers instead of two and delivers up to 4,400 tokens per second per user. Cerebras says that's up to 30 times faster than setups running on Nvidia GPUs. Memory capacity stays the same at 44 GB per wafer. Cerebras is using a new, modular "Backpack" design for faster assembly, along with disaggregated inference through partners like AMD and AWS Trainium. Analysts at [SemiAnalysis](https://newsletter.semianalysis.com/p/cerebras-faster-tokens-please) see the networking gains as fairly small. More details are coming at the Hot Chips conference. Cerebras hardware is used by [OpenAI for Codex Spark](https://the-decoder.com/openai-has-yet-another-new-coding-model-and-this-time-its-really-fast/), among others. ``` AI News Without the Hype – Curated by Humans Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section. Subscribe now ``` [Cerebras](https://investors.cerebras.ai/news-releases/news-release-details/cerebras-unveils-cs-4-30-times-faster-gpu-based-solutions)