{"slug": "cerebras-hands-prefill-to-amd-gpus-to-get-5x-more-tokens-per-megawatt", "title": "Cerebras hands prefill to AMD GPUs to get 5x more tokens per megawatt", "summary": "Cerebras Systems has put AMD Instinct GPUs on prompt prefill and its CS-3 wafer-scale systems on decode in a production system it says delivers 5x more tokens per megawatt, first on Cerebras Inference Cloud and then in partner data centers. The split-workload design, described by Cerebras CEO Andrew Feldman, pairs AMD GPUs for prompt processing with Cerebras CS-3 hardware for token generation.", "body_md": "# Cerebras hands prefill to AMD GPUs to get 5x more tokens per megawatt\n\n**The production system puts AMD Instinct GPUs on prompt processing and Cerebras CS-3 systems on decode, first on Cerebras Inference Cloud and then in partner data centers.**\n\n        By [Ryan Merket](https://runtimewire.com/author/ryan-merket)\n        · Published \n\nAndrew Feldman built Cerebras Systems around a chip so large it uses most of a silicon wafer. The company's newest production system divides the work between that chip and AMD GPUs.…", "url": "https://wpnews.pro/news/cerebras-hands-prefill-to-amd-gpus-to-get-5x-more-tokens-per-megawatt", "canonical_source": "https://runtimewire.com/article/cerebras-hands-prefill-to-amd-gpus-to-get-5x-more-tokens-per-megawatt", "published_at": "2026-11-17 20:20:18+00:00", "updated_at": "2026-11-17 21:15:51.499195+00:00", "lang": "en", "topics": ["ai-infrastructure", "ai-chips", "large-language-models"], "entities": ["Cerebras Systems", "AMD", "AMD Instinct", "Cerebras CS-3", "Cerebras Inference Cloud", "Andrew Feldman"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/cerebras-hands-prefill-to-amd-gpus-to-get-5x-more-tokens-per-megawatt", "markdown": "https://wpnews.pro/news/cerebras-hands-prefill-to-amd-gpus-to-get-5x-more-tokens-per-megawatt.md", "text": "https://wpnews.pro/news/cerebras-hands-prefill-to-amd-gpus-to-get-5x-more-tokens-per-megawatt.txt", "jsonld": "https://wpnews.pro/news/cerebras-hands-prefill-to-amd-gpus-to-get-5x-more-tokens-per-megawatt.jsonld"}}