{"version":1,"type":"story","url":"https://digestai.news/story/openai-s-jalapeno-chip-ai-driven-design-in-under-two-years","json":"https://digestai.news/story/openai-s-jalapeno-chip-ai-driven-design-in-under-two-years.json","markdown":"https://digestai.news/story/openai-s-jalapeno-chip-ai-driven-design-in-under-two-years.md","slug":"openai-s-jalapeno-chip-ai-driven-design-in-under-two-years","headline":"OpenAI's Jalapeño Chip: AI-Driven Design in Under Two Years","summary":"On August 25, OpenAI unveiled its debut AI accelerator chip, **Jalapeño**. Capable of delivering up to 13.4 petaflops and linking to a massive 232 gigabytes of memory at 15.4 terabytes per second, Jalapeño promises significant performance improvements over Nvidia’s GB300. Benchmarks show it can reduce end-to-end latency by up to 3.6x while consuming less power. While these figures are promising, the chip's rapid design timeline is equally noteworthy—moving from concept to silicon in under 20 months and RTL completion within nine months. This speed was accelerated significantly by OpenAI’s large language models (LLMs), which were used throughout the design process. The team at OpenAI designed a workflow that leveraged AI, particularly XLS—a high-level synthesis tool originally developed at Google—resulting in significant reductions in development time and cost. Jalapeño is now ready for deployment in pods containing 2,048 chips.","keyPoints":["Jalapeño accelerates to 13.4 petaflops with 232 GB of memory","Reduces end-to-end latency by up to 3.6x compared to Nvidia's GB300","Design timeline was under 20 months, including RTL from first concept"],"whyItMatters":"Jalapeño’s rapid design and deployment highlights the potential of AI in accelerating chip development, potentially reducing costs and speeding up innovation.","category":{"slug":"hardware","name":"Hardware & Compute","url":"https://digestai.news/category/hardware"},"entities":{"companies":["OpenAI","Broadcom","agentic"],"models":["o3","GPT-6 Astra"],"people":[]},"firstPublishedAt":"2026-09-14T14:06:31Z","updatedAt":"2026-09-14T14:06:31Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"IEEE Spectrum AI","title":"How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip","url":"https://spectrum.ieee.org/llms-for-chip-design","publishedAt":"2026-09-14T14:06:31Z","type":"press","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":{"title":"OpenAI Samsung Chip AI Saga","url":"https://digestai.news/thread/openai-s-samsung-chip-partnership-sparks-broadcom-stock-drop","storyCount":2},"cite":{"text":"Digest AI, \"OpenAI's Jalapeño Chip: AI-Driven Design in Under Two Years\", 14 September 2026, https://digestai.news/story/openai-s-jalapeno-chip-ai-driven-design-in-under-two-years","publisher":"Digest AI","title":"OpenAI's Jalapeño Chip: AI-Driven Design in Under Two Years","datePublished":"2026-09-14T14:06:31Z","url":"https://digestai.news/story/openai-s-jalapeno-chip-ai-driven-design-in-under-two-years"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}