OpenAI says its models improved Jalapeño chip design and AMD hardware
- OpenAI
- AMD
- chip design

Changes cut standard-cell area by about 13% in
Jalapeño’s vector unit and 10% in its matrix unitCollaborations with AMD achieved up to a 14 percent PPA improvement on flagship GPU IP
OpenAI is partnering with Synopsys to train GPT-Synopsys on professional EDA tools
OpenAI says it used its models to improve the design of its custom Jalapeño inference chip and hardware from AMD. The reported changes reduced standard-cell area in Jalapeño’s vector unit by about 13% and its matrix unit by about 10%.
OpenAI began using GPT‑o3 mini to help design Jalapeño in early 2025. The company says the optimizations were proven in silicon during the chip’s bring-up, the process of powering on and testing a newly manufactured chip.
In work with AMD, OpenAI reports up to 14% improvement in power, performance and area (PPA) relative to heavily optimized baselines on AMD’s flagship GPU intellectual property. AMD has included some model-generated changes in a recent design sent for manufacturing and added others to production design files for future versions. OpenAI also reports single-digit percentage PPA gains across key blocks of AMD’s upcoming flagship CPU; those gains are not specified as completed production changes.
OpenAI says its models used Synopsys electronic design automation (EDA) tools to check that circuit rewrites preserved behavior and to assess PPA trade-offs. The companies are also partnering to train GPT‑Synopsys to use those tools for chip design. The announcement does not say that GPT‑Synopsys is available as a product.




