OpenAI and Broadcom unveil LLM-optimized inference chip
OpenAI and Broadcom announced Jalapeño, described by OpenAI as its first custom inference accelerator and the first generation of a planned multi-generation compute platform.
Source metadata
- Publisher
- OpenAI
- Source type
- official record
- Verification
- verified
- Credibility tier
- direct statement
- Published
- Jun 24, 2026
- Captured / reviewed
- Aug 7, 2026
- Confidence
- medium
- Content status
- verified
- Original host
- openai.com
Citation
OpenAI. "OpenAI and Broadcom unveil LLM-optimized inference chip." Published June 24, 2026. Accessed August 7, 2026. https://openai.com/index/openai-broadcom-jalapeno-inference-chip.
Notes
Source detail pages are citation records, not standalone analysis. Related topic, entity, and archive links are editorial context and do not upgrade placeholder records into published claims.
Official product and partnership announcement. Performance-per-watt statements were preliminary, and deployment timing and gigawatt-scale plans were forward-looking; independent benchmarks and later deployment records are required before treating them as realized.
Why this matters
Published claims
No published public claims cite this source yet.
Archive context
No public archive entry explicitly references this source.
Indexed topics
Ecosystem
OpenAI / Sam Altman Ecosystem
Repository-assigned topic
Infrastructure
OpenAI Infrastructure & Compute
Repository-assigned topic
Infrastructure subcluster
GPU Supply Chain
Repository-assigned topic
Infrastructure subcluster
Data Centers
Repository-assigned topic
Indexed entities
OpenAI
organizationRepository-assigned entity