Broadcom and OpenAI Launch 'Jalapeño' AI Accelerator
Broadcom (AVGO) and OpenAI announced the launch of 'Jalapeño,' OpenAI's first custom large language model inference accelerator, co-developed with Celestica. The accelerator is already running GPT-5.3-Codex-Spark workloads at production target frequency and power.
Broadcom (AVGO) and OpenAI announced the launch of 'Jalapeño,' OpenAI's first custom large language model inference accelerator, co-developed with Celestica. The accelerator is already running GPT-5.3-Codex-Spark workloads at production target frequency and power.
The Product
Jalapeño is co-designed around real-world LLM needs, paired with Broadcom's Tomahawk networking and system integration. This creates an integrated AI compute stack that could reshape how hyperscalers source and deploy AI hardware.
Pricing and Availability
Pricing and commercial availability have not been disclosed. The accelerator is expected to be targeted at hyperscale cloud providers.
Competition
Jalapeño enters a competitive market dominated by NVIDIA (NVDA) GPUs, with AMD (AMD) and Intel (INTC) also offering alternatives. However, the custom design for OpenAI may give Broadcom a unique edge.
Potential Impact on Broadcom
This launch strengthens Broadcom's position in the AI accelerator market, potentially driving revenue growth and diversification. However, competition remains intense, and financial details are yet to be released.
Frequently Asked Questions
Found this useful? Share it