OpenAI Unveils Jalapeo, Its First Custom AI Chip to Rival Nvidia

OpenAI has introduced its first custom AI chip, Jalapeo, marking a significant challenge to Nvidia's dominance in AI hardware. Jalapeo is designed specifically for inference, the process of answering user queries, which is a critical function in artificial intelligence systems.

Key Highlights

• OpenAI introduces Jalapeo, its first custom AI chip for inference tasks.
• Jalapeo was developed with Broadcom and is already running GPT-5.3-Codex-Spark in labs.
• Early tests show Jalapeo offers better performance per watt than current leading AI chips.
• OpenAI plans to deploy the chip by year-end with server systems built by Celestica.

OpenAI's Jalapeo Chip Details

Jalapeo was developed in partnership with Broadcom and is already running the GPT-5.3-Codex-Spark model inside OpenAI's labs. Unlike Nvidia’s general-purpose AI chips, Jalapeo was tailored to meet OpenAI’s unique requirements. Early tests indicate the chip delivers better performance per watt compared to current leading AI hardware, which could lower the operational costs of services like ChatGPT.

The chip design process took only nine months, with AI tools assisting engineers to accelerate development. After completion, the design was sent to Taiwan’s TSMC for manufacturing. Broadcom’s CEO stated that Jalapeo performs on par with Nvidia’s latest Blackwell processors and Google’s Tensor Processing Units (TPUs), suggesting it could compete at the highest level if results hold at scale.

Industry Context and Future Plans

OpenAI’s move follows similar strategies by Google and Amazon, who have also developed custom AI accelerators to reduce dependence on external suppliers. OpenAI began exploring custom chips in 2023, and Jalapeo is its first public outcome. Rival company Anthropic is also planning to build its own custom AI chip, indicating a broader industry trend.

Jalapeo is the first step in OpenAI’s multi-generation hardware strategy. The company plans to start deploying the chip by the end of the year. Canadian manufacturer Celestica will produce the server systems that will house Jalapeo chips. OpenAI expects improvements in speed, cost, and reliability to result in faster ChatGPT responses, more capable coding assistants, more affordable AI services, and fewer outages during high demand.

Implications for AI Hardware Race

The development of Jalapeo highlights the growing importance of custom hardware in the AI sector. As AI models become more advanced, the underlying chips play a crucial role in performance and efficiency. OpenAI’s investment in custom chips could influence the direction of future AI infrastructure, as companies seek to optimize both software and hardware for better results.