OpenAI Announces First AI Chip to Rival NVIDIA: Jalapeno!

OpenAI, a world leader in artificial intelligence with ChatGPT, continues its investments in this technology without slowing down. OpenAI and Broadcom introduced Jalapeño, the first dedicated AI chip developed for LLM inference.

OpenAI, in collaboration with Broadcom, introduced its first dedicated AI processor, Jalapeño. OpenAI officially unveiled its first dedicated AI processor, which it has been working on for a long time.



OpenAI, the developer of ChatGPT, announced Jalapeño, its first chip specifically designed for large language models. OpenAI and Broadcom introduced Jalapeño Numerous models.

OpenAI has added permissions to open AI models (inference) specifically served with Jalapeno and has begun testing it. OpenAI and Broadcom unveiled Jalapeno, OpenAI's first intelligent processor, developed by OpenAI using linear machine learning models. Jalapeño was handed over to OpenAI CEO Sam Altman and President Greg Brockman by Broadcom President and CEO Hock Tan and President Charlie Kavvas. OpenAI executives stated in their announcements that they designed the chips from scratch in a 9-month period with the help of artificial intelligence models.



The Jalapeno chip, to be manufactured by Broadcom, will be used in ChatGPT and other applications for computationally intensive inference processes, providing users with AI models. OpenAI stated that the first prototypes of Jalapeno ran machine learning workloads at the targeted speed and power consumption levels in laboratory tests.

According to the companies, the chip can offer performance comparable to NVIDIA's Blackwell and Google's Tensor processors. OpenAI also highlighted that Jalapeno's development process reached the production stage in just 9 months.



OpenAI positions Jalapeno not as a one-off chip project, but as the first step in a multi-generational processing platform. OpenAI added that final performance measurements of the AI ​​processor are still ongoing and a detailed technical report will be shared in the coming months.

Broadcom CEO Hock Tan said that Jalapeno's performance is on par with Nvidia's Blackwell architecture processors and Google's TPU solutions. With the introduction of Jalapeno, OpenAI is crossing a significant threshold in establishing its own hardware ecosystem.

By developing its own chip, OpenAI aims to strengthen its platform, ranging from products to models, and to bring intelligence to a wider audience. OpenAI plans to widely deploy Jalapeño in its infrastructure by the end of 2026. OpenAI announced that it was carried out the design process of this new hardware entirely in-house.

The company states that in the initial stages of testing, Jalapeño achieved significantly better performance per watt than the most advanced solutions currently available. OpenAI stated that the chip has been tested in several workloads, including GPT-5.3-Codex-Spark.

OpenAI developed Jalapeno's chip architecture considering memory movement, network resources, model execution patterns, and the service infrastructure that OpenAI uses daily in its own products. With this technology, OpenAI aims to expand access to artificial intelligence and optimize system performance.



Jalapeno is designed as an ASIC (Application-Specific Integrated Circuit), meaning it is a processor developed specifically for a particular task. By developing its own chip, OpenAI aims to reduce its dependence on Nvidia and offer higher throughput.

This development will allow the company to offer its future AI products in a more scalable structure. OpenAI has chosen to develop its own internal AI chips to both reduce costs and create a strong alternative to the NVIDIA chips that currently dominate the market.