OpenAI Unveils First Custom AI Chip to Power ChatGPT and Future Models
OpenAI has introduced its first custom-designed artificial intelligence chip, marking a significant step in the company’s push to strengthen its AI infrastructure and reduce its reliance on third-party hardware suppliers.
The new processor, named Jalapeño, was developed in partnership with semiconductor company Broadcom and is designed to power services such as ChatGPT and OpenAI’s AI coding assistant, Codex. According to the company, the chip is also built to support future generations of AI models and more demanding computing workloads.
In a blog post announcing the launch, OpenAI said early testing has delivered encouraging results.
“While OpenAI is still measuring final performance, early testing shows that Jalapeño will deliver performance per watt substantially better than current state-of-the-art.”
The unveiling follows plans announced last year by OpenAI and Broadcom to jointly develop custom AI chips capable of supporting computing infrastructure equivalent to 10 gigawatts. Jalapeño is the first product to emerge from that collaboration.
Reducing Dependence on Nvidia
The move places OpenAI alongside major technology companies such as Google and Amazon, both of which have invested heavily in designing their own AI processors.
Unlike general-purpose chips, Jalapeño has been engineered specifically for modern large language models, allowing OpenAI to better optimise performance for its AI systems while reducing dependence on chipmaker Nvidia, whose hardware currently dominates the AI industry.
Industry analysts have previously noted that many technology firms are looking to diversify their hardware supply chains.
Speaking to CNN last year, Quilter Cheviot’s Global Head of Technology Research, Ben Barringer, said:
“Nobody wants to be beholden to Nvidia.”
“They are trying to diversify their chip footprint.”
Part of Broader AI Infrastructure Push
The announcement comes as OpenAI continues to invest heavily in the infrastructure needed to support increasingly advanced AI services.
As AI products evolve from simple chatbots into autonomous digital assistants capable of handling continuous workloads, demand for computing power and energy has risen sharply. OpenAI has previously highlighted the need to build large-scale AI infrastructure to lower the cost of operating its models.
The company is also preparing for a potential initial public offering (IPO), which reports suggest could value the AI firm at around $1 trillion, increasing pressure to demonstrate long-term revenue growth and technological leadership.
Focus on Lower Costs and Greater Accessibility
OpenAI says developing more of its technology stack internally will ultimately benefit users by making its AI services more efficient and affordable.
According to the company, custom hardware will help reduce operating costs while expanding access to advanced AI capabilities.
OpenAI co-founder and President Greg Brockman said the new chip represents an important step toward that goal.
“By designing more of the stack ourselves, we can serve more intelligence with greater efficiency and keep pushing advanced AI toward broader access.”
