/ciol/media/media_files/2025/10/15/image-2025-10-15-23-55-39.png)
OpenAI has recently announced a multi-year strategic collaboration with Broadcom to co-develop 10 gigawatts of custom AI accelerators and Ethernet-based networking systems. The collaboration is the critical step of OpenAI that is going to establish a proper infrastructure that would be able to support next-generation AI clusters and frontier models.
The announcement signifies the increased commitment of OpenAI to the design of its own hardware, which is not a new trend as per the long-term vision that includes software, models, and infrastructure to deliver unmatched AI performance.
OpenAI seeks to apply the knowledge of its model training by integrating concepts into its hardware design by creating its own AI accelerators. This may enable the company to enhance efficiency, performance, and scalability besides decreasing reliance on third parties in providing GPUs.
“Partnering with Broadcom is a critical step in building the infrastructure needed to unlock AI’s potential and deliver real benefits for people and businesses,” said Sam Altman, Co-founder and CEO of OpenAI.
Broadcom’s Role:
Broadcom will deploy racks of AI accelerator and networking systems, scaled entirely with its Ethernet and optical connectivity solutions. The deployment phase is expected to begin in the second half of 2026 and conclude by 2029.
“OpenAI has been at the forefront of the AI revolution since the ChatGPT moment,” said Hock Tan, President and CEO of Broadcom. “We are thrilled to co-develop and deploy 10 gigawatts of next-generation accelerators and network systems to pave the way for the future of AI.”
A Push Towards AI Hardware Self-Sufficiency
The collaboration highlights an emerging trend in the industry, namely, AI giants such as OpenAI, Google, or Amazon are starting to develop their own custom chips in order to minimise the latency, decrease the expenses, and enhance the effectiveness of the models.
Following the same route, OpenAI is placing itself in a better position to meet the huge computational needs of the growing ecosystem, which today has more than 800 million weekly active users and is well adopted by enterprises.
“Our collaboration with Broadcom will power breakthroughs in AI and bring the technology’s full potential closer to reality,” said Greg Brockman, co-founder and president of OpenAI.
For Broadcom, this deal reinforces the centrality of Ethernet technology in AI datacenter networking—a choice that promises cost-effective scaling and performance optimisation.
“Our partnership with OpenAI continues to set new industry benchmarks for the design and deployment of open, scalable and power-efficient AI clusters,” added Charlie Kawwas, Ph.D., President of Broadcom’s Semiconductor Solutions Group.
As OpenAI moves from software-driven innovation to full-stack AI infrastructure, this collaboration signals a defining moment in the AI hardware race—one that could reshape how intelligence is built, trained, and deployed across the globe.