October 14, 2025 – Broadcom has just unveiled a groundbreaking strategic partnership with OpenAI, centered around the development and deployment of custom 10 – gigawatt (GW) – scale AI chips. This ambitious initiative is set to kick off its deployment phase and start refining the network system in the second half of 2026, with the entire project slated for completion by the end of 2029.
According to the official press release, OpenAI will take the lead in designing the AI chips. The two companies will work hand – in – hand throughout the development and deployment processes. By leveraging these self – developed chips and systems, OpenAI aims to directly integrate its cutting – edge model and product development expertise into the hardware products. This integration is expected to unlock unprecedented levels of computing power and intelligence.
The resulting products will make extensive use of Broadcom’s Ethernet and related connectivity solutions. These solutions are specifically tailored to meet the ever – increasing demands of AI. The deployment of these products is anticipated to take place in OpenAI’s own data centers as well as those of its partners.
Sam Altman, the founder and CEO of OpenAI, expressed great enthusiasm about the collaboration. He stated, “Our partnership with Broadcom represents a crucial step in unleashing the full potential of AI infrastructure. It will lead to outcomes that truly benefit both humanity and businesses.”
On the other hand, Tan Hock Eeng, the CEO of Broadcom, also highlighted the significance of this collaboration. He remarked, “This collaboration marks a pivotal moment in humanity’s journey towards artificial general intelligence (AGI). Since the development of ChatGPT, OpenAI has consistently been at the forefront of the AI revolution. We are thrilled to work together on developing and deploying these 10 – GW next – generation AI chips, paving the way for the future of AI.”
