Broadcom and Meta partner on 2-nanometer AI chips through 2029
Broadcom will support Meta's development of custom artificial intelligence accelerators under an extended partnership running through 2029. The two companies will co-develop Meta's next-generation AI infrastructure using Broadcom's XPU platform and advanced networking technologies.
The deal addresses Meta's growing compute demands as it builds features powered by generative AI across WhatsApp, Instagram, and Threads. Meta founder and CEO Mark Zuckerberg said the partnership will help the company deploy "more than 1 gigawatt of our custom silicon to start and then multiple gigawatts over time."
What the partnership covers
Broadcom will work with Meta on chip design, packaging, and networking infrastructure. The collaboration extends beyond hardware supply to include system-level optimization and research and development.
The effort centers on Meta's Training and Inference Accelerator (MTIA) chips. Meta plans to develop four new MTIA generations within two years, using a modular design that allows new chip launches every six months or less-faster than the industry standard of one to two years.
Broadcom President and CEO Hock Tan will step down from Meta's board to take an advisory role, providing guidance on Meta's custom silicon roadmap and infrastructure investments.
Business context
Meta's AI hardware revenues grew 65 percent year over year to $20 billion during the 12-month period ending November 2, 2025. The company's rapid chip development cycle depends on the kind of tight integration Broadcom's platform enables-coupling logic, memory, and high-speed input-output in ways that allow quick iteration across multiple chip generations.
Broadcom has pursued similar partnerships with other tech companies. In May 2023, the company signed a multibillion-dollar deal with Apple to develop 5G radio frequency components.
For IT and development professionals: Understanding custom silicon strategies and AI infrastructure design is increasingly relevant to your work. Learn more about AI for IT & Development or explore the AI Learning Path for Software Engineers to build skills in AI system architecture and infrastructure optimization.
Your membership also unlocks:
