Meta and Broadcom Extend Partnership to Develop Multiple Generations of Custom AI Chips Through 2029

• April 15, 2026

Meta and Broadcom have announced an extension of their partnership, aiming to jointly develop several new generations of custom artificial intelligence (AI) chips. The renewed agreement is set to run through 2029, reinforcing their commitment to advancing AI processing capabilities tailored specifically for Meta’s demanding workloads.

Expanding Capabilities in AI Chip Development

This collaboration between Meta, known for its scale in social networking and AI research, and Broadcom, a leading semiconductor supplier, focuses on delivering specialized silicon solutions capable of managing intense AI operations. Meta has required substantial computational resources to support its AI infrastructure, and Broadcom is tasked with helping produce chips that meet these high performance standards.

In the first phase of this extended plan, Meta emphasizes a key milestone: the provision of more than 1 gigawatt (GW) of compute power. This target points to significant improvements in efficiency and scale, reflecting the growing demands of AI workloads not only for Meta but also for the broader technology environment where custom silicon plays an increasingly important role.

Custom AI chips are increasingly central to reducing reliance on off-the-shelf hardware, allowing companies like Meta to optimize performance and energy use for their specialized needs. Working closely with Broadcom allows Meta to innovate in hardware design tailored specifically for its AI models and data processing demands.

The strategic partnership between these two technology players continues to underline the broader industry trend toward bespoke chip development, a shift driven by the exponential growth of AI applications across sectors. As AI workloads become more complex and resource-intensive, the design of purpose-built processors is critical to maintaining competitive advantage and operational efficiency.

Details on the specific architectures or technical innovations planned for these upcoming machine learning chips have not been publicly disclosed. Likewise, pricing and deployment timelines remain confidential. However, the commitment to maintaining and expanding this partnership through 2029 signals a long-term investment by both companies in shaping the future AI hardware landscape.

Meta’s continued investment in custom chip technology aligns with its broader strategy to scale AI-powered services, including content recommendation, augmented and virtual reality platforms, and large-scale data analytics. Meanwhile, Broadcom benefits from the partnership by positioning itself as a key supplier in the advanced AI semiconductor market.

The expanding collaboration between Meta and Broadcom represents a significant development in the AI hardware sector, showcasing how leading tech firms are aligning with semiconductor manufacturers to meet the evolving needs of AI computation at scale.

Meta and Broadcom have extended their collaboration to design successive generations of custom AI processors with over 1GW compute power commitment.