Meta Broadcom deal extends to 2029 as custom AI chips push reaches multi-gigawatt scale

Reviewed byNidhi Govil

14 Sources

Share

Meta and Broadcom announced an expanded partnership running through 2029 to design custom AI chips, starting with over 1 gigawatt of computing capacity. Broadcom CEO Hock Tan is stepping down from Meta's board to become a chip strategy adviser. The deal marks Meta's commitment to in-house silicon as it competes with Google and OpenAI in the AI infrastructure race.

Meta Broadcom Deal Expands Through 2029 for Custom AI Chips

Meta Platforms and Broadcom announced an expanded partnership on Tuesday to design and build custom AI chips that will power the social media giant's artificial intelligence efforts through 2029

1

2

. The multi-year agreement includes a commitment to develop chips providing more than 1 gigawatt of computing capacity, enough to power roughly 750,000 US homes

5

. This initial deployment represents only the first phase of a sustained, multi-gigawatt rollout as Meta races to build AI infrastructure capable of delivering what Mark Zuckerberg calls "personal superintelligence" to billions of users across its platforms

2

.

Source: SiliconANGLE

Source: SiliconANGLE

Hock Tan Steps Down from Board to Focus on Chip Strategy

In a significant development accompanying the partnership expansion, Broadcom Chief Executive Officer Hock Tan announced he will leave Meta's board of directors when his term expires at the company's next annual meeting

1

. Tan, who joined Meta's board in early 2024, will transition into an advisory role focused specifically on Meta's custom chip strategy

5

. Wolfe Research analyst Chris Caso noted that Tan's decision to step down from Meta's board is an important development, suggesting the duration of the partnership between the two companies may be longer than explicitly stated

3

. Broadcom shares rose 3.4% in extended trading following the announcement, jumping nearly 4% in Wednesday trading

2

3

.

MTIA Chips and In-House AI Silicon Strategy

The partnership centers on Meta's Training and Inference Accelerator program, known as MTIA, which the company unveiled as part of a roadmap of four new chips it is making in-house

2

. Under the deal, Broadcom will provide technology supporting Meta's custom artificial intelligence accelerators, with the companies sharing a roadmap to co-design and scale hardware aimed at delivering real-time generative AI features

2

. The new MTIA chips will be the first AI silicon manufactured on a 2-nanometer process, representing a significant technical advancement

4

5

. Meta introduced MTIA in 2023 and added four new versions in March, demonstrating a quick turnaround between chip iterations

4

. The first chip in the program, the MTIA 300, already runs Meta's ranking and recommendation systems across Facebook, Instagram, and other apps, with three further chip generations planned through 2027

5

.

Source: Quartz

Source: Quartz

Reducing Dependence on Nvidia and AMD GPUs

The deal arrives as large cloud providers and hyperscalers seek ways to rely less on expensive, scarce GPUs from Nvidia and AMD by designing their own ASIC chips tailored to specific AI workloads

4

. While custom chips offer less flexibility than general-purpose GPUs, they can be cheaper and more efficient when running a fixed set of AI tasks

4

. Google launched the first major hyperscaler ASIC effort with its Tensor Processing Units in 2015, followed by Amazon's custom chips in 2018

4

. Unlike those companies, which expose their accelerators through cloud platforms, Meta uses its MTIA silicon entirely for internal workloads to power the AI features and recommendation systems that underpin its advertising business

5

. The Broadcom agreement follows Meta's multi-gigawatt GPU deals with AMD and Nvidia and a new custom-chip partnership with Arm Holdings

4

.

Multi-Gigawatt Rollout and Data Centers Expansion

Broadcom confirmed that Meta has committed to an initial deployment of 1 gigawatt of MTIA capacity and plans to ramp to multiple gigawatts of Broadcom-based accelerators as its AI footprint grows

4

. During Broadcom's March earnings call, CEO Hock Tan addressed recent analyst reports by stating, "Now, contrary to recent analyst reports, Meta's custom accelerator, MTIA roadmap is alive and well. We're shipping now and, in fact, for the next generation XPUs, we will scale to multiple gigawatts in 2027 and beyond"

4

. Meta plans to host this mix of GPUs and accelerators across 31 data centers, including 27 in the US

4

. Broadcom is working with Meta on chip design, advanced packaging, and networking for MTIA, giving the chipmaker a broad role in Meta's AI infrastructure

4

.

Source: Benzinga

Source: Benzinga

Massive Capital Expenditure and Industry Competition

The silicon strategy sits inside a much larger capital plan. In January, Meta said it could spend up to $135 billion on artificial intelligence this year as it tries to stay on pace with Google, Amazon, Anthropic, and OpenAI

4

5

. Mark Zuckerberg stated that Meta was partnering with Broadcom "across chip design, packaging, and networking to build out the massive computing foundation we need to deliver personal superintelligence to billions of people"

5

. Goldman Sachs analyst James Schneider wrote that the announcement demonstrates Broadcom's increasing traction with its XPU platform and networking solutions across major US hyperscalers, providing a broader and more diverse customer base with high exposure to both enterprise and consumer AI

3

. Broadcom has a separate long-term agreement with Google to develop and supply future TPUs, and beginning in 2027 Anthropic is slated to access about 3.5 gigawatts of that TPU capacity

4

.

Today's Top Stories

TheOutpost.ai

Your Daily Dose of Curated AI News

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Spend less time searching for the latest in AI and get straight to action.

© 2026 Triveous Technologies Private Limited
Instagram logo
LinkedIn logo