Meta’s new AI chips will begin production in September

2026-07-09 · Source: TechCrunch · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Intermediate, quick

Summary

Meta is set to begin production of its latest AI-specific chips in September, aiming to reduce GPU costs amidst a component shortage. At least one chip completed its testing phase in approximately six weeks. Developed under the Meta Training and Inference Accelerator (MTIA) program, these chips are designed with a modular approach, building on previous generations and incorporating evolving AI workload insights. Meta is collaborating with Broadcom on design, TSMC for manufacturing, Samsung for RAM, Sandisk for storage, and Sumitomo Electric for fiber-optic equipment. The company plans to use MTIA chips for training models for ranking and recommendation algorithms, broader AI workloads, and inference across its applications. This initiative is part of Meta's substantial investment in AI infrastructure, with projected capital expenditures between \$125 billion and \$145 billion this year, including plans to deploy 7 gigawatts of compute capacity.

Key takeaway

For AI Architects evaluating compute infrastructure, Meta's strategy signals a critical shift towards custom silicon. You should assess whether your organization's specific AI workloads, particularly for ranking and recommendations, warrant investing in or exploring specialized hardware development. This approach can significantly reduce long-term GPU costs and enhance performance for proprietary models, but requires substantial upfront capital expenditure and deep engineering expertise.

Key insights

Major tech companies are developing custom AI chips to reduce costs and optimize for specific workloads.

Principles

• Modular chip design adapts to evolving AI needs.
• Diversifying compute suppliers mitigates shortages.
• In-house chip development targets specific AI workloads.

In practice

• Use custom chips for ranking and recommendation algorithms.
• Integrate modular chiplets for future-proofing AI hardware.
• Partner with multiple vendors for component supply chain.

Topics

• AI Chips
• Custom Silicon
• Meta MTIA
• AI Infrastructure
• Supply Chain Diversification
• Inference Accelerators

Best for: Director of AI/ML, AI Architect, AI Hardware Engineer

Related on AIssential

• Meta Is Developing 4 New Chips to Power Its AI and Recommendation Systems — Meta is developing custom MTIA chips to power its internal generative AI and content ranking systems.
• Why everyone from OpenAI to SpaceX is building their own chips (and turning up the heat on Nvidia) — Major tech firms are developing custom AI chips to reduce supplier dependence and optimize performance.
• Meta to Spend Billions on AMD Gear, AI Scare Trade Continues — Diversification in AI chip procurement and specialized AI agents are reshaping market dynamics and software valuations.
• India’s $11B Semiconductor Bet, AI Power Shift: Meta Chips & The AI Cost Crisis — The global AI landscape is marked by a race for custom silicon, rising inference costs, and national strategic investments.
• Meta ramps up AI push with expanded Broadcom agreement - Latest news from Azerbaijan — Meta and Broadcom are deepening their custom AI chip partnership to scale Meta's AI computing infrastructure.
• Custom silicon is critical to scaling next-gen AI. We’re detailing the evolution of the Meta Training and Inference Accelerator (MTIA), our homegrown silicon family designed to power the next era of AI experiences. Traditional chip cycles span years, but model a - x.com — Rapid custom silicon iteration is vital for scaling AI given fast-changing model architectures.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by TechCrunch.