✨ From vibe coding to vibe deployment. UBOS MCP turns ideas into infra with one message.

Learn more
Andrii Bidochko
  • Updated: March 11, 2026
  • 1 min read

Meta Unveils Four New MTIA AI Chips to Power Generative AI and Content Ranking

Meta has announced the launch of four new MTIA (Meta AI) chips—MTIA 300, 400, 450, and 500—designed to accelerate generative AI models and improve content recommendation systems across its platforms. Built on a Broadcom‑designed RISC‑V architecture and fabricated by TSMC, the chips aim to give Meta greater control over its AI compute stack while still leveraging external vendors for certain workloads.

The MTIA 300 is already in production and will be used for training large‑scale AI models. The subsequent chips (400, 450, and 500) target inference workloads, offering progressively larger memory capacities and higher performance. Meta expects the MTIA 400 to be available by 2027, with the higher‑end models following later.

This move reflects Meta’s strategic push to own more of its AI infrastructure, reducing reliance on third‑party hardware providers such as Nvidia, AMD, and Google. By integrating these custom chips, Meta hopes to lower operational costs, improve latency, and enhance the personalization of feeds, ads, and other user‑facing experiences.

Read the original story on Wired for full details.

Related reads on ubos.tech: AI Chip Architecture, Meta’s AI Strategy.


Andrii Bidochko

CTO UBOS

Andrii Bidochko is an AI entrepreneur and researcher focused on AI agents, reinforcement learning, and autonomous systems. He writes about the technologies shaping the future of machine intelligence, from frontier models and agent architectures to real-world AI applications.

Sign up for our newsletter

Stay up to date with the roadmap progress, announcements and exclusive discounts feel free to sign up with your email.

Sign In

Register

Reset Password

Please enter your username or email address, you will receive a link to create a new password via email.