{"data":{"slug":"meta-mtia-v2","etype":"chip","name":"Meta MTIA v2","cat":"AI Accelerator","subtitle":"Meta's recommendation-inference chip — custom silicon for the ads engine.","summary":"MTIA is Meta's inference accelerator for the ranking and recommendation models that drive its feed and ads. Version 2 (2024) tripled compute over v1 within a 90 W envelope using large on-chip SRAM instead of HBM — a different design point from LLM chips, tuned for Meta's own workloads. Training-class successors have been reported as in development.","attrs":[{"g":"Overview","l":"Designer","v":"Meta"},{"g":"Overview","l":"Announced","v":"April 2024"},{"b":"Vendor spec","g":"Manufacturing","l":"Foundry / process","v":"TSMC N5"},{"b":"Vendor spec","g":"Compute","l":"INT8","u":"TOPS","v":"354 (dense) / 708 (sparse)"},{"b":"Vendor spec","g":"Memory","l":"On-chip memory","v":"256 MB SRAM; LPDDR5 off-chip"},{"b":"Vendor spec","g":"Power","l":"TDP","u":"W","v":"90"}],"kf":{"node":"TSMC N5","year":2024,"class":"AI Accelerator","tdp_w":90,"designer":"Meta"},"aliases":["MTIA","MTIA v2"],"ticker":null,"exchange":null,"hq":null,"founded":null,"website":null,"status":"Shipping","region":null,"asof":"2026-01-15","sources":[{"url":"https://ai.meta.com/blog/next-generation-meta-training-inference-accelerator-AI-MTIA/","label":"Meta MTIA v2 post"}]},"meta":{"found":true,"count":1,"served_from":"site cache (no live upstream calls)","as_of":"2026-10-14T05:53:08+00:00","disclaimer":"Research and information only — not investment advice. Structured records carry basis labels (vendor spec / reported / estimate); market quotes are delayed/indicative.","attribution":"Records compiled from vendor documentation, official filings and reputable public reporting; market quotes via financial data providers. See /data-sources."}}