← Back to the wire

MTIA 300: Meta’s First Training Chip with Built-in NICs and Communication-Offloading Engines

AnnouncementProductAug 24, 2026

Meta reports that on a 150-billion-parameter production recommendation model running across 40 accelerators, MTIA 300's total communication time was 3.9 times faster than an equivalent GPU cluster. MTIA 300 is the first of Meta's in-house accelerators optimized for training ranking and recommendation models, with network chiplets integrated into the chip package. Meta co-designed the chip with HCCL, its communication library, which compiles collectives into subgraphs executed autonomously by dedicated message engines.

Receipt № 15501 source · awaiting confirmation ◐

Evidence

1source· awaiting independent confirmation

No score is assigned. Sources and their independence are shown in the citation chain below.

Citation chain · 1 source

MetaCompanyMTIA 300ModelHCCLModel
Canonical: https://engineering.fb.com/2026/08/24/networking-traffic/mtia-300-meta-training-chip-built-in-nics/