Meta to deploy self-designed MTIA 450 and 500 inference accelerators through 2027, per Bloomberg
Beth_Kindig · x · 2026-09-18
Bloomberg reports Meta is preparing to deploy its MTIA 450 accelerator (codenamed Arke) into data centers in H1 2027, with the MTIA 500 (codenamed Astrid) following by end of 2027.
Meta says the two chips are expected to be its 'workhorses' for general-purpose inference, while future generations will target improved speed and throughput — part of Meta's push to rely less on external GPUs from Nvidia and AMD.
Related event: Meta to Deploy Its Own MTIA 450/500 AI Chips in 2027(2 posts)→
More from Infra
- SSAD 2026 closing talk: Dirk Hornung on OpenXLA as a ready-to-use end-to-end compiler stack — abursuc · 2026-09-18
- AMD Ryzen 7 CPUs gained over 50% Geekbench performance from 2022 to 2024 — lemire · 2026-09-18
- Lemire: 15% frequency gain can't explain Ryzen's 50% performance jump — lemire · 2026-09-18
- Zen's IPC jump: from 6 to 8 instructions per cycle, bigger caches and ROB — lemire · 2026-09-18
- Zen 5 leads the pack on SIMD data parallelism — lemire · 2026-09-18
- ADSP Episode 304: Mark Saroufim on the agentic era, autoresearch and productionizing GPU kernels — blelbach · 2026-09-18