AI hardware
Meta MTIA v2 (MTIA 200)
FP16 throughput
1.8×10¹⁴ FLOP/s
Memory
244 MB
TDP
90 W
Meta MTIA v2 (MTIA 200) is an AI accelerator designed by Meta (Other), released in April 2024. It is built on a 5 nm process at TSMC.
It delivers 1.8×10¹⁴ FLOP/s of tensor FP16/BF16 throughput. It carries 244 MB of memory at 931 GB/s of bandwidth. Its thermal design power is 90 W.
Full record
- Manufacturer
- Meta
- Type
- Other
- Release date
- Apr 10, 2024
- Tensor FP16/BF16
- 1.8×10¹⁴ FLOP/s
- FP32
- 2.8×10¹² FLOP/s
- INT8
- 3.5×10¹⁴ OP/s
- Memory
- 244 MB
- Memory bandwidth
- 931 GB
- TDP
- 90 W
- Process node
- 5 nm
- Die size (mm2)
- 421
- Foundry
- TSMC
More from Meta
SourceEpoch AI, 'Machine Learning Hardware'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/machine-learning-hardware. Licensed under CC BY 4.0.
← All ai hardware