AI hardware
Meta MTIA v1 (MTIA 100)
FP16 throughput
5.1×10¹³ FLOP/s
Memory
122 MB
TDP
25 W
Meta MTIA v1 (MTIA 100) is an AI accelerator designed by Meta (Other), released in May 2023. It is built on a 7 nm process at TSMC.
It delivers 5.1×10¹³ FLOP/s of tensor FP16/BF16 throughput. It carries 122 MB of memory at 745 GB/s of bandwidth. Its thermal design power is 25 W.
Full record
- Manufacturer
- Meta
- Type
- Other
- Release date
- May 18, 2023
- Tensor FP16/BF16
- 5.1×10¹³ FLOP/s
- FP32
- 1.6×10¹² FLOP/s
- INT8
- 10¹⁴ OP/s
- Memory
- 122 MB
- Memory bandwidth
- 745 GB
- TDP
- 25 W
- Process node
- 7 nm
- Die size (mm2)
- 373
- 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