DEPLOYDatabase

Chip comparison

Google TPU v5p vs Meta MTIA v2

Side-by-side specs

Straight from each vendor's datasheet. Winner in bold: higher throughput, memory, or transistor count; lower power draw or price. Sparse (2:4) throughput roughly doubles the FP8/INT8 numbers shown.

FieldGoogle TPU v5pMeta MTIA v2
Process nodeTSMC N5
TDP300 W90 W
Memory95 GB HBM2e128 GB LPDDR5
Memory bandwidth2,765 GB/s205 GB/s
BF16 (dense)459 TFLOPS
INT8 (dense)918 TOPS708 TOPS
Form factorOAM (per-chip in v5p pod)PCIe (internal Meta)
Announced2023-12-062024-04-10
Released2024-01-012024-05-01

Robots running Google TPU v5p

No robots publicly running it yet.

Robots running Meta MTIA v2

No robots publicly running it yet.

Data centers with Google TPU v5p

Common questions

Google TPU v5p vs Meta MTIA v2: which has more memory?

Meta MTIA v2 carries more HBM (Google TPU v5p: 95 GB HBM2e; Meta MTIA v2: 128 GB LPDDR5). Memory capacity sets the largest model that fits on one chip without partitioning; memory bandwidth sets inference throughput.

Google TPU v5p vs Meta MTIA v2: which is more widely used?

Google TPU v5p: 0 robots and 1 data center publicly running it. Meta MTIA v2: 0 robots and 2 data centers. Empty here means no operator has said publicly, not zero adoption in the market.

See also: every chip comparison · Google TPU v5p full page · Meta MTIA v2 full page.