DEPLOYDatabase

Chip comparison

Google TPU v6 (Trillium) vs Tesla HW4 (FSD Computer 2)

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 v6 (Trillium)Tesla HW4 (FSD Computer 2)
Process nodeSamsung 7LPP
TDP200 W90 W
Memory32 GB HBM316 GB GDDR6
Memory bandwidth1,600 GB/s
BF16 (dense)918 TFLOPS
FP8 (dense)1,836 TFLOPS
INT8 (dense)1,836 TOPS300 TOPS
Form factorOAM (Trillium pod, up to 256 chips)Auto-board (dual-SoC)
Announced2024-05-142023-01-01
Released2024-12-012023-02-01

Robots running Google TPU v6 (Trillium)

No robots publicly running it yet.

Robots running Tesla HW4 (FSD Computer 2)

Data centers with Google TPU v6 (Trillium)

Data centers with Tesla HW4 (FSD Computer 2)

No data centers publicly running it yet.

Common questions

Google TPU v6 (Trillium) vs Tesla HW4 (FSD Computer 2): which has more memory?

Google TPU v6 (Trillium) carries more HBM (Google TPU v6 (Trillium): 32 GB HBM3; Tesla HW4 (FSD Computer 2): 16 GB GDDR6). Memory capacity sets the largest model that fits on one chip without partitioning; memory bandwidth sets inference throughput.

Google TPU v6 (Trillium) vs Tesla HW4 (FSD Computer 2): which is more widely used?

Google TPU v6 (Trillium): 0 robots and 1 data center publicly running it. Tesla HW4 (FSD Computer 2): 3 robots and 0 data centers. Empty here means no operator has said publicly, not zero adoption in the market.

See also: every chip comparison · Google TPU v6 (Trillium) full page · Tesla HW4 (FSD Computer 2) full page.