DEPLOYDatabase

Chip comparison

Google TPU v4 vs NVIDIA DRIVE Thor

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 v4NVIDIA DRIVE Thor
Process nodeTSMC N7TSMC 4NP
TDP192 W130 W
Memory32 GB HBM2128 GB LPDDR5X
Memory bandwidth1,200 GB/s
BF16 (dense)275 TFLOPS
FP8 (dense)2,000 TFLOPS
INT8 (dense)275 TOPS2,000 TOPS
Form factorOAM (per-chip in v4 pod)Auto-board
Announced2021-05-182022-09-20
Released2022-05-012025-06-01

Robots running Google TPU v4

No robots publicly running it yet.

Data centers with NVIDIA DRIVE Thor

No data centers publicly running it yet.

Common questions

Google TPU v4 vs NVIDIA DRIVE Thor: which is faster for training?

NVIDIA DRIVE Thor has 7.27x the dense FP16/BF16 throughput of the other (Google TPU v4: 275 TFLOPS; NVIDIA DRIVE Thor: 2,000 TFLOPS). Real-world training speed also depends on memory bandwidth, interconnect, and software stack: see the spec diff for those inputs.

Google TPU v4 vs NVIDIA DRIVE Thor: which has more memory?

NVIDIA DRIVE Thor carries more HBM (Google TPU v4: 32 GB HBM2; NVIDIA DRIVE Thor: 128 GB LPDDR5X). Memory capacity sets the largest model that fits on one chip without partitioning; memory bandwidth sets inference throughput.

Google TPU v4 vs NVIDIA DRIVE Thor: which is more power-efficient?

Dense FP16/BF16 TFLOPS per watt: Google TPU v4 1.43 (275 ÷ 192 W); NVIDIA DRIVE Thor 15.38 (2,000 ÷ 130 W). NVIDIA DRIVE Thor wins at the die level; system-level efficiency also depends on cooling and interconnect.

Google TPU v4 vs NVIDIA DRIVE Thor: which is more widely used?

Google TPU v4: 0 robots and 2 data centers publicly running it. NVIDIA DRIVE Thor: 11 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 v4 full page · NVIDIA DRIVE Thor full page.