DEPLOYDatabase

Chip comparison

Cerebras WSE-3 (Wafer-Scale Engine 3) vs Huawei Ascend 910D

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.

FieldCerebras WSE-3 (Wafer-Scale Engine 3)Huawei Ascend 910D
Process nodeTSMC N5SMIC N+2 (~7nm)
Transistors (B)4000
Die size46225 mm²
TDP23000 W800 W
Memory44 GB on-die SRAM192 GB HBM3
Memory bandwidth21,000,000 GB/s4,000 GB/s
FP16 (dense)62,500 TFLOPS1,200 TFLOPS
BF16 (dense)1,200 TFLOPS
FP8 (dense)2,400 TFLOPS
Form factorwafer-scale (single-wafer system)OAM
Announced2024-03-132025-04-01
Released2024-04-012025-12-01

Robots running Cerebras WSE-3 (Wafer-Scale Engine 3)

No robots publicly running it yet.

Robots running Huawei Ascend 910D

No robots publicly running it yet.

Data centers with Cerebras WSE-3 (Wafer-Scale Engine 3)

No data centers publicly running it yet.

Data centers with Huawei Ascend 910D

No data centers publicly running it yet.

Common questions

Cerebras WSE-3 (Wafer-Scale Engine 3) vs Huawei Ascend 910D: which is faster for training?

Cerebras WSE-3 (Wafer-Scale Engine 3) has 26.04x the dense FP16/BF16 throughput of the other (Cerebras WSE-3 (Wafer-Scale Engine 3): 62,500 TFLOPS; Huawei Ascend 910D: 2,400 TFLOPS). Real-world training speed also depends on memory bandwidth, interconnect, and software stack: see the spec diff for those inputs.

Cerebras WSE-3 (Wafer-Scale Engine 3) vs Huawei Ascend 910D: which has more memory?

Huawei Ascend 910D carries more HBM (Cerebras WSE-3 (Wafer-Scale Engine 3): 44 GB on-die SRAM; Huawei Ascend 910D: 192 GB HBM3). Memory capacity sets the largest model that fits on one chip without partitioning; memory bandwidth sets inference throughput.

Cerebras WSE-3 (Wafer-Scale Engine 3) vs Huawei Ascend 910D: which is more power-efficient?

Dense FP16/BF16 TFLOPS per watt: Cerebras WSE-3 (Wafer-Scale Engine 3) 2.72 (62,500 ÷ 23000 W); Huawei Ascend 910D 3.00 (2,400 ÷ 800 W). Huawei Ascend 910D wins at the die level; system-level efficiency also depends on cooling and interconnect.

See also: every chip comparison · Cerebras WSE-3 (Wafer-Scale Engine 3) full page · Huawei Ascend 910D full page.