TPU v7 Ironwood vs B300 Blackwell Ultra vs Ascend 950DT

Pyyan / Compare / TPU v7 Ironwood vs B300 Blackwell Ultra vs Ascend 950DT

3 of 5

GPU & AI Chips · verified 24 Sept 2026

×TPU v7 IronwoodGooglecurrent

×B300 Blackwell UltraNVIDIAcurrent

×HAscend 950DTHuaweicurrent

Specification	TPU v7 Ironwood	B300 Blackwell Ultra	Ascend 950DT
Summary	Inference-oriented, Google Cloud only.	15 PFLOPS dense FP4.	144GB on Huawei's own HBM, and the part China is building gigawatts around.
Memory	192GB HBM3e	288GB HBM3e	144GB HiZQ 2.0
Bandwidth	7.37 TB/s	8 TB/s	4 TB/s
Price	—	~$40–50K (verify)	Above ¥250,000 (about $37K), reported
Vendor	Google	NVIDIA	Huawei
Category	GPU & AI Chips	GPU & AI Chips	GPU & AI Chips
Official	Google ↗	NVIDIA ↗	Huawei ↗

Highlighted rows are where these differ.

Best for GCP inference.

Best for frontier training.

4 TB/s memory bandwidth and a 2 TB/s interconnect, 2.5x the Ascend 910C
HiZQ 2.0 is Huawei's own memory standard rather than a licensed HBM3 or HBM3e part
Indicated price rose 20% to 50% in two months to September 2026, on a memory shortage
Successor Ascend 960DT pulled forward to Q1 2027, Huawei said on 17 September

Best for serving models where Nvidia cannot be bought.