Compare XPUs

Select up to 5 XPUs to compare side-by-side

Select XPUs to Compare

Clear all (4)Jump to results

Filter by Vendor

Showing 72 XPUs • 4 selected

Alibaba

Hanguang 800

AMD

MI100

23.1 TFLOPs

AMD

MI210

181 TFLOPs

AMD

MI250X

383 TFLOPs

AMD

MI300A

980.6 TFLOPs

AMD

MI300X

1,307 TFLOPs

AMD

MI325X

1,400 TFLOPs

AMD

MI350X

2,100 TFLOPs

AMD

Radeon PRO W7900

122 TFLOPs

AMD

Radeon RX 7900 XT

104 TFLOPs

AMD

Radeon RX 7900 XTX

122 TFLOPs

AWS

Inferentia2

190 TFLOPs

AWS

Trainium

190 TFLOPs

AWS

Trainium2

680 TFLOPs

Baidu

Kunlun II

Biren Technology

BR100

Cambricon

MLU370

256 TFLOPs

Cerebras

WSE-3

Enflame Technology

CloudBlazer T20

Etched

Sohu

10,000 TFLOPs

FuriosaAI

Warboy

Google

TPU v4

275 TFLOPs

Google

TPU v5e

197 TFLOPs

Google

TPU v5p

459 TFLOPs

Google

TPU v6e (Trillium)

918 TFLOPs

Graphcore

Bow IPU

Graphcore

IPU-M2000

Groq

LPU Inference Engine

Huawei

Ascend 910B

Iluvatar CoreX

BI-V150

300 TFLOPs

Intel

Data Center GPU Max 1100

177 TFLOPs

Intel

Data Center GPU Max 1550

419 TFLOPs

Intel Habana

Gaudi 2

432 TFLOPs

Intel Habana

Gaudi 3

1,835 TFLOPs

Multi-Metric Comparison

Relative performance across 5 key metrics (normalized to 100 = best in comparison)

Compute Performance (BF16)

Memory Capacity

Power Consumption

Power Efficiency

Specifications

Specification	Google TPU v5p	Tenstorrent Wormhole	NVIDIA GB200 Superchip	Qualcomm Cloud AI 100
Architecture	TPU v5	Tensix Core	Blackwell	Qualcomm AI
Form Factor	Mezzanine	—	SXM	PCIe
VRAM	—	24 GB	384 GB	16 GB
Memory Bandwidth	—	—	8,000 GB/s	134 GB/s
TFLOPs (FP32)	—	—	90	50
TFLOPs (FP16)	—	—	5,000	400
TFLOPs	459	364	5,000	50
TFLOPs (FP8)	—	—	—	—
TDP	400 W	160 W	2700 W	75 W
Launch Date	Aug 2023	Oct 2023	Mar 2024	Sep 2020

Efficiency Metrics

Metric	TPU v5p	Wormhole	GB200 Superchip	Cloud AI 100
TFLOPs per Watt (FP32-eq)	0.57	1.14	0.93	0.67
Memory Bandwidth per GB	—	—	20.8 GB/s	8.4 GB/s

Performance Equivalence

How many units of each GPU are needed to match the performance of the others?

To match 1x Google TPU v5p

Tenstorrent Wormhole

Compute (FP32-eq)

1.26x

Need 1.26x Wormhole

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.09x

GB200 Superchip is 10.89x faster

Qualcomm Cloud AI 100

Compute (FP32-eq)

4.59x

Need 4.59x Cloud AI 100

To match 1x Tenstorrent Wormhole

Google TPU v5p

Compute (FP32-eq)

0.79x

TPU v5p is 1.26x faster

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.07x

GB200 Superchip is 13.74x faster

VRAM

0.06x

GB200 Superchip has 16.00x more

Qualcomm Cloud AI 100

Compute (FP32-eq)

3.64x

Need 3.64x Cloud AI 100

VRAM

1.50x

Need 1.50x Cloud AI 100

To match 1x NVIDIA GB200 Superchip

Google TPU v5p

Compute (FP32-eq)

10.89x

Need 10.89x TPU v5p

Tenstorrent Wormhole

Compute (FP32-eq)

13.74x

Need 13.74x Wormhole

VRAM

16.00x

Need 16.00x Wormhole

Qualcomm Cloud AI 100

Compute (FP32-eq)

50.00x

Need 50.00x Cloud AI 100

FP32 Compute

1.80x

Need 1.80x Cloud AI 100

VRAM

24.00x

Need 24.00x Cloud AI 100

Memory Bandwidth

59.70x

Need 59.70x Cloud AI 100

To match 1x Qualcomm Cloud AI 100

Google TPU v5p

Compute (FP32-eq)

0.22x

TPU v5p is 4.59x faster

Tenstorrent Wormhole

Compute (FP32-eq)

0.27x

Wormhole is 3.64x faster

VRAM

0.67x

Wormhole has 1.50x more

NVIDIA GB200 Superchip

Compute (FP32-eq)

0.02x

GB200 Superchip is 50.00x faster

FP32 Compute

0.56x

GB200 Superchip is 1.80x faster

VRAM

0.04x

GB200 Superchip has 24.00x more

Memory Bandwidth

0.02x

GB200 Superchip has 59.70x more

Pricing

Price Type	TPU v5p	Wormhole	GB200 Superchip	Cloud AI 100
CAPEX (Street Price)	—	—	—	—
OPEX (per hour)	$5.00/hr	—	—	—
Price per TFLOPs (FP32-eq)	—	—	—	—