AWS logo vs

AWS Trainium2 vs NVIDIA Groq 3 LPU

The Trainium2 delivers 8% more FP8 throughput than the Groq 3 LPU (1,299 vs 1,200 TFLOPS dense).

Trainium2: NeuronCore-v3, 2024 Groq 3 LPU: Groq LPU, 2026
FP8 dense lead
+8%
Trainium2: 1,299 vs 1,200 TFLOPS
Download comparison sheet

Share this page

AWS Trainium2 vs NVIDIA Groq 3 LPU

A post with this link shows the card on the Image tab as its preview.

Printable version →
AWS logo

AWS Trainium2

Leads on FP8 Full specs →

What is the AWS Trainium2?

The Trainium2 is an ASIC from AWS on the NeuronCore-v3 architecture, an accelerator built for AI workloads. Trainium2 full specifications →

Can I rent the Trainium2 in the cloud?

Flopper is not tracking any live rental listings for the Trainium2 yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the Trainium2 compare with other accelerators?

This page sets the Trainium2 against the Groq 3 LPU. The Trainium2 is the older generation of the two. This is a cross-vendor matchup, AWS against NVIDIA. Flopper also compares it head to head with the H100 SXM5 80GB, across every published precision, memory and power. Trainium2 vs H100 SXM5 80GB →

Where do the Trainium2 figures come from?

Every Trainium2 figure on this page traces to AWS's own datasheet or specification table. How Flopper verifies specifications →

NVIDIA Groq 3 LPU

Full specs →

What is the NVIDIA Groq 3 LPU?

The Groq 3 LPU is an LPU from NVIDIA on the Groq LPU architecture, an accelerator built for language-model inference. Groq 3 LPU full specifications →

Can I rent the Groq 3 LPU in the cloud?

Flopper is not tracking any live rental listings for the Groq 3 LPU yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the Groq 3 LPU compare with other accelerators?

This page sets the Groq 3 LPU against the Trainium2. The Groq 3 LPU is the newer generation of the two. This is a cross-vendor matchup, NVIDIA against AWS. Flopper also compares it head to head with the GroqChip 1, across every published precision, memory and power. Groq 3 LPU vs GroqChip 1 →

Where do the Groq 3 LPU figures come from?

Every Groq 3 LPU figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

Specifications

AWS logo
Trainium2
Groq 3 LPU
Architecture
NeuronCore-v3Groq LPU
Launch Year
20242026
Form Factor
——
Memory
96 GB—
Memory Bandwidth
2.9 TB/s—
TDP
——
Process Node
——

Performance (TFLOPS)

AWS logo
Trainium2
Groq 3 LPU
No verified data available No verified data available
181 TFLOPS No verified data available
667 TFLOPS
2,563 TFLOPS sparse
No verified data available
667 TFLOPS
2,563 TFLOPS sparse
No verified data available
667 TFLOPS
2,563 TFLOPS sparse
No verified data available
1,299 TFLOPS
2,563 TFLOPS sparse
1,200 TFLOPS
FP6
No verified data available No verified data available
No verified data available No verified data available
No verified data available No verified data available

Figures are dense. Where a vendor publishes a figure it sits under the dense one and never decides a winner. Hover any precision for its definition, or read the GPU glossary.

FLOPS by Precision

What actually differs

The Groq 3 LPU is the newer part: Groq LPU (2026) against the Trainium2's NeuronCore-v3 (2024).

Weight the FP8 and FP16 rows for LLM work and the FP64 row for HPC.

New to the terms? The glossary explains FP32, TF32, BF16, FP16, FP8, Sparsity and TFLOPS. For how the generations relate, see NVIDIA architectures explained.

Frequently asked questions

Is the Trainium2 faster than the Groq 3 LPU?

At FP8 precision the Trainium2 reaches 1,299 TFLOPS dense against 1,200 TFLOPS for the Groq 3 LPU. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the Trainium2 or the Groq 3 LPU?

Choose the Trainium2 for training and high-throughput inference: it has 8% more FP8 throughput than the Groq 3 LPU.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.

© 2026 Flopper.io - Compare the GPUs Powering AI