Skip to main content

Groq LPU

Vendor: Groq (acquired by NVIDIA)

Category: LPU Language Processing Unit

Architecture: TSP (Tensor Streaming Processor)

Introduction

Groq LPU (Language Processing Unit) is a processor purpose-built for large language model inference. Using a deterministic architecture with extremely low inference latency, its token generation speed for models like LLaMA far exceeds traditional GPUs. In December 2025, NVIDIA acquired Groq for approximately $20 billion, with LPU technology to be integrated into NVIDIA's product line. The third-generation LPU (LP30) will be released in 2026.

Specifications

ModelComputeMemoryInterfaceTDPProcess
LPU v1750 TOPS (INT8) / 188 TFLOPS (FP16)230MB on-chip SRAMEthernet Interconnect300W14nm
LPU v3 (LP30)1.2 PFLOPS (FP8)500MB on-chip SRAMNVLink-C2CTBASamsung 4nm

Official Website

Visit Official Website

Driver Downloads

Linux

OS Support

WindowsLinuxmacOSAndroid
✅ (GroqCloud API)

Version History

VersionRelease DateDescription
LPU Runtime 1.02024Llama 3 8B reaches 800+ tokens/s

Performance Benchmarks

ModelTaskPerformance Metric
LPU v1Llama 2 70B Inference~330 tok/s (FP16, GroqCloud)
LPU v1Mixtral 8x7B Inference~180 tok/s/chip
LPU v1Llama 3 8B Inference~800 tok/s

Pricing Information

ModelReference PriceNotes
LPU v1Free APIGroqCloud free tier
LPU v1EnterpriseGroqCloud pay-as-you-go

Quick Setup

GroqCloud (API)

pip install groq

LPU v1 is not sold separately; accessible only via the GroqCloud API.

Code Examples

Python (Groq API)

from groq import Groq

client = Groq(api_key="your-key")
response = client.chat.completions.create(
model="llama-3.1-70b-versatile",
messages=[{"role": "user", "content": "Hello"}],
max_tokens=100
)
print(response.choices[0].message.content)

Architecture Highlights

  • TSP (Tensor Streaming Processor): A tensor processor optimized for sequential execution, completing one full matrix operation per clock cycle
  • Deterministic Latency: Inference latency is fully predictable, ideal for real-time AI services
  • SRAM-Intensive: 230MB on-chip SRAM, avoiding DRAM access latency

Model Compatibility

Model/FrameworkSupport StatusNotes
Llama Series✅ NativeOfficially deployed by Groq
Mixtral✅ NativeMoE model support
Large Language ModelsGroqCloud API
CNN/TrainingInference only, Transformer only

Large-Scale Cluster Deployments

Based on global AI supercomputing cluster statistics, Groq LPU has accumulated over 19,725 chips deployed across 1 cluster in publicly disclosed deployments.

Chip Model Statistics

Chip ModelTotal DeployedClusters
GroqChip LPU v119,7251

Notable Deployment Clusters Top 10

#Cluster NameTotal ChipsChip ModelOperator
1Aramco Groq Inference Cluster19,725GroqChip LPU v1 ×19,725Saudi Aramco, Saudi Arabia

If you are evaluating alternatives, the following products may also fit your scenario: