# Groq LPU

9 Hacker News commenters recommended Groq LPU and 2 criticized it, November 2025 to November 2026. It was recommended in 5 of 13 months.

**Verdict:** Recommended for fast, efficient AI inference, using SRAM instead of HBM. Critics point to small per-chip SRAM and hard deployment requirements.

## Why people recommend it

- Faster, more efficient inference than GPUs (4 comments): "A cluster of LPUv2s would probably be faster and cheaper than an Infiniband cluster of Epycs." ([source](https://news.ycombinator.com/item?id=46251208)) "DRAM-based GPUs/TPUs are quite inefficient for inference compared to SRAM-based Groq/Cerebras" ([source](https://news.ycombinator.com/item?id=47109377)) "No HBM needed on this card, and 32x faster inference than today's best GPUs at inference!" ([source](https://news.ycombinator.com/item?id=48044961))
- HBM-free design avoids constrained supply (2 comments): "Groq's technical decisions are unsound in a normal world. But being HBM-free may have some merit in a world where HBM supply is constrained." ([source](https://news.ycombinator.com/item?id=46380672)) "No HBM needed on this card, and 32x faster inference than today's best GPUs at inference!" ([source](https://news.ycombinator.com/item?id=48044961))
- Credible alternative to GPUs (2 comments): "Google's TPU and Groq's LPU are the only real, commercially viable ways to provide all the compute required for the inference people want." ([source](https://news.ycombinator.com/item?id=48452805)) "The most rapid path that Intel has to selling competitive GPUs, would be to licence designs from Groq" ([source](https://news.ycombinator.com/item?id=46889615))

## What people warn about

- Small SRAM per chip (2 comments): "Groq's chips only have 230MB of SRAM per chip vs 80GB on an H100" ([source](https://news.ycombinator.com/item?id=46404612)) "they scale the number of chips to fit a single model when they have no dram" ([source](https://news.ycombinator.com/item?id=46388575))

## Alternatives hackers recommend

- [Google TPU](https://hackerapproved.com/p/google-tpu.md): 73 recommend
- [Taalas HC1](https://hackerapproved.com/p/taalas-chip.md): 59 recommend
- [Taalas hardwired LLM chip](https://hackerapproved.com/p/taalas.md): 45 recommend
- [Cerebras wafer-scale chip](https://hackerapproved.com/p/cerebras-wafer-scale-chip.md): 29 recommend
- [Cerebras CS-4](https://hackerapproved.com/p/cerebras-cs4.md): 12 recommend

## Where to get it

- [Visit the official site](https://groq.com/)

## Cite as

Hacker Approved, "Groq LPU", https://hackerapproved.com/p/groq-lpu/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).