# NVIDIA GB10

16 Hacker News commenters recommended NVIDIA GB10 and 13 criticized it, November 2025 to November 2026. It was recommended in 8 of 13 months.

**Verdict:** Recommended for running local AI models and Linux development in a small box; the most common caveat is high price and slow memory bandwidth.

## Why people recommend it

- Runs local AI models well (4 comments): "For personal needs I use a local Qwen3.8-Next-Flash setup on a GB10 cluster." ([source](https://news.ycombinator.com/item?id=49673633)) "It is great for inference for single user/single session." ([source](https://news.ycombinator.com/item?id=48355929)) "unreasonably effective on the GB10 IMO" ([source](https://news.ycombinator.com/item?id=48564264))
- Cheap enough to cluster (2 comments): "new AI desktops with GB10s. They are relatively cheap and you can cluster them" ([source](https://news.ycombinator.com/item?id=48639861)) "For personal needs I use a local Qwen3.8-Next-Flash setup on a GB10 cluster." ([source](https://news.ycombinator.com/item?id=49673633))
- Efficient power use (2 comments): "I mean the GB10 is pretty efficient for the power it has, but imho is nowhere near the power efficiency of Apple Silicon" ([source](https://news.ycombinator.com/item?id=48356609)) "gb10 uses way less power." ([source](https://news.ycombinator.com/item?id=46322979))

## What people warn about

- Memory bandwidth is slow (3 comments): "GB10 boxes have way more compute than they have memory bandwidth" ([source](https://news.ycombinator.com/item?id=49794245)) "memory bandwidth is not just slower than on a GPU, as expected, but is significantly slower than Apple’s unified memory" ([source](https://news.ycombinator.com/item?id=48425003)) "the GB10 is incredibly bandwith starved. You get 128gb ram, but only 270GB/s bandwidth" ([source](https://news.ycombinator.com/item?id=47543418))
- Expensive compared with alternatives (4 comments): "The GB10 itself is pretty good and I love using mine for broad Linux development. But it's too expensive for consumer level pricing" ([source](https://news.ycombinator.com/item?id=48355902)) "The GB10 workstations are not that competitive in price with M5 MacBooks and the Mac Studio" ([source](https://news.ycombinator.com/item?id=47396328)) "at about $4k per GB10, that's $36k for a 1TB cluster" ([source](https://news.ycombinator.com/item?id=48642864))
- Slower than Apple Silicon (2 comments): "I mean the GB10 is pretty efficient for the power it has, but imho is nowhere near the power efficiency of Apple Silicon" ([source](https://news.ycombinator.com/item?id=48356609)) "It is not even close. Like 0.75x SPECint 2017 of the M5." ([source](https://news.ycombinator.com/item?id=48362057))

## Alternatives hackers recommend

- [Apple Mac mini](https://hackerapproved.com/p/mac-mini.md): 390 recommend
- [NVIDIA DGX Spark](https://hackerapproved.com/p/nvidia-dgx-spark.md): 170 recommend
- [AMD Strix Halo](https://hackerapproved.com/p/amd-strix-halo.md): 130 recommend
- [Framework Desktop](https://hackerapproved.com/p/framework-desktop.md): 93 recommend
- [Intel NUC](https://hackerapproved.com/p/intel-nuc.md): 42 recommend

## Where to get it

- [See price on Amazon](https://www.amazon.com/s?k=NVIDIA+DGX+Spark+GB10&tag=hackerapprove-20) (affiliate link)
- [Official site](https://www.nvidia.com/en-us/products/workstations/dgx-spark/)

## Cite as

Hacker Approved, "NVIDIA GB10", https://hackerapproved.com/p/nvidia-gb10/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).