# AMD Ryzen AI Max+ 395

44 Hacker News commenters recommended AMD Ryzen AI Max+ 395 and 14 criticized it, November 2025 to November 2026. It was recommended in 11 of 13 months. It meets the Hacker Approved badge rule.

**Verdict:** Recommended for unified memory and integrated GPU that run local AI models for less than a comparable Mac. Common caveat: slow token speeds and tight 128GB memory for big models.

## Why people recommend it

- Unified memory for local AI (3 comments): "AMD made a CPU/GPU soc with unified memory specifically for ai inference." ([source](https://news.ycombinator.com/item?id=48209686)) "For inference you want the one with 128GB soldered RAM." ([source](https://news.ycombinator.com/item?id=47039657)) "would run the 4-bit quant of a trillion param model well enough for personal use" ([source](https://news.ycombinator.com/item?id=48651123))
- Strong integrated GPU (2 comments): "None of them have something even close to the memory bandwidth or integrated GPU performance of the Ryzen AI Max processors." ([source](https://news.ycombinator.com/item?id=49024169)) "first usable integrated graphics solution I've seen" ([source](https://news.ycombinator.com/item?id=47457697))
- Cheaper than comparable Macs (3 comments): "Or get an AMD 395 laptop or mini PC for half the price of an equivalent mac device" ([source](https://news.ycombinator.com/item?id=48355819)) "Desktops with the amd ai+ 395 are still fairly affordable for what they can do." ([source](https://news.ycombinator.com/item?id=48723407)) "If you are not planning to batch, you can run it much cheaper with Ryzen AI Max SoC devices." ([source](https://news.ycombinator.com/item?id=46900747))
- Runs large models locally (3 comments): "it loads a big enough model and spits out tokens just fast enough that you can read what it's doing" ([source](https://news.ycombinator.com/item?id=48388610)) "it does have decent performance at like 130b parameters" ([source](https://news.ycombinator.com/item?id=46225068)) "would run the 4-bit quant of a trillion param model well enough for personal use" ([source](https://news.ycombinator.com/item?id=48651123))

## What people warn about

- Slow token generation speed (3 comments): "It's showing about 8-9 tk/sec which seems quite slow for something like a web search" ([source](https://news.ycombinator.com/item?id=48653018)) "Order of magnitude slower than dedicated GPU but at half the cost." ([source](https://news.ycombinator.com/item?id=48210013)) "Great for that purpose. Not exactly a speed demon with LLMs" ([source](https://news.ycombinator.com/item?id=47929273))
- Tight 128GB memory for big models (3 comments): "The AMD 395 chip supports up to 128GB unified RAM. So still not enough even at 1-bit quant unfortunately." ([source](https://news.ycombinator.com/item?id=48637504)) "Trying to run LLM models somehow makes 128 GiB of memory feel incredibly tight." ([source](https://news.ycombinator.com/item?id=47369087)) "it's only useful for models under 64gb of vram due to the active parameter counts on larger MoEs" ([source](https://news.ycombinator.com/item?id=48734644))
- Setup effort on Linux (2 comments): "I spent a while fiddling around with the right Linux kernel, kernel firmware, ROCm installs, etc." ([source](https://news.ycombinator.com/item?id=48806245)) "Up until recently they didn't even support their cashcow Ryzen 395+ MAX properly." ([source](https://news.ycombinator.com/item?id=47748763))
- Memory bandwidth limits performance (2 comments): "Both the AMD Ryzen 395/495 and the Spark have terribly slow bandwidth" ([source](https://news.ycombinator.com/item?id=50159347)) "memory bandwidth quickly becomes the limiting factor" ([source](https://news.ycombinator.com/item?id=49157847))

## Alternatives hackers recommend

- [Apple M5](https://hackerapproved.com/p/apple-m5-chip.md): 48 recommend
- [AMD Ryzen](https://hackerapproved.com/p/amd-ryzen.md): 25 recommend
- [Intel N150](https://hackerapproved.com/p/intel-n150.md): 21 recommend
- [Apple M4](https://hackerapproved.com/p/apple-m4-chip.md): 21 recommend
- [Intel N100](https://hackerapproved.com/p/intel-n100.md): 18 recommend

## Where to get it

- [See price on Amazon](https://www.amazon.com/s?k=AMD+Ryzen+AI+Max%2B+395+Strix+Halo+mini+PC&tag=hackerapprove-20) (affiliate link)
- [Official site](https://www.amd.com/en.html)

## Cite as

Hacker Approved, "AMD Ryzen AI Max+ 395", https://hackerapproved.com/p/amd-ryzen-ai-max-plus-395/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).