MINISFORUM showcased its upcoming flagship AI NAS, the N5 MAX, at the OpenClaw developers' gathering "How to Use OpenClaw – OpenClaw Developers Shenzhen Gathering" held in Shenzhen on March 14, 2026, demonstrating a range of deployment scenarios on local LLMs[1]. Roughly 800 developers, AI enthusiasts and industry partners attended, and the event positioned the OpenClaw ecosystem on edge AI not as an abstract concept but alongside concrete use cases[1].

800 attendees, with AMD and MINISFORUM on stage

The gathering took place on a Saturday in Shenzhen and centered on the question of "how to use" OpenClaw. Developers, AI enthusiasts and industry partners gathered to discuss how OpenClaw is spreading on edge AI[1].

The opening speaker, AMD Greater China Region AI Marketing Director Zhongyang Zan, demonstrated three OpenClaw AI agents running simultaneously on a single laptop and showed how edge AI could expand the workflows of small operators — what he called "One-Person-Company" — through enterprise collaboration platforms[1]. He also pointed to the MINISFORUM AI NAS N5 MAX, equipped with the AMD Ryzen AI Max+ 395, as a system that can run the same kind of scenario more safely on the edge and without token fees[1].

Semantic photo search and game generation on a local LLM

In the second half of the event, MINISFORUM Product Director Jiang Xue took the stage and ran OpenClaw on a local LLM via the N5 MAX with no internet connection[1].

Two demonstrations stood out. The first was an AI-powered semantic photo search: a natural-language query asked for "photos taken on a specific date" and the system pulled the matching files from the local album[1]. The second was a "Snake Game" generated on the spot by OpenClaw, which was then playable directly on the demo screen[1]. Both demos ran entirely on the N5 MAX without making external API calls, which structurally eliminates inference costs and the risk of data leaving the device — the central message of the demonstration[1].

MINISFORUM cited three pillars supporting this kind of deployment[1]:

  • Zero token cost for AI model usage thanks to local execution
  • Privacy protection, with sensitive data processed on the device
  • Stronger data security, eliminating the need to upload personal information to external servers

Inside the N5 MAX, and its role as an "edge AI NAS"

On the hardware side, the N5 MAX uses AMD's flagship APU, the Ryzen AI Max+ 395 (codename Strix Halo). According to Tom's Hardware, the APU houses 16 Zen 5 cores clocked up to 5.1 GHz, an integrated Radeon 8060S GPU with 40 CUs, an XDNA 2 NPU and 64 MB of L3 cache[2]. System memory is configurable from 32 GB to 128 GB, and given that NAS workloads benefit from loading LLMs into the largest possible memory pool, the publication suggests MINISFORUM is likely to ship the N5 MAX in the 64 GB to 128 GB range[2].

The chassis appears to be the same one used by the company's N5 and N5 AI Pro, which Tom's Hardware expects will mean five 3.5/2.5-inch HDD bays plus three M.2 slots (two of them U.2-capable), with each HDD bay accepting drives up to 30 TB[2]. On the software side, MINISFORUM's in-house "MinisCloud OS" coordinates the AI NAS functionality, and OpenClaw is launched through pre-integrated one-click deployment tools[3]. OpenClaw itself is not a standalone LLM but an AI agent framework that routes messages to an LLM and lets the LLM choose which tools to invoke, allowing photo search, video editing, email automation and social-media posting to share a single foundation[2][3].

One caveat: Tom's Hardware notes that the rapid uptake of OpenClaw has surfaced new security management challenges, citing cases of malicious extensions appearing on the third-party hub "ClawHub"[2]. Local execution itself is helpful for privacy, but vetting the extensions and tool chains an agent can call remains an operational responsibility for the user. Pricing and regional release timing have not yet been formally announced by MINISFORUM and are awaiting further updates[2].

Summary

The N5 MAX demo at this event quietly pulls generative AI back from the "use it in the cloud" assumption toward "complete it on the NAS in your own room." For independent developers and small offices that have been weighing token costs and network-side privacy concerns, having a Strix Halo-class APU inside a NAS chassis is a meaningful option. That said, pricing, release date and Japan availability are still unannounced, so anyone planning a purchase should wait for MINISFORUM's official announcement and also confirm the security policy around third-party extensions.

出典:https://www.minisforum.com/blogs/news/minisforums-ai-nas-n5-max-showcased-diverse-scenario-deployment-in-local-llms-at-openclaw-developers-gathering

出典:https://www.tomshardware.com/pc-components/nas/minisforums-new-flagship-nas-comes-with-openclaw-pre-installed-strix-halo-powered-n5-max-can-run-a-local-ai-llm

出典:https://www.prnewswire.com/news-releases/minisforums-n5-max-marks-a-breakthrough-in-nas-industry-with-built-in-openclaw-in-local-computing-302710891.html