32GB max RAM · 8 bays · 1GbE
AINAS Score 32
No verified price

AI CapableASUSTOR
Why AI Capable: Intel Celeron N5105 (4C/4T) with Intel UHD Graphics, no NPU, no PCIe or GPU expansion and a 16GB memory ceiling; small CPU-only models via ASUSTOR's Ollama app at best.
Updated
Can the ASUSTOR Nimbustor 4 Gen2 (AS5404T) run local AI?
The ASUSTOR Nimbustor 4 Gen2 (AS5404T) can run small local models, with limits. Its Intel Celeron N5105 and 16GB memory ceiling suit 7B to 8B class models (Estimated), so treat it as a storage-first NAS that can host light AI tools.
There is no way to add a discrete GPU, so all inference runs on the CPU and basic integrated graphics.
Local LLM support is documented by the manufacturer.
The Nimbustor 4 Gen2 suits a buyer who wants a general four-bay NAS that can also run light AI tasks on the side. It has four bays, four M.2 slots and two 2.5GbE ports, with an Intel Celeron N5105 and ADM 5. Docker and virtual machines are both supported. ASUSTOR's Ollama and Open WebUI package is vendor confirmed, and face grouping in Photo Gallery runs on the NAS. Surveillance Center support is vendor confirmed, with a licence table that goes up to 44 IP cameras. It ships with 4GB and takes up to 16GB of DDR4-2933 SO-DIMM. Qwen3 8B fits at Q4 (Estimated). It carries a 3-year warranty and draws power from a 90W supply. The AINAS Score is 32.
The limit that rules it out for some buyers is memory and the lack of a GPU path. Memory tops out at 16GB, and there is no PCIe slot, no Thunderbolt, no USB4 and no OCuLink, so a graphics card cannot be added. Anyone who wants to run 14B models comfortably, or models of 24B and above, will be stuck. Mistral Small 3 24B does not fit at Q4 (Estimated), while Gemma 3 12B and Qwen3 14B are marginal at Q4 (Estimated). Inference runs on four CPU cores with no accelerator, and ASUSTOR recommends a separate computer for RAG inference. Buyers with those aims should look for a much higher memory ceiling, a PCIe slot or OCuLink for a GPU, or an NPU with a large amount of unified memory.
Amazon UK
4GB
£476£529
Verified
In stock
Buy on AmazonProduct prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon.co.uk or Amazon.com, as applicable, at the time of purchase will apply to the purchase of this product.
Each capability shows how it was confirmed: Verified by AINAS, Vendor confirmed, Community confirmed or Unverified.
| Capability | Supported | Evidence | Note |
|---|---|---|---|
| Local LLM | Limited | Vendor confirmed | The Nimbustor 4 Gen2 (AS5404T) falls under App Central's AS54 entry for the Ollama and Open WebUI app, so models run on the four N5105 cores with no accelerator. Out of the box it has 4GB, too little for most models, and ASUSTOR's 16GB ceiling sets the upper limit. The app ships without a model. |
| Ollama | Yes | Vendor confirmed | ASUSTOR's Ollama and Open WebUI package names AS54 as supported, and it installs Ollama as a Docker container on this four-bay unit. |
| Open WebUI | Yes | Vendor confirmed | On the AS5404T, Open WebUI comes inside the same App Central package as Ollama rather than as its own app. |
| RAG | Limited | Vendor confirmed | ASUSTOR positions its NAS as a home for RAG knowledge bases, with Chroma, pgvector or Weaviate in Docker, and Open WebUI can chat over uploaded documents. For the inference itself ASUSTOR recommends a separate computer. |
| AI search | No | Unverified | Nothing in the ADM 5.1 material for this model describes AI or semantic search across stored files. |
| GPU expansion | No | Vendor confirmed | The AS5404T has no PCIe slot on its spec page, unlike the Lockerstor 4 Gen2's 10GbE card slot, and ASUSTOR says none of its NAS carry a discrete GPU. |
| Hardware acceleration | Limited | Vendor confirmed | ASUSTOR uses the N5105's Intel UHD Graphics for 4K H.265 decoding and transcoding. The chip has no NPU, and AI inference on the iGPU is not documented. |
| AI photos | Yes | Vendor confirmed | AS54 is on the support list for Photo Gallery 4.2 (ADM 5.1.1 or later), whose face grouping runs on the NAS rather than in the cloud. |
| AI video | No | Unverified | No AI video analysis appears in ASUSTOR's ADM documentation for the Nimbustor 4 Gen2. |
| Surveillance AI | Yes | Vendor confirmed | AS54 is listed by both Surveillance Center 2.8.4 (app 1663) and Surveillance Center 3 (app 1661); the spec page licence table goes up to 44 IP cameras. |
Calculated from the Nimbustor 4 Gen2 (AS5404T)'s memory and each model's published size at each quantization. Rows marked Estimated are calculations, not benchmarks. Check any model.
| Model | Best quantization | Runs? | Memory needed | Expected performance | Verification |
|---|---|---|---|---|---|
| Llama 3.2 3BLlama, 3.21B | Q8 | Yes | 13GB | Not measured | Estimated |
| Llama 3.1 8BLlama, 8.03B | Q5 | Yes | 15GB | Not measured | Estimated |
| Qwen3 8BQwen, 8.19B | Q4 | Yes | 15GB | Not measured | Estimated |
| Gemma 3 12BGemma, 11.8B | Q4 | Marginal | 15GB | Not measured | Estimated |
| Qwen3 14BQwen, 14.8B | Q4 | Marginal | 16GB | Not measured | Estimated |
| Phi-4 14BPhi, 14.7B | Q4 | Marginal | 16GB | Not measured | Estimated |
| Mistral Small 3 24BMistral, 23.6B | None fits | No | 25GB at Q4 | Not measured | Estimated |
| Gemma 3 27BGemma, 27B | None fits | No | 29GB at Q4 | Not measured | Estimated |
| Qwen3 30B A3BQwen, 30.5B | None fits | No | 31GB at Q4 | Not measured | Estimated |
| Qwen3 32BQwen, 32.8B | None fits | No | 32GB at Q4 | Not measured | Estimated |
| DeepSeek-R1 32B (Qwen distill)DeepSeek, 32.8B | None fits | No | 32GB at Q4 | Not measured | Estimated |
| Llama 3.3 70BLlama, 70.6B | None fits | No | 60GB at Q4 | Not measured | Estimated |
| Llama 4 Scout 17Bx16ELlama, 109B | None fits | No | 89GB at Q4 | Not measured | Estimated |
| Qwen2.5 72BQwen, 72.7B | None fits | No | 65GB at Q4 | Not measured | Estimated |
| Mistral Large 2 123BMistral, 123B | None fits | No | 96GB at Q4 | Not measured | Estimated |
One verdict per workload, from the Nimbustor 4 Gen2 (AS5404T)'s use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The Nimbustor 4 Gen2 is a poor choice for local LLMs. Support is limited and vendor confirmed: models run on the four N5105 cores with no accelerator, 4GB is included, and the 16GB ceiling caps model size.
Private RAG is a poor fit here. ASUSTOR says Chroma, pgvector or Weaviate can run in Docker and Open WebUI can chat over documents (limited support), but it recommends a separate computer for the inference itself.
AI search is poor on this system, and we cannot confirm it exists at all. The capability row is Unverified, and the ADM 5.1 material describes no AI or semantic search across stored files.
AI photos rate poor overall, although one part works. Photo Gallery 4.2 lists this model, and its face grouping runs on the NAS rather than in the cloud (vendor confirmed), but the weak CPU and memory limit wider AI use.
Surveillance rates poor in our scoring, even with basic recording support. Surveillance Center 2.8.4 and Surveillance Center 3 both list this model (vendor confirmed), with a licence table up to 44 IP cameras, but no NPU or GPU is available for analysis.
Home lab use is poor for this system. Docker and virtual machines are supported, but memory stops at 16GB, there is no ECC, and the two 2.5GbE ports are the only networking.
Business use is poor on the Nimbustor 4 Gen2. It has no ECC memory and a 16GB ceiling, and AI features are limited to a quad-core Celeron, though the 3-year warranty and two 2.5GbE ports help ordinary file serving.
Creative AI is poor here. There is no PCIe slot and ASUSTOR says its NAS carry no discrete GPU, and the Intel UHD Graphics is used for 4K H.265 transcoding, not documented AI inference.
The Nimbustor 4 Gen2 (AS5404T) ships with 4GB of DDR4-2933 SO-DIMM and supports up to 16GB. Upgrading to the maximum lifts the model ceiling from only very small models of around 1B to 3B parameters to 7B to 8B class models (both Estimated).
As a rule of thumb, a 4-bit model needs about 0.6GB per billion parameters plus room for context. After leaving around 8GB for the NAS operating system, 16GB leaves space for a model of roughly 13B parameters at most (Estimated).
No. The Nimbustor 4 Gen2 (AS5404T) has no PCIe slot or OCuLink port for a graphics card. Inference runs on the CPU, with Intel UHD Graphics handling media tasks, so keep to small models.
With limits. The Nimbustor 4 Gen2 (AS5404T) can host a small RAG stack such as AnythingLLM or Open WebUI with a compact embedding model, but memory and compute restrict how large the answering model can be. It suits a personal document library more than a team knowledge base.
Networking is 2 x 2.5GbE. Fine for one or two users; indexing a large archive over the network will be slower.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| 4GB | 4GB | None | Diskless; 1x 4GB DDR4 pre-installed, 2 SO-DIMM slots. |
The closest ASUSTOR model by bay count and AINAS Score. Only the specifications that differ are listed.
The Nimbustor 4 Gen2 (AS5404T) scores best for Home Lab (36/100) and Business (36/100). It suits home lab owners who want one box for storage, containers, media and a few AI services.
It is a weaker fit for anyone who wants to run models larger than about 8B parameters and for buyers who expect to add a graphics card later.
Other AI Capable systems closest in price.
Yes, at Q4 only. Qwen3 8B needs 5.2GB at Q4 and fits (Estimated), but at Q8 it needs 8.9GB and is marginal (Estimated), with little headroom. It runs on the four N5105 cores with no accelerator through ASUSTOR's Ollama and Open WebUI package, which ships without a model. ASUSTOR lists local LLM support as limited. Fitting more than the 4GB included means adding memory.
Yes. It uses DDR4-2933 SO-DIMM modules and takes up to 16GB, as 2x 8GB. The sold configuration has one 4GB module pre-installed and two SO-DIMM slots in total. The memory is not ECC. ASUSTOR's compatibility list covers modules from 4GB to 16GB for this model, and 32GB is listed as unofficial only.
No. The spec page lists no PCIe expansion slot, and the system has no Thunderbolt, USB4 or OCuLink port that could carry an external GPU. ASUSTOR also says none of its NAS carry a discrete GPU. The Intel UHD Graphics handles 4K H.265 decoding and transcoding, and AI inference on it is not documented.
No. The Intel Celeron N5105 has four cores and no NPU, so models run on the CPU with no accelerator. Its Intel UHD Graphics handles 4K H.265 decoding and transcoding for media, but ASUSTOR does not document AI inference on the iGPU. The Ollama and Open WebUI package is vendor confirmed and comes without a model, so you choose one small enough for the 16GB ceiling, such as Qwen3 8B at Q4 (Estimated).
Yes. ASUSTOR's Docker Engine package lists the AS54 and needs ADM 5.1.0 or later. Its Ollama and Open WebUI package installs Ollama as a Docker container. Virtual machines are also supported through VirtualBox, which needs ADM 5.1.3 or later. ASUSTOR describes running Chroma, pgvector or Weaviate in Docker for RAG knowledge bases.
Yes. Ollama support is documented by the manufacturer. See the capability table above for the source.
Yes. Open WebUI support is documented by the manufacturer. See the capability table above for the source.
Yes. AI photos support is documented by the manufacturer. See the capability table above for the source.
Yes. Surveillance AI support is documented by the manufacturer. See the capability table above for the source.
Partly. Hardware acceleration support is documented by the manufacturer. See the capability table above for the source.
It has 2 x 2.5GbE built in. The fastest port is 2 x 2.5GbE.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.