2GB max RAM · 2 bays · 1GbE
AINAS Score 16
£169.15£238
Amazon UK · Verified

Traditional NASWD
Why Traditional NAS: Intel Pentium N3710 quad-core with no NPU and no GPU expansion, 16GB DDR3L maximum, and My Cloud OS 5 with no Docker support; comparison only.
Updated
Can the WD My Cloud PR4100 run local AI?
No. The WD My Cloud PR4100 is a storage-first Traditional NAS: Intel Pentium N3710 and at most 16GB of memory leave no room for a local language model.
It is here as a reference point against systems that can run AI.
The My Cloud PR4100 suits a buyer who wants a plain four-bay file server for a home or small office and has no plans for local AI. It has four drive bays, an Intel Pentium N3710 with 4 cores, and 4GB of DDR3L memory that can be expanded to 16GB through SO-DIMM slots. Two 1GbE ports handle file sharing. WD states built-in hardware transcoding for Plex Media Server, and that claim is vendor confirmed. The warranty is 2 years. It runs My Cloud OS 5, and the AINAS Score is 16, which reflects that it is a traditional NAS kept on the site for comparison only.
The main thing that rules the PR4100 out is that it cannot run AI software at all. My Cloud OS 5 has no documented Docker support, there is no NPU, and there is no PCIe slot or external GPU path (vendor confirmed). Memory could hold a small model, but nothing on the system can run it. Anyone who wants a local LLM, private RAG or AI search should look for a NAS with container support, at least 32GB of memory, an NPU or a GPU path such as a PCIe slot or OCuLink, and 2.5GbE or faster networking. Memory also tops out at 16GB, so even with software support the model range would stay small (Estimated).
No verified price
We have not verified a listing for the WD My Cloud PR4100 at a retailer we track. The specifications below stand on their own; a price appears here as soon as a listing is confirmed.
Product 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 | No | Unverified | No container support documented for My Cloud OS 5 and very limited memory. |
| Ollama | No | Unverified | No Docker or app path for Ollama documented. |
| Open WebUI | No | Unverified | No Docker or app path documented. |
| RAG | No | Unverified | Not supported. |
| AI search | No | Unverified | No AI search feature documented. |
| GPU expansion | No | Vendor confirmed | No PCIe slot or external GPU path. |
| Hardware acceleration | Yes | Vendor confirmed | WD states built-in hardware transcoding for Plex Media Server. |
| AI photos | No | Unverified | My Cloud OS 5 mentions improved photo/video viewing; no AI recognition documented. |
| AI video | No | Unverified | No AI video feature documented. |
| Surveillance AI | No | Unverified | No surveillance app documented on fetched pages. |
Calculated from the My Cloud PR4100'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 My Cloud PR4100's use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The PR4100 is a poor choice for running a local LLM. The local LLM capability is marked no and Unverified, with no container support documented for My Cloud OS 5 and very limited memory.
Private RAG is a poor fit here. RAG is listed as not supported and Unverified, and no Docker or app path is documented for Ollama or Open WebUI, so there is nothing to build a pipeline on.
AI search is a poor fit on the PR4100. No AI search feature is documented, and that row is marked no and Unverified, so it should not be counted on.
AI photo tagging is a poor fit. My Cloud OS 5 mentions improved photo and video viewing, but no AI recognition is documented, and the capability is marked no and Unverified.
Surveillance is a poor fit for the PR4100. No surveillance app is documented on the pages checked, and the capability is marked no and Unverified, so camera recording and detection are unconfirmed.
Home lab use is a poor fit. The PR4100 has no Docker support documented and no virtual machine support, and its Pentium N3710 with 2 x 1GbE ports leaves little room for experiments.
Business use is rated poor on the AINAS workload list. It has four bays and a 2 year warranty, but no Docker, no NPU and no GPU path, and new units are no longer listed by WD.
Creative AI is a poor fit. GPU expansion is not supported (vendor confirmed, no PCIe slot or external GPU path), and local LLM support is unconfirmed. The only acceleration WD states is Plex transcoding.
Out of the box the My Cloud PR4100 has 4GB; the official ceiling is 16GB of DDR3L SO-DIMM. A memory upgrade is the cheapest way to make it more useful for AI.
Working backwards from 16GB: reserve about 8GB for the operating system and storage services, and at roughly 0.6GB per billion parameters at 4-bit, the largest model that fits is around 13B (Estimated).
There is no supported way to add a GPU to the My Cloud PR4100: no PCIe slot and no OCuLink. Inference runs on the CPU, so keep to small models.
Not well. A useful private RAG setup needs more memory or compute than the My Cloud PR4100 offers.
Built-in networking: 2 x 1GbE. Adequate for a household, slower for bulk indexing.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| 0TB (diskless) | 4GB | None | WDBNFA0000NBK. |
| 8TB | 4GB | None | WDBNFA0080KBK. |
| 16TB | 4GB | None | WDBNFA0160KBK. |
| 24TB | 4GB | None | WDBNFA0240KBK. |
| 32TB | 4GB | None | WDBNFA0320KBK. |
| 40TB | 4GB | None | WDBNFA0400KBK. |
The closest WD model by bay count and AINAS Score. Only the specifications that differ are listed.
Its strongest AINAS profiles are Business at 20 and Home Lab at 19, which points to small offices that need shared storage with private AI tools and room to grow.
Look elsewhere if you are anyone who wants to run models larger than about 8B parameters, or buyers who expect to add a graphics card later.
Other Traditional NAS systems closest in overall score.
No. Qwen3 8B is the largest model listed on this page that would fit in the memory at Q4 (Estimated), but nothing on the PR4100 can run it. My Cloud OS 5 has no documented Docker support, and local LLM support is marked no and Unverified. Memory would hold the model, yet no software path exists. Treat any model fit on this page as a calculation only.
Yes. It ships with 4GB of DDR3L, supplied as 2GB plus 2GB, and WD documentation says it is upgradable to 16GB through two SO-DIMM slots. The memory is not ECC. Even at 16GB, the system has no documented way to run local AI software, so extra memory helps general NAS tasks only.
No. The PR4100 uses non-ECC DDR3L SO-DIMM memory. It has two SO-DIMM slots and a documented maximum of 16GB. Buyers who need error-correcting memory for data integrity should look for a NAS that lists ECC support in its specification.
No. WD confirms there is no PCIe slot and no external GPU path. There is also no OCuLink, USB4 or Thunderbolt port. The only acceleration WD states is built-in hardware transcoding for Plex Media Server, which is vendor confirmed and is not a route to GPU compute.
No. Docker is not documented for My Cloud OS 5. WD lists third-party apps and Plex, but no container support, and virtual machines are not supported either. That rules out installing Ollama or Open WebUI on the system itself, as no Docker or app path is documented.
Six. WD offers the PR4100 diskless or with drives fitted at 8TB, 16TB, 24TB, 32TB and 40TB. Every configuration has the same 4GB of DDR3L memory, expandable to 16GB, the same 4-core Intel Pentium N3710 and two 1GbE ports, so the choice changes capacity only. None of them adds container support, so the AI picture stays the same whichever capacity is chosen.
No. Ollama support is not yet verified. See the capability table above for the source.
No. Open WebUI support is not yet verified. See the capability table above for the source.
No. AI photos support is not yet verified. See the capability table above for the source.
No. Surveillance AI support is not yet verified. See the capability table above for the source.
Yes. Hardware acceleration support is documented by the manufacturer. See the capability table above for the source.
It has 2 x 1GbE built in. The fastest port is 1GbE.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.