192GB max RAM · 12 bays · 2 x 10GbE
AINAS Score 64
No verified price

AI ReadyQNAP
Why AI Ready: Xeon E-2378 or E-2336 with PCIe Gen 4 x8 expansion and a 128GB memory ceiling meets the Xeon ready rule; the best listed GPU is the 4GB RTX A400 and QNAP documents no GPU inference on this model.
Updated
Can the QNAP TS-h2287XU-RP run local AI?
Partly. The QNAP TS-h2287XU-RP runs local models on Intel Xeon E-2378 (E-2336 on E2336-32G) alone, and its 128GB ceiling allows 70B class models at higher-quality quantization, or large mixture-of-experts models at 4-bit (Estimated), but replies will be slow without graphics acceleration.
A graphics card can be added through its PCIe x8 slot, which is the biggest single speed upgrade for local models.
Local LLM support is documented by the manufacturer.
The TS-h2287XU-RP suits businesses and serious home lab users who want a 22-bay rackmount NAS that can also host local AI in containers. It ships with 64GB of DDR4 ECC UDIMM memory, expandable to 128GB, and an 8-core Intel Xeon E-2378. Docker and virtual machines are both supported, and QNAP documents Ollama and Open WebUI through Container Station. Network ports are 2 x 10GbE and 2 x 2.5GbE, and a 3U chassis holds 16 3.5-inch and 6 2.5-inch drives. Memory is not the constraint: Mistral Large 2 123B fits at Q4 (Estimated), and the AINAS Score is 63. Warranty is 5 years.
The main limit is GPU inference. The three PCIe Gen 4 slots (one x8, two x4) take cards up to 250mm long, but the largest listed GPU is the 4GB RTX A400, and QNAP documents no GPU inference on this model. Large models will therefore run on the CPU, and QNAP rates local LLM support as limited. Buyers who want GPU-accelerated generation should look for a system with a 16GB or larger VRAM GPU path, or an OCuLink or Thunderbolt link to an external GPU. This model has neither. The E2336-32G configuration also drops to a 6-core Xeon E-2336 and 32GB, which leaves less headroom for the CPU work these models depend on.
Amazon US
TS-h2287XU-RP-E2336-32G
Ships from the US; delivery and import charges shown at checkout
$6,209.00
Verified
In stock
Buy on AmazonAmazon US
TS-h2287XU-RP-E2378-64G
Ships from the US; delivery and import charges shown at checkout
$9,999.00
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 | QNAP Ollama guide applies (x86_64, 16GB or more for 7B models, GPU optional and must be in Container Station mode). 128GB allows CPU inference; the largest listed GPU is the 4GB RTX A400. |
| Ollama | Yes | Vendor confirmed | QNAP Container Station Ollama tutorial for QTS and QuTS hero on x86_64. |
| Open WebUI | Yes | Vendor confirmed | QNAP Open WebUI tutorial, QTS 5.2 or QuTS hero h5.2 or later. |
| RAG | Limited | Unverified | Qsirch lists on-premises RAG with no model named. Not confirmed for this model. |
| AI search | Yes | Vendor confirmed | Qsirch AI Semantic Search: 64-bit x86, 5.0.1 or later, 8GB RAM. Qsirch says the AI Semantic Index and AI Mode need an x74 series NAS or later; not confirmed for this model, and AI Mode also needs a 16GB or larger VRAM GPU. |
| GPU expansion | Limited | Vendor confirmed | PCIe Gen 4 x8 slot, 250 mm card limit. GPU list tops out at RTX A400 4GB and Quadro P1000 4G. |
| Hardware acceleration | Limited | Vendor confirmed | No integrated graphics; spec page says graphics optional via a PCIe adapter. Coral Edge TPU on the AI accelerator list; RTX A400 lists Hardware Transcoding. |
| AI photos | Yes | Vendor confirmed | QuMagie listed on QTS and QuTS hero. |
| AI video | Yes | Vendor confirmed | QVR Face listed under QTS only. |
| Surveillance AI | Yes | Vendor confirmed | QVR Surveillance 2 free channels plus 6 on registration; QVR Pro 8 free channels (QTS only). |
Calculated from the TS-h2287XU-RP'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 | Q8 | Yes | 19GB | Not measured | Estimated |
| Qwen3 8BQwen, 8.19B | Q8 | Yes | 19GB | Not measured | Estimated |
| Gemma 3 12BGemma, 11.8B | Q8 | Yes | 24GB | Not measured | Estimated |
| Qwen3 14BQwen, 14.8B | Q8 | Yes | 28GB | Not measured | Estimated |
| Phi-4 14BPhi, 14.7B | Q8 | Yes | 28GB | Not measured | Estimated |
| Mistral Small 3 24BMistral, 23.6B | Q8 | Yes | 38GB | Not measured | Estimated |
| Gemma 3 27BGemma, 27B | Q8 | Yes | 44GB | Not measured | Estimated |
| Qwen3 30B A3BQwen, 30.5B | Q8 | Yes | 47GB | Not measured | Estimated |
| Qwen3 32BQwen, 32.8B | Q8 | Yes | 50GB | Not measured | Estimated |
| DeepSeek-R1 32B (Qwen distill)DeepSeek, 32.8B | Q8 | Yes | 50GB | Not measured | Estimated |
| Llama 3.3 70BLlama, 70.6B | Q8 | Yes | 98GB | Not measured | Estimated |
| Llama 4 Scout 17Bx16ELlama, 109B | Q4 | Yes | 89GB | Not measured | Estimated |
| Qwen2.5 72BQwen, 72.7B | Q8 | Yes | 101GB | Not measured | Estimated |
| Mistral Large 2 123BMistral, 123B | Q5 | Yes | 112GB | Not measured | Estimated |
One verdict per workload, from the TS-h2287XU-RP's use-case scores and capability rows. Unknown means we cannot confirm it, not that it probably works.
The TS-h2287XU-RP is a good fit for local LLM work. QNAP's Ollama guide applies and is vendor confirmed, and 128GB allows CPU inference, though the local LLM row is marked limited because the largest listed GPU is 4GB.
Private RAG is an acceptable fit, not a strong one. The RAG capability is unverified and limited, since Qsirch lists on-premises RAG without naming a model, while Ollama and Open WebUI are vendor confirmed.
AI search is a good fit. Qsirch AI Semantic Search is vendor confirmed, but the AI Semantic Index is not confirmed for this model, and AI Mode needs a GPU with 16GB or more of VRAM.
AI photo management is a good fit. QuMagie is vendor confirmed on both QTS and QuTS hero, so the photo features are not tied to one of the two operating systems.
Surveillance is a good fit. QVR Surveillance offers 2 free channels plus 6 on registration, and QVR Pro adds 8 free channels, with QVR Pro on QTS only. Both are vendor confirmed.
Home lab use is a good fit. Docker and virtual machines are both supported, and 22 bays, 64GB of ECC memory and 2 x 10GbE give plenty of room for storage and test services.
Business use is a good fit. ECC memory, two 550W power supplies, a 3U rackmount chassis and a 5 year warranty suit an office or server room.
Creative AI is rated a good fit, with a caveat. GPU expansion and hardware acceleration are both limited, as the listed GPUs top out at the 4GB RTX A400 and the Quadro P1000 4G.
Out of the box the TS-h2287XU-RP has 64GB; the official ceiling is 128GB of DDR4 ECC UDIMM with ECC support. A memory upgrade is the cheapest way to make it more useful for AI.
Working backwards from 128GB: 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 200B (Estimated).
ECC memory support is a plus for always-on systems that also hold important files.
It can. A 3x PCIe Gen 4 (2 x4, 1 x8); cards up to 250 x 111.15 mm (slot 1 up to 23.76 mm thick, slots 2 and 3 up to 18.76 mm) is the route, and a card with 16GB or more of its own memory changes what the TS-h2287XU-RP can run. Measure the space and check the 550W power budget before ordering.
For a personal library, yes. A compact embedding model and a small chat model will run, but the TS-h2287XU-RP is better as the document store behind a RAG system whose model runs on stronger hardware.
The manufacturer also provides AI search across your files (documented by the manufacturer), which covers simple document questions without setting up your own stack.
Built-in networking: 2 x 10GbE + 2 x 2.5GbE. Plenty for a shared knowledge base with several users.
| Configuration | RAM | SSD | Notes |
|---|---|---|---|
| TS-h2287XU-RP-E2336-32G | 32GB | None | Xeon E-2336 6C/12T, 32GB ECC. |
| TS-h2287XU-RP-E2378-64G | 64GB | None | Xeon E-2378 8C/16T, 64GB ECC. |
The closest QNAP model by bay count and AINAS Score. Only the specifications that differ are listed.
Its strongest AINAS profiles are Home Lab at 72 and Business at 68, which points to home lab owners who want one box for storage, containers, media and a few AI services.
Other AI Ready systems closest in price.
Buying guides that list it
Workloads it suits
Yes, at Q4 only (Estimated). It needs 73GB at Q4, which fits within the 128GB memory ceiling, but 130GB at Q8 does not fit (Estimated). It would run on the CPU, because QNAP documents no GPU inference here. It is a memory calculation, not a benchmark.
128GB at most. The system ships with 64GB of DDR4 ECC UDIMM, and the memory can be upgraded. The E2336-32G configuration ships with 32GB instead. ECC is supported, which suits a NAS holding business data.
Yes, but the choice is limited. There are three PCIe Gen 4 slots (two x4, one x8) taking cards up to 250mm long. QNAP's GPU list tops out at the 4GB RTX A400 and Quadro P1000 4G, and there is no integrated graphics.
No. It has no onboard M.2 slots. NVMe storage is optional through a QM2 PCIe adapter, which takes up one of the three PCIe slots. The bays hold 16 3.5-inch and 6 2.5-inch drives.
Two. The TS-h2287XU-RP-E2378-64G has an 8-core, 16-thread Xeon E-2378 and 64GB of DDR4 ECC UDIMM, and the TS-h2287XU-RP-E2336-32G has a 6-core, 12-thread Xeon E-2336 and 32GB. Both share the 22 bays, the three PCIe Gen 4 slots, 2 x 10GbE and 2 x 2.5GbE, and both take up to 128GB, so both reach the same Estimated fits, such as Llama 3.3 70B at Q4 and Q8, once memory is raised to 128GB. Because large models run on the CPU here, the 8-core version suits local AI better.
Yes. Docker is supported, and so are virtual machines. QNAP's Container Station guides for Ollama and Open WebUI are vendor confirmed, and the system runs QTS 6.1.0 or QuTS hero h6.1.0.
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 10GbE + 2 x 2.5GbE built in, and the 3x PCIe Gen 4 (2 x4, 1 x8); cards up to 250 x 111.15 mm (slot 1 up to 23.76 mm thick, slots 2 and 3 up to 18.76 mm) can take a faster network card. The fastest port is 2 x 10GbE.
Every field records its source. Specifications were last checked on . Spot an error? Send a correction with a manufacturer link.