Minisforum / Sweet spot tier
Minisforum MS-01
Expandable mini-PC in the same memory bracket as the mid tier of this index. Its distinguishing trait against closed unified-memory systems is serviceability: RAM and storage are user-upgradeable, which extends the useful life of the machine as local models grow.
- Cores
- 14 cores and 20 threads
- Graphics
- 96EU Torch core
- Processor
- Intel Core i9-13900H
- Usb Ports
- Two USB4, two USB TYPE-A
The decision
Where it wins
- Dual 10GbE networking ports
- PCIe x8 slot for hardware expansion
- Robust cooling for a 1L chassis
Where it doesn't
- Limited to half-height single-slot GPUs
- External power brick is bulky
Specifications
Structured product data, grouped by system.
Core hardware03
| Processor | Intel Core i9-13900H |
|---|---|
| Network Chip | Intel X710 for 10GbE, i226 for 2.5G |
| Max Clock Speed | 5.4GHz |
AI & acceleration01
| Cores | 14 cores and 20 threads |
|---|
Power & thermal02
| Pl1 Power Consumption | 60W |
|---|---|
| Pl2 Power Consumption | 80W |
Physical & connectivity03
| Usb Ports | Two USB4, two USB TYPE-A |
|---|---|
| Form factor | mini-pc |
| Network Ports | Dual 10GbE, two 2.5G (one with vPro) |
Other specifications03
| Graphics | 96EU Torch core |
|---|---|
| Expandable | Yes |
| Heat Dissipation | Metal body construction, large heat dissipation module, independent heat sinks for 10GbE interfaces |
Read the record
- Source on record
- minisforum.com
An enterprise-grade workstation condensed into a 1L chassis. The inclusion of a PCIe x8 slot makes it uniquely adaptable for local LLM inference when paired with a low-profile GPU. Thermal limits require careful component selection, but the dual 10GbE networking provides excellent data throughput.
The MS-01 condenses an enterprise-grade workstation into a 1-litre chassis, and its distinguishing trait against closed unified-memory boxes is serviceability: RAM and storage are user-upgradeable, so the machine's useful life extends as local models grow.
What makes it adaptable for local AI
A PCIe x8 slot — rare at this size — means you can add a low-profile GPU for CUDA inference rather than being stuck with integrated graphics. Dual 10GbE networking gives it the data throughput for a homelab node or an edge box feeding other machines.
The trade-offs
- GPU choice is constrained to half-height, single-slot cards — real acceleration, but not a 5070 Ti.
- The external power brick is bulky, which undercuts some of the small-chassis appeal.
- Thermals in 1 litre demand careful component selection.
Who should buy
Edge AI nodes, dense homelab virtualisation, and anyone who wants a small machine they can upgrade rather than replace. If you want maximum model capacity in one box, a 128GB unified-memory mini-PC runs larger models — just not faster, and not upgradeably.
Frequently Asked Questions
Who is Minisforum MS-01 for?↓
What are the drawbacks of Minisforum MS-01?↓
What does Minisforum MS-01 do well?↓
How much does Minisforum MS-01 cost?↓
What is the Fathom Layer design score for Minisforum MS-01?↓
Alternatives to consider
See all alternatives →- Another option Mac Mini M4 Pro (64GB)
64GB unified memory accessible by GPU
64 GBDesign score 9.5/10 - Another option Nvidia GeForce RTX 3090 (Used)
24GB of GDDR6X memory for handling large datasets and complex scenes
Explore its evidenceDesign score 8.5/10 - Another option GMKtec EVO-X2 (128GB)
128GB massive memory pool
128GB LPDDR5X 8000MHzDesign score 8.0/10
Where to buy
Purchase links may earn Fathom Layer a commission, at no extra cost to you. As an Amazon Associate we earn from qualifying purchases.
Further reading
Featured badge
Building this product? Add the badge to your site to show it’s in the index.
<a href="https://fathomlayer.com/compute/local-ai-workstations/minisforum-ms-01" target="_blank" rel="noopener noreferrer"><img src="https://fathomlayer.com/fathom-badge.svg" alt="Featured on Fathom Layer" width="250" height="54" /></a>