DGX Spark and Every GB10 System Compared
Dell Pro Max with GB10 leads this page with an AIPC Score of 55.5, giving a local model about 120GB of usable AI memory at $5,903 for the 128GB / 2TB configuration. The cheapest tracked option here is Lenovo ThinkStation PGX at $5,859. 8 systems on this page, 4 with a verified price.
Updated
| # | System | Type | AI Memory | GPU/NPU | AI Score | Price | Verified | View |
|---|---|---|---|---|---|---|---|---|
| 1 |  Dell Pro Max with GB10 | Mini PC | 120GB | Blackwell GPU (GB10) | 55.5 | Price awaiting re-verification | Check Price at Dell → | |
| 2 |  Lenovo ThinkStation PGX | Mini PC | 120GB | Blackwell GPU (GB10) | 55.5 | Price awaiting re-verification | Check Price at Lenovo → | |
| 3 |  NVIDIA DGX Spark | Mini PC | 120GB | Blackwell GPU (GB10) | 55.3 | Price awaiting re-verification | Check Price at NVIDIA → | |
| 4 |  MSI EdgeXpert MS-C931 | Mini PC | 120GB | Blackwell GPU (GB10) | 55.1 | Price awaiting re-verification | Check Price at MSI → | |
| 5 |  ASUS Ascent GX10 | Mini PC | 120GB | Blackwell GPU (GB10) | 54.2 | No verified price | — | View details → |
| 6 |  HP ZGX Nano G1n | Mini PC | 120GB | Blackwell GPU (GB10) | 54.2 | No verified price | — | View details → |
| 7 |  Gigabyte AI TOP ATOM | Mini PC | 120GB | Blackwell GPU (GB10) | 54.2 | No verified price | — | View details → |
| 8 |  Acer Veriton GN100 | Mini PC | 120GB | Blackwell GPU (GB10) | 54.2 | No verified price | — | View details → |
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 at the time of purchase will apply to the purchase of this product.
One chip, several badges
DGX Spark and its partner machines all use the same NVIDIA GB10 module: a 20-core Arm processor and a Blackwell graphics processor sharing 128GB of coherent memory at 273 GB/s. The differences between them are chassis, storage options, networking and support, not performance. That is why they score within a point of each other on this site.
Why coherent memory matters
The graphics processor addresses nearly the whole pool directly, so the practical reservation is the operating system, around 8GB. A 128GB machine therefore gives a model about 120GB, which is more than any of the unified-memory alternatives at this size.
Usable AI memory decides what a machine can load at all, and memory bandwidth decides how fast it answers once loaded. Those two numbers matter more than the processor name on the box. A machine with a large unified pool and modest bandwidth will run a very large model slowly; a machine with a fast narrow pool will run a small model quickly and refuse a large one outright.
The software argument
These run the full CUDA stack. If your tooling assumes NVIDIA, which most inference and fine-tuning tooling still does, that removes an entire category of problem you would otherwise spend weekends on.