Power rating
Published detail240 W external adapter capacity; 140 W SoC TDP. Neither number is typical wall consumption. [1]
Product image
Manufacturer product image. The selected configuration may vary.
Product photo & illustrative 3D · configuration may varySource
Sourced specifications
Desktop AI systemNVIDIA DGX Spark
GB10 Grace Blackwell Superchip
A compact NVIDIA development system for exploring local models, applications and workflows.
128 GB unifiedReference memory
arm64CUDA ecosystem
Compared withASUS Ascent GX10Dell Pro Max with GB10HP ZGX Nano G1nLenovo ThinkStation PGXAcer Veriton GN100GIGABYTE AI TOP ATOMMSI EdgeXpertApple Mac Studio · M3 UltraApple Mac Studio · M5 UltraFramework Desktop · Ryzen AI Max+MINISFORUM MS-S1 MAXHP Z2 Mini G1aAMD Ryzen AI HaloNVIDIA DGX Station
BEFORE IT ARRIVES
Power, cooling & your space.
What this configuration needs on your premises.
Lab assessment · based on the evidence below
Office candidate
A compact office candidate. Its documented operating ceiling is 30°C: an unventilated cupboard or warm equipment shelf needs particular attention.
Electrical input
Published detail48 V DC, 5 A at the computer. Use the approved adapter and grounded power cord; check its AC input label for your region. [2]
Outlet & circuit
Not verifiedNo machine-specific branch-circuit requirement verified. Check the nameplate, regional cord and other loads on the same circuit.
Machine cooling
Published detailAir-cooled heatsink and two fans, documented in ChargerLAB’s disassembly of the 4 TB model. Heat pipes also serve supporting components. [3]
Operating limits
Published detail5–30°C operating ambient. Keep the system and adapter ventilated. [2]
Noise in the workspace
Not verifiedNo comparable noise measurement under an AI workload verified. Ask for dB(A), distance, workload and ambient conditions; audition it before placing it beside people.
ACTUAL CONSUMPTION
Power at the wall
PSU capacity, chip TDP and measured consumption describe different things. Only measurements with a stated configuration and test method appear here.
NVIDIA DGX Spark · regulatory model P4242
- Idle in vendor test
- 38 W
- Maximum in vendor test
- 233.2 W
NVIDIA compliance measurements under IEC/EN 62623 at 230 V, 50 Hz. These are vendor test points, not measurements of an LLM workload or of another GB10 manufacturer’s box. [2]
Room cooling
Plan for the computer and its power supply to heat the room. Check ventilation during long jobs and when several machines share a small office.
Lab planning guidance. Office suitability depends on the installed configuration, shared circuits, heat removal and acceptable noise; a workstation label alone does not settle it.
Space & installation
- Enclosure
- 150 × 150 × 50.5 mm
- Weight
- 1.2 kg
- Keep air inlets and exhausts unobstructed, with room for cables and servicing. An enclosed cupboard needs its own ventilation; the enclosure dimensions are not an airflow clearance.
Estimate electricity use & room heat
Enter average power at the wall for the hours you are modeling, or an explicit planning assumption. Supply wattage and chip TDP are not consumption measurements.
Add wall power to calculate a scenario.
Scenario only. Hours outside this period are excluded: include idle time for an always-on system. Displays, networking, UPS losses and room air conditioning are extra. Heat assumes the computer and power supply release their heat into this room; it does not size a circuit or an HVAC system. Currency selects your tariff’s unit, with no conversion.
Sources & review scope (3)
Manufacturer documentation and attributed first-hand hardware inspections, scoped to the named model or configuration. “Not verified” identifies a gap in this review, not a claim that the manufacturer has no documentation. Installation assessments are the Lab’s interpretation.
Opening model configurations…
01 / PRIVACY
Your data has a home.
Keep sensitive processing on your infrastructure. Inspect every connection in the complete stack.
02 / SECURITY
You set the boundaries.
Control access, connectivity and updates. A local machine still needs a secure deployment.
03 / TOKEN ECONOMICS
Own the capacity.
Generate tokens on your own compute. Understand the full cost of the useful work it delivers.
SHARED WORK AROUND THIS PLATFORM
Setups you can build on.
Reference-platform guidance and field records. Exact OEM configurations need their own validation.
Understand TechField recordAn AI application stack on two office Sparks
UT’s move from rented cloud GPUs to machines in its own office, and the engineering record behind it.
NVIDIA · indexed by the LabRecipeA browser chat interface, served by your Spark
NVIDIA’s Open WebUI and Ollama guide, connected to the hardware and operating questions it raises.
NVIDIA · indexed by the LabRecipeAn inference endpoint for your local applications
Connect applications to a model server on a GB10 or GB300 reference platform using vLLM.
THE PLATFORM BEHIND THE MACHINE
GB10 / DGX Spark
8 catalog configurations
BEYOND THE SPECIFICATION
Lessons for running this machine.
Published guidance and checklists to help you investigate the full deployment.
Infrastructure
Ollama, vLLM, NIM: nine months of deployment learning.
UT’s reported journey across serving stacks, hardware integration and customer feedback. A starting record for the decisions and evidence still to publish.
UT field note
Startup & recoveryA NIM container that took about 30 minutes to start.
An operator-reported startup observation from UT’s GB10/GB300 work. The next step is to identify the exact configuration and separate the startup phases.
UT field note
Engineering applicationsLocal AI beside a Wi-Fi module test bench.
An execution agent and confidential engineering test assets show how on-premises AI extends beyond office applications.
UT field note
THE CONFIGURATION
Inside the machine
- Configuration
- 128 GB unified / 4 TB NVMe
- CPU / accelerator
- GB10 Grace Blackwell Superchip
- Memory
- 128 GB unified
- Memory bandwidth
- 273 GB/s
- Architecture
- arm64
- Operating system
- NVIDIA DGX OS
- Storage
- 4 TB NVMe
ON YOUR PREMISES
Space, power & connectivity
- Form factor
- Desktop AI system
- Dimensions
- 150 × 150 × 50.5 mm
- Weight
- 1.2 kg
- Power rating
- 240 W external adapter capacity; 140 W SoC TDP. Neither number is typical wall consumption.
- Connectivity
- 10 GbE · ConnectX-7 · Wi-Fi 7
- Availability
- Listed for order · recheck with supplier
BEFORE YOU CHOOSE
Memory capacity does not establish interactive speed or simultaneous request capacity.
UT has documented a Spark deployment. Confirm exact software versions and support scope.
Manufacturer specificationsHelp improve this record. Suggest a correction
IN THE FIELD
Deployment experiences
Operating records from customers, contributors and the Lab.
Customer deploymentField record
Semiconductor company · Wi-Fi productsA GB10 beside every test bench: autonomous execution, retest and insight in a Wi-Fi NPI lab.
Two agents run the lab’s existing Python test automation through the controller API, judge results against the formal pass/fail protocols, retest within protocol and report. After a three-month pilot, each test bench received its own GB10-class workstation.
Read the deployment recordCustomer deploymentField record
Accounting & advisory firmAn accounting and advisory firm runs its legal document automation on premises.
A GB300-class appliance in production and a GB10-class appliance for development, installed at the firm, run the Understand Tech platform locally. The first application automates legal documents; the firm holds its source code.
Read the deployment record
Operator deploymentField record
Two office Sparks serving a company’s AI stack.
Understand Tech moved its applications and shared inference from two rented cloud GPU instances to two DGX Spark units on the office floor: retrieval, agents and coding tools for the team.
Read the deployment recordCustomer deploymentField record
Biomedical research institute · university-hospital networkRare-disease diagnosis support from dental photographs and radiographs, planned on an on-site appliance.
A six-month pilot of a vision-plus-retrieval pipeline that turns intra-oral images into standardised phenotypes and weighted, source-cited rare-disease hypotheses. Production is designed for a DGX Spark-class appliance on site, air-gap capable, for health-data sovereignty.
Read the deployment recordWhere does your AI run?
In an office, a factory, an engineering lab or beside your test equipment. Share a deployment and keep the people behind it credited.
Contribute a deploymentTHE SHARED TECHNICAL LIBRARY
Knowledge around this machine.
Supplier recipes, community experience and UT’s deployment lessons, connected to the hardware.
OperationsSupplier
Manage a Spark fleet and prepare offline installation
NVIDIA documents enterprise lifecycle integration and custom installation patterns.
For your deployment
Example tools are not preinstalled. Adapt to your management system and check the documented support boundaries.
RecipesSupplier
NIM inference on DGX Spark
NVIDIA’s model-serving playbook.
For your deployment
Check the selected NIM profile, image digest, memory use and readiness time.
RecipesSupplier
Serve models with vLLM
NVIDIA’s vLLM setup reference.
For your deployment
Pin the container and test your model, prompt lengths and simultaneous requests.
THE LOCAL AI WATCH
All updates Updates for this machine
Loading updates…
MEMBER RATINGS · SELF-DECLARED, NOT MEASURED
How members rate the DGX Spark.
One rating per member, tied to a Lab profile. Stars count immediately; written verdicts are published after a quick review by the Lab team. A rating is an opinion about fit for a job, not a benchmark.
THE PEOPLE BEHIND THE MACHINES
Experience with DGX Spark?
Questions, ideas and experience from putting AI machines to work. You do not need a finished deployment to contribute.
Loading the discussion…