DGX Spark getting started: from the box to the first running model

Before you switch it on

To get started, attach the peripherals or the Ethernet cable first and plug in the 240 W adapter last, because the Spark starts as soon as it has power. Setup runs at the desk with a display or headless over the Spark’s Wi-Fi hotspot, and there is no default password: the wizard creates the first account. The box holds the Spark, a 240 W power adapter, a power cord and a Quick Start card. It does not hold a display, a keyboard, a mouse or the cable for linking a second Spark. The system itself is 150 by 150 millimetres and weighs 1.2 kg; NVIDIA gives 5 to 30 °C as its ideal operating temperature, and its safety notes ask for a grounded socket, a ventilated spot for the adapter and some distance from radiators and other heat sources.

Prepare the network and the order of connections. Setup needs a fast, reliable internet connection over Wi-Fi or Ethernet, and NVIDIA advises against captive portals, such as hotel or airport Wi-Fi, and against phone hotspots. NVIDIA’s guide is explicit that “the DGX Spark device starts up immediately when power is applied” and asks for every peripheral to be attached first, so the power adapter goes in last. The power button is on the back, marked ON/OFF in NVIDIA’s Quick Start Guide; after a shutdown, press it to switch the Spark on again. We found no hold time for a forced power-off in NVIDIA’s user guide, so shut the Spark down from DGX OS.

Setup with a monitor or headless over Wi-Fi

NVIDIA offers two routes, and the choice only affects the setup itself; afterwards the Spark can be used either way.

At the desk. Connect a display, a keyboard and a mouse, wired or over Bluetooth, then power on. If a display on USB-C stays dark, NVIDIA’s advice is to try HDMI instead.

Headless, over the network. Power on without a display, and the Spark “creates a Wi-Fi hotspot that you will use to connect to the system”. The network name and password are on a sticker attached to the Quick Start Guide, which also lists the address of the setup page to open in a browser; the card gives it as http://spark-xxxx.local, with the Spark’s own name in place of xxxx. If the Spark cannot be found after it joins your network, which NVIDIA says can happen on complex corporate networks, the fallback is a display, keyboard and mouse. Since the October 2026 release, a Wi-Fi Assistant in DGX Dashboard lets a Spark without display, keyboard and mouse join a new Wi-Fi network after setup.

Either way, the wizard asks for the language and time zone, the keyboard layout for the desk setup, acceptance of the terms, analytics preferences (which can be skipped) and the Wi-Fi network, a step it skips when an Ethernet cable already provides internet. It also asks you to create the first user account, which gets administrator rights. Then it downloads and installs the software image and restarts, possibly more than once. NVIDIA’s guide says “Do not shut down or reboot the system during this process” and gives up to 10 minutes for installing the updates and joining the network. Since the June 2026 release, further over-the-air updates are no longer installed by default during setup, so the Spark is usable sooner and the remaining updates come afterwards through the Dashboard.

For fleets there is a third route. Since April 2026, IT can skip the setup wizard entirely and install the systems with cloud-init, and PXE boot is supported.

For a rollout of several Sparks, for example two or four, you can order the Sparks alone, their setup alone, or both together: the Sparks delivered with everything installed and configured; tell us how many you need and you get a written quote.

What is installed when setup finishes

The Spark runs DGX OS, NVIDIA’s Ubuntu 24.04-based operating system. As of 5 November 2026, the current Founders Edition software is still NVIDIA’s October 2026 release (release notes updated 4 November 2026): DGX OS 7.6.0 with Linux kernel 7.0 (7.0.0-1019), GPU driver 580.178.04, CUDA 13.0.3 and TensorRT 10.14.1. NVIDIA notes that GB10-based partner systems may not receive updates at the same time. Docker is installed and the NVIDIA Container Toolkit is, in NVIDIA’s words, “preinstalled and configured”; to run containers without sudo, give your account Docker access with sudo usermod -aG docker $USER, then run newgrp docker or log in again. The GPU reports CUDA compute capability 12.1, so software built from source should target it; NVIDIA’s playbooks give the build settings. Canonical’s security maintenance for Ubuntu 24.04 runs to 2034 with Ubuntu Pro; pro status shows whether an Ubuntu Pro entitlement is attached to the system.

NVIDIA Sync and the DGX Dashboard

Most work on a Spark happens from another computer. NVIDIA Sync is a tray application for Windows, macOS and Ubuntu that finds the Spark on the local network by its hostname and sets up key-based SSH access; NVIDIA states that the password it asks for “is used temporarily for set up and is not logged or stored”. From Sync you open a terminal, a code editor such as Cursor, NVIDIA AI Workbench or the DGX Dashboard on the Spark, and you can add your own applications, whose ports Sync forwards to your laptop. For access from outside the office, Sync has Tailscale built in. NVIDIA estimates its remote-access playbook at 5 minutes.

The DGX Dashboard is the Spark’s own web interface, reached through Sync or through an SSH tunnel on port 11000. It shows the system’s state, runs updates and starts JupyterLab. You log in with your Linux account; updates need administrator rights. Each account gets its own JupyterLab on its own port, with a working directory under its home folder and a Python environment that is created and filled with recommended packages on first start. NVIDIA calls the Dashboard “the primary and recommended way to perform system updates”; the command line, with apt for packages and fwupdmgr for firmware, installs the same updates. Login, JupyterLab ports and remote access are covered step by step in our DGX Dashboard guide.

The quickest route to a first model

NVIDIA’s playbooks at build.nvidia.com/spark are step-by-step guides, and since the June 2026 release the playbook site opens after the first boot. The quickest route to a chat model is the first in this table.

build.nvidia.com/spark and NVIDIA’s dgx-spark-playbooks repository, October 2026. The Open WebUI playbook needs about 7 GB for the container and about 15 GB for gpt-oss-20b.

Start with Open WebUI and gpt-oss-20b to confirm the machine works end to end, then move to vLLM when other people or applications need access. Before anyone else on the network uses the chat, check the versions in the container: Ollama 0.35.0 or later fixes a flaw in its model download API (CVE-2026-103663, 8 October 2026; critical, CVSS 9.4, where the Ollama service can write to /usr/lib/ollama, otherwise 6.9), Open WebUI 0.12.2 or later fixes the advisories published in September and October 2026, and the Ollama port 11434 should not be reachable from the network. For how many people a Spark can serve, see our article on sharing one DGX Spark; for which models fit, see what fits in 128 GB.

If the Spark is to serve a team, we set up the serving engine with access control on it; write to us with the number of users and the models.

Common first-day issues and what they mean

NVIDIA DGX Spark user guide, known issues and release notes; answer by NVIDIA staff on NVIDIA’s developer forum (November 2025).

Factory reset from a USB drive

The Founders Edition can be returned to its factory state from a USB drive. NVIDIA’s recovery procedure needs a USB drive of at least 16 GB, a wired USB keyboard and a display on the Spark, and the recovery archive from NVIDIA’s website, which comes with a script to create the drive on Windows, Linux or macOS. In the UEFI you restore the defaults, switch Secure Boot off, boot from the drive and start the recovery, which erases the internal drive completely; Secure Boot is switched back on automatically at the end. Keep the data you care about somewhere else before you start. Partner GB10 systems use their makers’ own recovery images, as our EU buying guide explains.

What we supply

We supply the NVIDIA DGX Spark Founders Edition (128 GB, 4 TB) across the EU on one contract and invoice, with manufacturer warranty.

FAQ