Nvidia DGX Spark Price Hikes $700 Due to Memory Shortages

by priyanka.patel tech editor March 10, 2026

written by priyanka.patel tech editor March 10, 2026

Nvidia has announced a price increase for its DGX Spark AI mini-PC powered by the GB10 Grace Blackwell Superchip. As per Nvidia’s official developer forums, the company is raising the price due to the constrained memory supplies worldwide. The MSRP has now been revised from $3,999 to $4,699, a substantial increase of $700. Nvidia also notes that while the new pricing is already reflected on Nvidia.com, other sales channels may take longer to update.

Go deeper with TH Premium: GPUs

(Image credit: Noctua)

First showcased at CES 2025 as Project Digits, Nvidia started selling the DGX Spark in October 2025 after facing a delay in May. An AI supercomputer that can easily fit on your desk, the DGX Spark is highly portable, measuring just 5.9” by 5.9” by 2”, and is ideal for AI developers, researchers, and data scientists. According to Nvidia, it has been custom-tailored to the needs of local AI inference and development and can run inferences on many of today’s popular AI models.

DGX Spark has been made in collaboration with MediaTek, featuring 10 Arm Cortex-X925 and 10 Cortex-A725 CPU cores and a Blackwell GPU capable of delivering up to 1 petaFLOP of AI inferencing performance. There’s also 128GB of LPDDR5X coherent unified memory that can be shared between the CPU and GPU, as well as a 4TB NVMe SSD for storage. The device itself features a USB-C power input, three USB-C 20Gbps ports with DisplayPort alt mode support, an HDMI 2.1a port, a 10Gb Ethernet port, and two QSFP ports for the onboard ConnectX 7 NIC running at up to 200 Gbps.

With the ability to support up to 200-billion-parameter models locally (assuming FP4 quantization), one can even connect two of these units using the built-in Nvidia ConnectX 7 NIC to gain double the performance. The DGX Spark runs on Nvidia’s DGX OS, which is based on Ubuntu and supports the CUDA software stack, which has become crucial for AI development. Nvidia has also roped in a range of software partners to ensure that their tools work well with DGX Spark, including Anaconda, Cadence, ComfyUI, Docker, Google, Hugging Face, JetBrains, LM Studio, Meta, Microsoft, Ollama, and Roboflow.

If you’re interested in the DGX Spark, several OEM partners, including Dell, Asus, MSI, and HP, also offer similarly designed systems built around comparable hardware, delivering similar levels of performance. In fact, the Asus Ascent GX10 is a solid alternative to the DGX Spark as it is cheaper at $3,266.53, with the only difference being a smaller 1TB SSD.

Nvidia confirmed the updated pricing for the DGX Spark took effect this week, but that existing orders will be honored at the pricing at the time of ordering. No hardware changes have been made to the device as part of the increase.