# NVIDIA launches 64 GB DGX Spark for local AI development

> DGX Spark systems with 64 GB memory ship on 2026‑10‑23, letting developers run large language models on‑premises.

Oossa · 2026-10-21 · https://oossa.com/en/nvidia-launches-64-gb-dgx-spark-for-local-ai-development

NVIDIA announced that 64 GB DGX Spark systems will start shipping on Friday, Oct. 23, 2026. The new machines can run models such as Nemotron 3.5 Bolt 74B, Qwen 3.8 27B and Gemma 4 26B‑A4B entirely on‑premises. They are available through partners like Acer, ASUS, Dell, GIGABYTE, HP, Lenovo and MSI.

## The facts

- DGX Spark 64 GB systems ship on 2026‑10‑23.
- They support models up to 74 billion parameters, e.g., Nemotron 3.5 Bolt 74B.

## Why it matters

Small AI teams can now run frontier‑class language models locally, reducing dependence on cloud services.

## Sources & references

1. [Bringing the Spark to Europe: Local Intelligence Powers Next Wave of Agents](https://blogs.nvidia.com/blog/local-ai-gtc-rtx-dgx-spark-station-sync/) – NVIDIA, 2026-10-21
2. [Tanium Brings Enterprise-Grade Manageability to NVIDIA Local AI Infrastructure](http://www.businesswire.com/news/home/20261021365024/en/Tanium-Brings-Enterprise-Grade-Manageability-to-NVIDIA-Local-AI-Infrastructure/?feedref=JjAwJuNHiystnCoBq_hl-WsGkFyUNMTSNrPsg57HLEqqcp-o_pnudlUwsb5apQ1S4gUE65BTfjH3-pSuqdv0gW3cb3F4oTIgUqCPafFkgu5AneicJQDNboWvodvG82E0) – Business Wire, 2026-10-21

Last updated: 2026-10-21