Dell launches first desktop powered by Nvidia GB300 AI superchip

Tuesday 17 March 2026 - 12:39

By: Dakir Madiha

Dell launches first desktop powered by Nvidia GB300 AI superchip

Dell Technologies has introduced the first desktop computer equipped with Nvidia’s GB300 Grace Blackwell Ultra Desktop Superchip, marking a shift toward bringing data center level artificial intelligence computing to enterprise workstations.

The announcement was made at Nvidia’s GTC 2026 conference in San Jose, where Dell confirmed it is the first original equipment manufacturer to commercialize a system using the GB300 processor. The new Dell Pro Max delivers up to 20 petaflops of AI performance in FP4 precision and includes 784 GB of unified memory, targeting companies developing autonomous AI agents.

Dell executives said the system is designed to address growing demand for local AI processing, particularly for organizations handling sensitive data. The desktop integrates Nvidia’s OpenShell runtime, which enables policy enforcement and security controls for AI agents operating without reliance on cloud infrastructure.

The GB300 chip combines a 72 core Arm based Grace CPU with Nvidia’s Blackwell Ultra GPU architecture. The GPU integrates two chips connected through high speed interconnects, delivering 640 tensor cores and up to 8 terabytes per second of memory bandwidth. The system includes 288 GB of HBM3e GPU memory and 496 GB of LPDDR5X CPU memory, linked through NVLink C2C for unified access.

Networking is handled by Nvidia’s ConnectX-8 SuperNIC, offering dual 400 gigabit connections. The desktop is built on the Alienware Area 51 chassis and uses a closed loop liquid cooling system. Dell’s MaxCool technology improves heat dissipation significantly, supported by a metallic thermal compound that reduces resistance between the chip and cooling system.

Dell also introduced a smaller Pro Max configuration powered by the GB10 chip, offering one petaflop of performance and 128 GB of memory. Both systems are designed to run Nvidia’s NemoClaw framework, an open source platform that allows enterprises to deploy AI agents capable of executing complex, multi step tasks with integrated privacy and security controls.

The GB300 system can run models with trillions of parameters entirely offline, enabling faster response times and stronger data protection. Dell and Nvidia are also developing an isolated network version for government clients handling classified information. Pricing has not been disclosed, though shipments to select customers have begun.