Dell Expands AI Factory With NVIDIA, Plans New Servers for 2027

Dell Technologies expanded its Dell AI Factory with NVIDIA on Oct. 21, adding tools for local AI agents and planning new data center hardware as customer orders show broader use of the platform.

Dell counted more than 6,500 AI Factory customers based on its analysis of August 2026 customer order data, up from 5,000 last quarter. The AI Factory platform already spans deskside workstations and PowerEdge servers for enterprises running AI on their own infrastructure.

The new deskside templates cover agentic coding, agentic VSS, NVIDIA NemoClaw, Hermes Agent on NemoClaw and an AI lab platform for higher education. Available on GitHub, they outline the hardware, models and software for each deployment. Dell designed them to scale from one developer to 1,000 or more systems.

Dell introduced its Deskside Agentic AI offering in May 2026 for running agent workflows locally. That earlier offering extended NVIDIA OpenShell support from workstations to PowerEdge servers. The latest templates provide configurations for specific tasks within that hardware range.

Dell also plans a 64GB unified-memory version of its Pro Precision with GB10. The Windows Pro Precision with GB300 has no announced release date. Its listed maximums are 748GB of coherent memory and 20 petaflops of FP4 compute.

For larger installations, Dell is preparing a containerized design that combines computing, storage, networking, rack-level power and liquid cooling. It uses NVIDIA’s DSX POD design and is scheduled for early 2027.

The planned PowerEdge M9882 fits 256 NVIDIA Vera CPUs in a 48U rack. Another model, the XE9810, has 16 direct-liquid-cooled compute trays powered by Groq 3 LPUs. Dell has scheduled both servers for the first quarter of 2027.

The two-socket, air-cooled PowerEdge R8822 is scheduled for October 2026. The next dated workstation release is the 64GB Pro Precision with GB10 configuration on Oct. 23.