Nebula Labs launches 'Nebula Edge' — a private LLM appliance for studios

Tech · 5 min read

Nebula Edge pairs custom inference hardware with a software stack that supports multiple LLMs and multimodal models in a private, offline environment. The appliance is sold through a subscription model that includes model hosting, security updates, and optional model fine-tuning services.

The company positioned Nebula Edge as a response to studios facing privacy constraints and unpredictable cloud costs when using third-party APIs. Nebula claims sub-10ms token latency under load for common generation tasks and built-in tooling for dataset management and evaluation.

Early customers include a major indie game studio and a film visual effects house testing on-prem image and dialogue generation workloads. Nebula Labs also announced a $40 million preemptive expansion of its Series A credit line to support manufacturing and logistics for appliance deliveries.