Bright AI runs DeepSeek-Coder-V2-Lite 16B locally on an RTX 2050 laptop

Hello NVIDIA Developer Community,

We are developing Bright AI, a software technology focused on making large local AI models practical on affordable hardware.

Our recorded demonstration shows DeepSeek-Coder-V2-Lite-Instruct, a 16-billion-parameter AI model, running locally on an NVIDIA RTX 2050 laptop.

Running a model of this size on a laptop in this hardware class is normally considered impractical. Bright AI made it possible to run the model locally and communicate with it.

We believe this technology could help bring powerful local AI to affordable computers, robots, drones, autonomous machines, and other edge devices while reducing dependence on expensive hardware and large data centres.

We are looking for technical feedback, potential technology and hardware partners, software licensing opportunities, and investors interested in helping bring this technology to market.

Recorded demonstration:

Thank you.