BharatGPT-mini has been demonstrated running on a laptop without any internet connection or cloud dependency. CoRover.ai showcased the small language model at Intel's AI PC Innovation Event held at Conrad Bengaluru on April 24. The model operated as a conversational AI agent on Intel's Core Ultra Series 3 processor, handling text, voice, image, and video in multiple Indian and international languages—all without an internet connection.
On-Device Processing with Intel OpenVINO
The model processes everything locally using Intel's OpenVINO toolkit, which optimizes inference across the chip's CPU, NPU, and accelerators. This ensures efficient performance without relying on external servers.
Who Needs Offline AI?
Patchy connectivity is a key reason for offline AI. Additionally, sectors like defense, healthcare, and banking require data to remain on-device due to legal or practical constraints. On-device processing keeps all data local, addressing these needs.
CoRover CEO Ankush Sabharwal, speaking at the event, described a three-tier architecture: on-device for speed and privacy, on-premise for control, and cloud for scale. He emphasized that each layer serves a different purpose, with on-device handling scenarios where the other two are unavailable.
Existing Deployments and Future Potential
CoRover already has deployments across government, BFSI, education, and defense verticals. The Intel demo extends this capability to individual PCs—field offices, classrooms, and desks where connectivity is unreliable or data sensitivity is high.
Sabharwal's broader pitch: cloud AI made technology accessible, but on-device AI makes it usable in places that have so far been left out.
