Kyber targets low-latency remote control and rapid updates across large-scale deployments where operators, compute, and action are geographically separated, building an infrastructure layer for real-time management.

The company is positioned for physical AI across fleets of devices, with a vision of hundreds of millions of robots and drones, and has raised $5 million in a round led by Lightspeed.

Key challenges include scaling without performance drops and choosing between an open-source core with a commercial enterprise version to attract developers and large enterprises.

Kyber operates on an open-source core model complemented by productized enterprise offerings and deployment support through forward-deployed engineers.

Investors are watching Kyber as current robotics and drone systems are expensive and custom-built, with Kyber aiming to scale to millions of devices and compete with remote IT providers.

The platform blends video-streaming optimization with IoT device tuning to maximize performance based on each device’s compute capacity, enabling scalable operation from a few devices to millions.

Kyber is headquartered in Paris with expansions to San Francisco and Singapore, signaling an international push for deployments and enterprise adoption.

Upcoming indicators include onboarding large enterprise clients, performance in large real-world deployments, and resilience across varied networks and regions.

Observability and reliability are critical as AI agents manage fleets, with ongoing updates and verification across distributed environments.

The company’s VLC roots inform its focus on low-latency streaming and remote control, with founder Kempf’s background as a VLC creator and open-source advocate.

Led by Jean-Baptiste Kempf, Kyber is an infrastructure platform to control remote devices in real time by synchronizing video, audio, sensor data, and control inputs with minimal latency.

Kyber aims to standardize low-latency communication for managing large fleets, making distant machines feel local and advancing physical AI.