After VLC, Jean-Baptiste Kempf sets out to build the nervous system of the machine economy

Artificial intelligence is gradually learning how to act upon the physical world. Yet between a model capable of making a decision and a machine capable of executing it, a critical technological layer remains largely missing.

Autonomous drones, industrial robots, self-driving vehicles, and robotic medical systems must continuously exchange video streams, sensor data, and control commands with latencies measured in milliseconds. This is a requirement that today’s Internet infrastructure, designed primarily for communication and content delivery, was never optimized to handle.

As the AI industry shifts toward robotics, autonomous systems, and defense applications, this challenge is becoming increasingly strategic. After models, semiconductors, and data centers, a new category of companies is now emerging to build the infrastructure that will allow AI to interact with the real world.

This is precisely the layer targeted by Kyber, a company founded by Jean-Baptiste Kempf, creator of VLC Media Player and lead architect of FFmpeg, one of the world’s most widely used video-processing software libraries. The company has announced a $5 million Seed round, led by Lightspeed Venture Partners with participation from OVNI Capital and Kima Ventures.

Behind this funding round lies an ambition that extends far beyond robotics software. Kyber aims to build the real-time infrastructure that will enable both human operators and autonomous agents to control, monitor, and train machines from anywhere in the world.

Infrastructure built for real-time operations

Most existing systems today rely on technologies originally developed for the consumer Internet, particularly WebRTC, the protocol widely used for video communications.

These tools were designed to enable communication between people, not to continuously coordinate robots, drones, or autonomous systems operating in mission-critical environments.

In the physical world, a few hundred milliseconds can represent several meters traveled by a drone or autonomous vehicle before a command is executed. In industrial, medical, or military applications, such delays can quickly become operationally unacceptable.

Kyber claims it can reduce latency to approximately eight milliseconds by synchronizing within a single infrastructure all the data streams required for an autonomous machine to operate.

The goal is to provide developers with a unified software layer capable of simultaneously managing video, audio, sensor data, and control commands while ensuring temporal consistency across all streams.

Following the path of VLC and FFmpeg

Jean-Baptiste Kempf’s ambition follows the same logic that made VLC and FFmpeg successful.

Neither VLC nor FFmpeg became major businesses through direct consumer monetization. Instead, they evolved into foundational infrastructure technologies embedded in millions of applications, services, and devices worldwide.

Their adoption was largely driven by open source, which enabled widespread integration by developers and enterprises. Over time, these software components became de facto standards within the global video ecosystem.

Kyber is adopting a similar “open-core” strategy. Part of the technology will be released as open source to encourage ecosystem adoption, while advanced features, management tools, enterprise support, and commercial licensing will form the company’s monetization model.

For investors, this approach is particularly attractive. In software markets, technical standards established early in a technology cycle often capture a significant share of the value ultimately created by the broader ecosystem. The question is no longer how many robots or drones will be sold, but how many of them will rely on the software layer that connects operators, sensors, and AI systems.

An investment thesis aligned with the next wave of AI infrastructure

Lightspeed’s backing reflects this perspective. The American venture capital firm has significantly increased its investments in AI infrastructure, enterprise software, and autonomy-related technologies.

For Antoine Moyroud, Partner at Lightspeed Venture Partners, the diagnosis is straightforward: AI models are advancing rapidly, but the infrastructure required to connect them to the physical world remains underdeveloped.

This view aligns with a broader trend across the industry. After financing foundation models, GPUs, and data centers, investors are increasingly turning their attention to the software layers that will enable autonomous systems to operate in real-world environments.

Robotics, defense, industrial automation, and telecommunications currently represent the most immediate target markets.

Building infrastructure for the machine economy

The rise of autonomous systems could fundamentally reshape the nature of digital infrastructure.

Over the past two decades, the Internet has primarily been designed to connect people, content, and applications. The next phase may require infrastructure capable of connecting millions of machines that can perceive their environment, make decisions, and act autonomously.

In this context, communication protocols, real-time infrastructure, and control layers could become as strategically important to the machine economy as cloud platforms became to the software economy.

This is the hypothesis underlying Kyber’s strategy. After helping build some of the foundational technologies of digital video, Jean-Baptiste Kempf is now seeking to develop one of the infrastructure layers that could support the next phase of artificial intelligence, and potentially create a successor to WebRTC for the era of robots, drones, and physical AI.