VLC’s Jean-Baptiste Kempf raises $5M to control robots via Kyber

Jean-Baptiste Kempf, the lead developer of VLC Media Player, just raised $5 million for his startup Kyber. The pitch is not another media app or another streaming feature. Kyber positions itself as infrastructure for controlling remote devices in real time, and the ambition is to connect hundreds of millions of robots.

This matters because “real time” is the difference between a robot that looks smart in a demo and one that can actually operate in the messy world. When you control remote devices, you are not just sending commands. You are coordinating state, timing, connectivity, and reliability across networks that are unpredictable by design. In other words: the software layer between humans, systems, and far-away machines is becoming a product category, and Kempf is betting he can own part of it.

Kyber’s $5 million round was led by Lightspeed Venture Partners. The investor is familiar with fast-moving bets, including leadership in Mistral AI’s record-breaking seed round, and Lightspeed has since invested in Anthropic as well. That combination of AI-native credibility and early-stage scale is a meaningful signal for how the market is thinking about “infrastructure.” Boards tend to fund infrastructure when it has two properties: (1) it removes complexity from a lot of downstream applications, and (2) it becomes a default dependency. Kyber’s framing, as described in the source, is exactly that kind of dependency play.

The round also includes OVNI Capital and Kima Ventures, according to the source. While the Next Web excerpt does not list the exact ownership breakdown or term details, the presence of multiple early-stage investors usually indicates more than pure interest. It can suggest that multiple teams see Kyber’s problem as urgent, not speculative. In robot-control infrastructure, urgency is baked in. Robotics is moving from lab settings to logistics, industrial operations, and consumer-adjacent devices, where even small latency or reliability failures can cause expensive downtime or safety issues. If Kyber is aiming for a massive number of connected robots, the operational requirements become non-negotiable.

It also helps to remember what Kempf brings to the table. VLC is a rare example of software that achieved broad adoption by making complicated playback and codec ecosystems feel effortless. The through-line in Kyber is the same instinct: take fragmented, hard-to-integrate systems and provide a layer that makes them work together reliably. Translating that habit from media distribution to remote device control is not automatic, but it is at least coherent. For executives, the question is whether Kyber can become the “plumbing” that multiple robot makers and operators do not want to rebuild from scratch.

Now zoom out to second-order implications. If remote-control infrastructure gets centralized, it changes who holds leverage. Robot manufacturers and fleet operators can care about autonomy, but they still need a dependable control plane. An infrastructure layer like Kyber can become the place where observability, command routing, and network handling converge. That often turns into a governance and compliance story too, not just an engineering one.

Regulatory and policy pressure is already rising across remote operations, data handling, and critical systems, even before robots are everywhere. The source does not specify any regulatory jurisdiction for Kyber, but the regulatory reality for remote control products is clear: when you control devices over networks in real time, regulators and customers will ask about security, auditability, and safeguards. Even if Kyber is early, its infrastructure choices can lock in future compliance costs or simplify them. Boards should think of this as risk management as much as product strategy.

Finally, there is the market timing. Investing in robot infrastructure is a classic move when the category is still forming, because the winners can set standards before everyone else has aligned. Lightspeed leading the round, with prior activity that includes a record-breaking seed investment in Mistral AI and later investment in Anthropic, suggests it is comfortable backing foundational layers rather than only application-layer features. For peers, the strategic stakes are simple: if Kyber succeeds at the “hundreds of millions of robots” ambition, it becomes a force multiplier for anyone trying to scale remote operations. If it stumbles, it still proves that “infrastructure for real-time device control” is now on the investor scoreboard, which can accelerate competitors and partnerships.

In short, this is Kempf leaving the VLC glory zone and jumping into the real-time control layer for remote machines. The $5 million round is not huge compared to mega-rounds in AI, but in infrastructure startups, early capital can decide whether a platform becomes a default dependency. And once you are positioning for massive robot connectivity, that dependency question stops being theoretical fast.