By Stef van der Ziel, founder of Jet-Stream. Pioneering streaming since 1994.

A recent post about the DPP platform cost benchmark asked why Europe hasn’t built a shared streaming technology stack by now, and answered its own question: ego and control.

Correct. But incomplete. Allow me to add thirty years of first-hand detail, because I have a confession to make: We keep building the pan-European streaming stack. It’s becoming something of a habit. The technology always worked. It’s the “pan-European” part that has a remarkable talent for self-destruction.

Let me take you through the highlights.

1994: Four viewers

In 1994 we started the first livestream on the world wide web. Total audience: four viewers. From there, growth went exponential, and everything that follows is a direct consequence of that curve.

1996: We accidentally broke the Dutch internet

In 1996 we streamed a live event that attracted rather more viewers than the infrastructure appreciated. Everything ran from central servers, and that central bottleneck didn’t just choke our stream. It congested ISP XS4ALL and rippled through the Dutch internet.

Lesson learned, and learned early: centralized streaming doesn’t scale. Distribution has to be… distributed. Revolutionary stuff, I know 😉

1997: The first streaming CDN. Federated, obviously.

So in 1997 we built what I’d argue was the first streaming CDN in the world. Temporary, built for a single event, but real. And federated from day one: streaming servers inside multiple ISP networks, central load balancing with active redirects. No bottleneck. The AMS-IX relieved instead of hammered. Traffic stayed where the viewers were.

The industry would spend the next two decades enthusiastically rediscovering this architecture. We’re patient people.

2000: It works! Quick, let’s kill it.

In 2000 we made it permanent: a national federated CDN, called StreamGate. Dutchview, Jet-Stream, several ISPs, and several Dutch broadcasters worked together. Edge servers inside the ISP networks, an active central load balancer on top. Technically a resounding success. It scaled beautifully. Backbones and the AMS-IX were offloaded. Hundreds of thousands of concurrent viewers. Hundreds thousands of public money not paid to Akamai.

Then this platform had to become formal, and here Europe revealed its true superpower.

The wholesale departments of the ISPs wanted serious money for traffic that originated inside their own networks, serving their own subscribers. Yes, you read that correctly: they wanted to charge for the privilege of us saving them money. The public broadcaster refused to pay for anything, even though lots of money was saved, which meant the commercial broadcasters wouldn’t pay either. The AMS-IX was proposed as a neutral governing entity. Even neutrality couldn’t save it.

Sound architecture, sound economics, zero willingness to share a win.

We shrugged, took the technology forward on our own as VideoExchange, a control plane for federated streaming with edges at ISPs and hosters, for live and on-demand, and grew revenue by 1,800% in five years. So no, the technology and the model were not the problem.

The standards effort: patents to the rescue

Broadcasters and telcos then asked the industry to standardize CDN federation. Jet-Stream participated, assuming standardization was about… standards. Cisco and Alcatel promptly flooded the effort with patents, on concepts where we held years of prior art from actually running federated CDNs in production. Independence went overboard, trust evaporated, and CDN federation standardization quietly died.

A standard captured by vendors is not a standard. It’s a moat with paperwork.

The warning: applause included, action sold separately

Some years later I gave a talk at the Mediapark Jaarcongres. The pitch: “I just flew back from London, and on the plane sat people wearing Netflix shirts. They are already here, maybe even secretly in this hall. While you are all watching each other’s TV market share, they are storming your market, and soon they’ll be fighting Amazon, Google and HBO right on your turf. To them, the Netherlands is a small battlefield. They don’t even find you interesting.”

And I said: “stop competing with each other. We have the AMS-IX with world-class connectivity. We have the technology. We can build large CDNs with European telcos. We lead in content format innovation. We have every opportunity to take on the world together, but we have to act now.”

There was warm applause. There was, and I want to be precise here, absolutely nothing else. The media world went back to staring at each other, and OTT took everything, exactly as predicted. I’d say I hate to say I told you so, but by now I’ve had a lot of practice.

The EBU: so close, they published our secrets

Then came the big one, the opportunity that could have redrawn the European map.

By then Jet-Stream had evolved into a multi-CDN platform: we could deploy our own CDN infrastructure at ISPs and blend it with existing commercial CDNs, with the goal of steadily growing European-owned capacity inside European networks and get more European control versus American influence. But technology alone doesn’t shift a market. You need to bundle purchasing power. Enter the EBU: European broadcasters, aggregated demand, shared interest. Perfect.

EBU was totally into this plan. Jet-Stream invested roughly one million euros in further development and won the EBU tender for all streaming and multi-CDN technology. European technology plus European purchasing power. Control over streaming distribution back in European hands. Sovereignty.

What could possibly go wrong? I’m glad you asked. The multi-year contract was reopened for renegotiation. IPR ownership was demanded on parts of our own technology. And despite the exclusive award, a parallel track appeared with a French party, which turned out not to be able to deliver, was sold within a year, and took the whole project down with it.

But my favorite part came after. Under strict NDA we had shared the crown jewel of our platform: our “good enough” multi-CDN algorithm. Instead of chasing the fastest CDN at any price, it selected CDNs on a minimum guaranteed performance threshold and then optimized on lowest cost. Unique, exclusive, shared with nobody. EBU published it. Openly, in a paper of their own, presented as their own invention. I suppose I should be flattered.

And here’s the punchline: with our algorithm in hand, published and free for the taking, with all of Europe’s aggregated demand behind them, they still didn’t manage to build the European multi-CDN they set out to build. Ego and control explain a lot in this industry. Occasionally, plain incompetence likes to contribute too.

The pattern, for those keeping score

1997 and 2000: federation works technically, dies on partners wanting to charge each other for saving each other money. Standardization: dies on a vendor patent land grab. The strategic warning: applauded, filed, forgotten. The EBU: technology and purchasing power finally aligned, undone by internal politics, an IPR grab, and the remarkable feat of failing even with the answers handed over.

So when someone asks “why hasn’t Europe built a pan-European stack by now?”, it was never a technology gap. Europe has built this capability repeatedly. Europe has also dismantled it repeatedly, from the inside, with impressive consistency, every time it started to work.

Meanwhile, Netflix didn’t beat Europe on technology. Europe had the technology first. Netflix beat Europe because it never had to negotiate with itself.

Why I’m genuinely optimistic this time

The question “would you move your own stack onto someone else’s?” assumes the choice is between my stack and your stack. That framing is exactly what killed every attempt above. Everyone protects their hub, their sunk cost, their control.

But there’s a third model, the one we started with in 1997: federation. Not one broadcaster’s platform absorbing the others, but shared, sovereign infrastructure with clear governance, where nobody hands their crown jewels to a competitor. The technology is mature. The economics, as the DPP numbers show, are brutally in favor. We’ve only been saying so since the nineties.

And something has fundamentally changed. Europe is finally waking up to how dependent it has become on American infrastructure. We were shouting this fifteen years ago, and presenting working alternatives, back when it was considered a charming niche obsession. Today it’s on the agenda of governments, broadcasters and enterprises across the continent. The realization is landing that by working together, Europe can build a much stronger, more sovereign and more robust internet than any of us can alone.

The awareness is there. The urgency is there. The technology has been there all along.

All it takes is setting aside the egos, the not-invented-here syndrome and the petty individual interests, in favor of the bigger picture that benefits everyone. And embracing innovation instead of opportunistically borrowing it. That last part still needs some work.

After thirty years, my conclusion is simple: the next attempt shouldn’t start with technology, and certainly not with a committee. It should start with governance that makes defection unattractive, protects the IP of the parties who bring the innovation, and makes cooperation profitable, from day one. Get that right, and the stack is the easy part.

I know. We’ve built it four times. Happy to make it five. Actually, we’re already building it. This time, Europe might finally be ready to keep it.

Stef van der Ziel is founder and director of Jet-Stream, streaming pioneers since 1994, operating sovereign European streaming infrastructure from the Netherlands.

Go back to see all blogposts