Niklas Zennstrom is CEO and co-founder of Skype. Kelly Larabee sat in on the conference, and Zennstrom called in via phone. While it was a challenge to hear much of his presentation given crowd noise in the hall, what follows is a partial (partial!) transcript of his Supernova remarks:
Thank you for letting me participate in this way. What I want to talk about is when telephony becomes software, which is what’s really happening right now. We’re going to see a transition. What does that mean? One key factor is that when telephony becomes a software application, it will live on the edges of a network. There will be no centralized control. Telephony development cycles are still very long. But their carriers’ services are very much commodities, and there isn’t much differentiation.
When we can change telephony to an application, we’ll be able to change the economics of it. That will increase software innovation, new products, services, features. And that will be in the hand of small, nimble, software development companies. There will be higher competition, and the services will live on the Internet.
Another issue is peer-to-peer vs. client-server-client. What you will see is that virtually all voice over IP systems will look alike. And there will be incremental costs for managing the networks. Computing will move to the users, and we will see several benefits. We will see different economic models. If we look at ourselves as software developers rather than carriers, our marginal costs will be close to zero and we can provide these services for free. In the future, we will not be able to charge for telephony over the Internet. You will have to find other revenue sources. That could be anything: value-added services or advertisement.
There will also be increased robustness. If you have a switch go down or a gatekeeper, that’s not a problem to the end user. They can go to the Internet and find another route of communication. That will not be a problem for a decentralized network. A third benefit will be increased quality. If I can to call Germany using an American provider, it would go to a server in America and then one in Germany. That means there would be latency. If it’s a decentralized system, the call would go a shorter path between end points.
And another benefit would be privacy. End users will feel more safe. If you call them up using a decentralized server, who provides the service will not be able to record the conversation or calling patterns for any reason.
What are the implications for the telecommunications market? Before, services were integrated in the network. Now they become a software application. Before, we have an “intelligent network” even though it’s not that intelligent. After, the network would be dumb and all intelligence would be at the end points. In the telephony world, because services define the network, it becomes a commodity. Telecom companies acquiring customers need to spend a lot of money to convince people their services are better or cheaper. Decentralized systems bring low or no acquisition cost. It’s very easy for users to recommend services to their friends. We don’t spend anything on marketing.
In the old network, voice is the main traffic voice. Now, voice can be marginalized in terms of volume. There will be other kinds of traffic. Now, we have regional players. With a global network, there will be truly global competition. The players don’t need to be like AT&T. They can be very companies with very few people. And the last thing, which is most important, is that it will move from a provider-driven market to a customer-driven market.
Economically, value will shift to more productive areas. Disruptive innovation has to be good for the economy. That does not mean that the segment you’re operating in will grow. The market may shrink, but there may be other opportunities. Today, voice represents about half of the revenue. As voice becomes free, it could be good for telecom operators because we could grow broadband penetration. As a telco, you can provide new services. There are other segments that will develop. On top of all that, there is the voice over Internet segment, which is not part of the voice segment. You don’t make money on minutes. Voice over IP will be significantly smaller. It has to be, or the innovation did not serve any purpose.
I want to talk a little about regulation. Decentralized voice over IP has an implication on regulation. The regulators have been trying to regulate this saying it’s the same service, but it’s not. You regulate a market that is subject to monopoly. You want to make sure consumers are protected. There is no such monopolist in voice over IP. Traditional telecom regulations will not help the market grow, will not help consumers, and will not drive regulation. You should not and you cannot apply regulations designed for network operators to software providers.