There's $20 billion a year in venture capital, and our last fund was about $1.5 billion. Our limited partners are 400-year-old universities, pension funds, and sovereign wealth funds -- literally countries -- and they have a 40-, 50-year time horizon with big payouts for pensions in the out years.
We take this money and route it into the craziest companies. We ignore the day-to-day and try to figure 10 years out what markets will look like.
For most of human history, this kind of money wasn't available, but we have this moment in time when there are very few places people can invest to get this return.
Today's startups don't take much money to start. A Series A round used to be $20 million, but now it's half a million. That's created a generation of seed investors that put a small amount of money in a large number of companies.
But when something works, it's more expensive to scale. GitHub raised $100 million because they had critical mass and needed to scale to host the world's software. They want to be the new place for source code control and software sharing.
On Microsoft buying Nokia's phone business:
I think we should all hope Microsoft is successful, because our industry does best when there is competition at the platform layer. It would be really good if Apple, Google, and Microsoft were all successful. Microsoft has a lot of smart people, and I hope they do well.
Not long ago, Apple was considered dead. When Netscape was coming up, no one thought Apple could come back. There's an inherent unpredictability in our business. Whatever we think is obvious today is almost certainly not the case.
By the end of the decade, 5 or 6 billion people will have smartphones. We have the opportunity to think about a world where everybody has a supercomputer in their pocket and is connected all the time.
We are seeing glimmers of the next set of platforms after the smartphone, too.
I'm seed funding apps companies for Google Glass because I think it has been underestimated. It's not just glasses, but a built-in earpiece with voice recognition and synthesis. There will be a wide variety of devices to augment the human body for many uses, including making us healthier.
On ARM servers:
The same processors in smartphones will power servers in the future. Apps in the datacenter are built in Java or PHP or Ruby. There are intermediate runtimes between the apps and the processors, so the apps generally don't know what the processor is.
Most of these apps are not CPU but IO bound. There are huge issues of heat and space in the datacenter. The new chips will save money, power, and space, and apps will be unaffected. They may even be able to afford to buy more servers and run their apps even better. I think this will happen really quickly.
The backend is changing very fast, and we think there are huge opportunities for technologies that power that. The cost of the backend for running a modern datacenter app is falling very fast, so the app developers can get a lot more creative and assume they will have a hundred or thousand times more performance behind them. It's going to be a completely different world.
Andreessen: You could hire Spielberg to produce university courses.
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.