One of the items that was discussed and mentioned in my previous blog on Michio Kaku was the fact that computing is “everywhere and nowhere”. Part of that is the result of Moore’s Law, which the good doctor had a chart on. Many of us have seen this chart at various times over the years. The point was that the cost of compute is driving towards a penny. When things are this cheap, they proliferate as businesses and consumers find new ways to use computing power. Data centers and the cloud are to become commonplace, or so posited the theoretical physicist.
As part of Dr. Kaku’s talk, there was a Q&A session at the end. Now it is pretty cool to get to have the opportunity to ask him a question in an open, but small session. I, of course, was the equivalent of Arnold Horshack in my eagerness to ask away at my own version of Mr. Kotter. Here was the question (along with the obligatory comment to make me seem smart – which probably just made me appear arrogant….):
“With the proliferation of data centers and compute everywhere, what is going to power it all? It does not seem to me that we have made any significant strides in battery technology since Edison fought about it with Westinghouse.”
Now, I was fully expecting a different answer that I got. After all, this guy makes a living out of seeing the future. What he said was equally shocking, as well as great for the prospect of the data center industry. Here are the main points that he discussed:
1. It is true that power is a great quandary. There is nothing better than the ‘liquid sunshine’ that is gasoline in terms of joules/gram. Fossil fuels will continue to be the best, most efficient source of power.
2. Battery technology is stuck. We are 50 years plus away from new innovation that can start to break the stronghold that fossil fuels have. We may be able to use nanotechnology for better batteries, but we are a long way off.
So what does this mean for us? I take away the following: The power outlet business (aka data centers) is not going away. As a matter of fact, if the Great Karnack can’t even see a scenario in which there is a massively disruptive technology change for powering compute and storage, we have a great run in front of us! Apparently, Moore’s law only worked for silicon, not the power source.