Little more than two years ago, Pat Gelsinger's CEO tenure at VMware looked shaky. The software company's stock price was down by half since his arrival in 2012, an apparent victim of the industry shift to cloud computing. But the naysayers were wrong; VMware is back and reborn. I recently caught up with Gelsinger in London.
Q: Is technology progress speeding up or slowing down?
Gelsinger: Both. Moore’s Law hasn’t stopped, but it’s slowing. We were on this doubling-every-two-years curve. But now it’s more like doubling every three years.
Q: Why?
Gelsinger: Because going from 14-nanometer chips to 7-nanometer ones is hard. That’s getting close to physical limits, and you need tons of capital to manufacture them.
Q: But you also said technology progress is speeding up. Explain.
Gelsinger: In the cloud era if you give me $1,000 and a credit card, I’ll give you the world’s largest—for the next two minutes—supercomputer, a larger supercomputer than was ever assembled before. Entrepreneurs around the world can take advantage of that.
Q: You use the phrase the “four superpowers of technology.” What do you mean?
Gelsinger: The four superpowers are cloud, mobile, IoT and AI. Cloud is supercomputing power at unlimited scale. Mobile is unlimited reach. IoT is unlimited access. AI is unrivaled intelligence. It’s the cascading effect of those four things coming together at this point in time that’s ushering in the fastest piece of technical evolution that’s ever occurred in the history of mankind. I don’t see anything changing in it.
Q: AI both amazes and frightens people. Why is AI coming at us so fast?
Gelsinger: The foundations of AI have changed very little in 30 years. I first started thinking about AI in 1986, when I was the architect of Intel’s 80486. One day my marketing manager came running into my office, saying: “We have to make the 80486 a great AI chip.” At the time, you had companies like Thinking Machines, and you had the Smalltalk AI language, and LISP and Prolog were coming out. It was sort of the dawn of the AI era.
Q: But kind of a false dawn.
Gelsinger: Well, the extraordinary pace of change in AI today is because you now have data at scale and computing at scale. The core ideas of the AI algorithms haven’t changed all that much.
Q: George Gilder made that very point in a recent interview. AI is really about hardware.
Gelsinger: Right. I can take those same crappy algorithms, but I can now run them at such massive scale and against such extraordinary data sets, that I can start to produce interesting results. That’s what’s changed.
Q: Has AI’s effect on business ideas penetrated the minds of CEOs and boards outside of technology?
Gelsinger: I hope so. It’s existential for every business. The largest car company in the world, Uber, owns no cars; the largest hotel company, Airbnb, owns no hotels. The ability to harness digital assets at scale changes everything. A CEO and board need to look at their business and ask, “Okay, how do we use our assets, our industry position and these new superpowers for our business?” Good luck if they don’t do that. Someone else will knock them off.
Q: It takes a strong CEO to go to the board of directors and say, “We’re transforming our business. You’re not going to like the numbers in the short term.”
Gelsinger: What you say is, “We’re going through a three-year valley of death, but if we don’t do it, we’re going to get eaten alive. If the CEO goes to the board and the board doesn’t support him, then the CEO should think hard about resigning. Seriously.
Q: When I interviewed Michael Dell last May, he said cloud computing would co-exist with on-site technology at many companies for a long time, that companies will always want to hedge their bets with a hybrid strategy.
Gelsinger: Let me give you a different way to think about this. Throughout the history of technology, we’ve seen the pendulum swinging between centralized and decentralized, from mainframes to mini computers to client server. Cloud has been a force for centralization, and I believe that at this point people have become a little cloud crazy. They’re going to wake up and say, “Oh, crap. I’ve spent 20 years curating that application, and now I have to re-platform another cloud? And when I’ve finished re-platforming, what will I have? I’ll have the same 20-year-old app!” That’s a lot of investment just so somebody else can host it for me. But the thing is, with edge computing and IoT, I believe the pendulum will swing back in the other direction.
Q: Why?
Gelsinger: Think about it. Does my surveillance system need to send every image of the cat to the cloud? No. I want only the interesting images that look like a person I don’t think I’ve seen before. Also, am I going to take the 300-millisecond round trip to the cloud before I stop my robotic arm from hitting one of my workers? No, I’m not going to take that to the cloud. And a lot of work environments are going to be more remote, more hostile environments. I’m not going to take all my mining data from a remote location in Canada to the cloud when I have weak telephony for those environments. IoT and edge will push us away from massive cloud centralization. And the speed of light is still the speed of light, even in the four superpower era. So I can’t fix cloud-latency problems.
Q: Switching subjects, what’s it like being a part of the Dell constellation?
Gelsinger: Well, it’s purposely a loosely coupled model. I run VMware. I report to the board, and Michael Dell happens to be the chairman of my board. So at that level, I’m running my company, just as Rob Mee is running Pivotal Software. But we come together on a periodic basis to coordinate, to work certain issues and to agree strategically. That’s why Michael calls them strategically-aligned businesses. If I agree that this is the direction we’re headed, we might disagree on lots of stuff today but we’re working toward a common view of the future. And from my perspective, that’s working pretty well. We’ve set out some synergy objectives for Dell to sell more of VMware’s products and for us to build more solutions together, and we’re ahead of schedule in getting that done. So I view this as working.
Q: When I interviewed Michael in May, I got the overwhelming sense of how happy he was not to be the CEO of a public company.
Gelsinger: (Laughs) He’s happy. And to some degree, he’s created a private-public company with its tracking stock (DVMT). DVMT reports its financials, but, essentially, the street has no particular voting rights or any of those other kinds of concerns.
Q: What percentage of VMware does Dell Technologies own?
Gelsinger: Dell Technologies has a 90%-plus voting control of VMware, but it has a minority of about 19% of class A shareholders, and then the tracking stock, which is about 50% coverage of the VMware shares that Dell holds. So overall, Dell owns 32% of VMware.
Q: Before you became CEO of VMware, you spent 30 years at Intel. What did you learn from Intel founders Bob Noyce, Gordon Moore and Andy Grove?
Gelsinger: Well, I knew them all. But I considered Gordon to be the humblest billionaire who ever existed. One day my wife and I were at Macy’s, and I looked up and there was Betty Moore looking through the $1 bin.
Q: The other two, Noyce and Grove, fought a lot. How did they put that aside?
Gelsinger: Andy was a mentor of mine for 30 years. Meeting with him was like going to the dentist—without getting Novocain. He was brutal, tough and hard. And he enjoyed the debate. If you were to watch Andy and Bob argue, you’d think, “Oh my gosh, they’re going to kill each other; they hate each other.” But Andy loved the debate. If you couldn’t defend your point of view, you’d better get the blank out of there. Andy wanted somebody in the room with a point of view, the data to back it and the willingness to defend that point of view.
