GE CEO Jeff Immelt On How The Industrial Internet Is Helping Slash Downtime

At the company’s Minds + Machines event in San Francisco today, GE unveiled a range of new Industrial Internet offerings.

GE CEO Jeff Immelt On How The Industrial Internet Is Helping Slash Downtime
[Photo: Chip Somodevilla/Getty Images]

In a video touting General Electric’s new initiative to get more than 20,000 people building for, the company’s new dedicated cloud environment for app developers, we see a young engineer telling his friends he’s going to work for GE. His friends are utterly flummoxed.


GE is without a doubt one of the world’s most important companies, working on everything from airplane engines to locomotives to health care and much, much more. But in Silicon Valley, at least, it’s hardly considered the kind of company that gets hot-shot engineers coming out of Stanford excited.

GE wants to change the perception that it’s not a serious player in the battle for engineering talent.

Today, the company is dead set on changing that perception. It’s also dead set on convincing its major industrial clients that it is the only player they should be dealing with when it comes to building the kinds of applications that will run giant factories and improve manufacturing efficiency.

At its annual Minds + Machines event in San Francisco today, GE unveiled a series of new initiatives. They included Predix, which it hopes will become the pre-eminent industrial app marketplace; Brilliant Factory, a new digital manufacturing technology that GE says can deliver 10% to 20% improvements in unplanned downtime; Digital Power Plant, new technology meant to securely digitize physical electricity systems; and GE’s embedding of cyber-security firewall technology from Wurldtech into thousands of gas, steam, wind, and aero-derivative turbines.


In advance of the announcements, GE CEO Jeff Immelt sat down with Fast Company to talk about the company’s four-year Industrial Internet initiatives, and how they will shape both GE itself and its many large customers in the future.

Fast Company: Today’s a big day for GE’s Industrial Internet initiatives. In your view, what’s the most important thing you’re announcing today?

Jeff Immelt: This is our fourth year working on the Industrial Internet. What we’re showing this year is that it’s real, that basically, everybody has their own thesis about the Industrial Internet, or the Internet of Things, but it’s about better productivity. We have users here this year from railroads and airlines. So this is a statement to say, “We’re here. We’re a first-mover, we’re generating benefits for customers” and so I think that’s the big statement. That’s the overarching thing. And technically, we’re going to be one of the players that’s both offering an operating system, in the case of Predix, and also applications, in an open setting. So you have a macroeconomic story: Here’s what the Industrial Internet means, but then you have a technical story that says GE’s going to be one of the players that’s driving both horizontal platforms and vertical applications.


How hard is it to get these legacy industries to move into this new era of the Industrial Internet?

In some ways it’s hard, and in some ways it’s easy. I think what’s hard is just things move slowly, right? If you walk into a CEO of any one of these companies, and say “You have to change your technology every three years,” you get thrown out on your ass. So you’ve got legacy that you have to deal with. What’s easy is that they can eat the savings. If you can say we can reduce your downtime by 50%, an industrial company says “That’s worth $50 million, I’m going to invest in that. I know what the payback is.”

In some ways it’s hard, because you always have to deal with legacy. And in some ways it’s easy because the economic benefit is transparent. I remember in 1984 being in a sales office at GE the first time we installed Windows, and nobody had any idea what that would mean. Was it productivity, was it communications? When you walk in with an industrial customer, you know exactly what you’re selling, and they know exactly they’re buying.


I assume you’re using a lot of what you’re selling inside GE?

We have our own digital thread that goes from engineering through service, and we’re our own show site in terms of what we’re doing. And any apps that we’re using inside GE, we’re willing to offer to any of our customers, so I think that’s one of the big thrusts of a company like GE. We’re our own biggest customer, we’re our own biggest show site. I talk about going from inside out. We take the things we’re doing internally and drive them externally.


Talk about the concept of the “Brilliant Factory,” which you announced today. You say you can achieve 10%-20% reductions in unplanned downtime. How do you do that?

In the case of uptime, you’ve got sensor-based models on all your manufacturing equipment that are able to predict peak performance, able to predict downtime. So all the same things we can do on a jet engine, we can do on a milling machine, on die casts, on investment castings, and things like that. The same technology works in our factories that works in our install base. So that’s a story of sensors and analytics around equipment and avoiding equipment failure, how you schedule the plant and things like that.

If you have a customer that says they want to adopt the Industrial Internet, how long does it take?


To do asset performance management, let’s say a locomotive that’s filled with sensors today, to give them a suite of things that’s going to improve their uptime, you’re talking weeks, you’re building on an existing system. You can port the data right away. They can build some of their own analytics. We’re basically talking weeks. Let’s say the same railroad wanted to do something more enterprise-wide, that would be called Movement Planner, or Computer Dispatch, that might take six months or longer. It’s kind of everywhere in between what a customer wants.

Are these decisions being made at the CEO level?

I think it’s the CIO. I think you’re seeing, in industrial companies, you’re seeing kind of a complete repositioning of what a CIO does. A CIO has gone from being an outsourcer, a cost center, to really being the intelligence center for the industrial companies. I know that’s happening inside GE. I’d say it’s CIO, CFO, chief maintenance officer, up to the CEO, that’s where the nexus of the decision will be made.


This is the Industrial Internet, not consumer. But how much of a model is the consumer Internet and consumer Internet of Things and companies like Google and Facebook?

I think everybody sees what’s happened on the consumer side, and says, “Gosh, the same levels of productivity could happen industrially.” The data’s harder to get. It’s more unstructured. So there’s some technical issues industrially. But whereas everybody can get on Facebook, Facebook’s one of those great horizontal applications, that’s not necessarily going to be the case (industrially because) health care’s going to look different than locomotives, and locomotives are going to look different than elevators. So you’re going to have a little bit more deep domain in industrial versus consumer. It makes it a combination of horizontal technology and deep domain technology. But one of the things I would say today, even with all the Internet usage, IT consumption by industry, the rate of productivity gains in industry is actually quite poor. So in the early days, I’d say consumer companies have benefitted a lot more than industrial companies have, by investments in the Internet.

Do you think that’s going to change?


I think that’s what the Industrial Internet has to be all about. Sometimes when I read magazines, they say, “is there a bubble in tech?” I think the work around the Industrial Internet is just beginning. So I think in many ways, the investment has to come here right now, in terms of where the future’s going to be.

How hard is it to develop this idea of an Industrial Internet outside of Silicon Valley?

You have to tentacles here, for sure. Because this is where the talent is, and this is where some of the ideas are. But at the end of the day, making a jet engine is harder, or as hard as writing code, doing the analytics. So the Industrial Internet is going to be a combination of the physical and the analytical. We bring a tremendous amount of credibility to that world. But you have to be able to interface with Silicon Valley, because this is where the talent is, and it’s talent we want to bring inside GE.


How much do you feel there’s resistance within GE to making this transformation from legacy technology to the Industrial Internet?

When we first started, four or five years ago, it was slower to come. I think now everybody sees this as one of the real advantages, or one of the future aspects of the company. So I think everybody’s on board inside the company. With all that being said, we have to infuse talent inside the company that’s brought up a different way, or ultimately, we’re not going to be successful.

The LINK YouTube video that you’ve got for the Predix announcement touches on this idea that GE’s not the most obvious company for developers, for engineering talent. This guy’s very excited that he’s going to work for GE, and his friends are like, Ooookay. Tell me about that dynamic?


I think it was meant to be proactive and tongue-and-cheek to say we want to get inside the head of people to say, “look, if you went to the University of California at Berkeley or Stanford and thought the only outlet for your technical skills was going to be a consumer Internet company or a payment systems company, this is a new outlet for your skills” and it’s very mission-based. We’re trying to bridge two worlds and say, this industrial world is actually quite relevant, and can be quite meaningful and a lot of fun. So that’s the idea, really.

Still, you’ve got companies like Google, Facebook, Twitter, Apple and the like saying you can change the world right now. Again, how do you compete with them?


I think modeling, big unstructured data, doing it in a way that’s going to drive new industries like health care. I think that’s a very sellable pitch to the kind of talent that we want to attract, and we’re seeing that already. We’re getting a ton of good people, a ton of good partners, and in the end, we have the gateway to the industrial world, which is hard to get. We have what is hard, if you want to do the Industrial Internet, it’s going to have to come through an industrial gateway, and we’re good at that. We have that.

You talk about having tentacles in Silicon Valley. Do you have people here?

We have a big location in the East Bay, and then we have a big venture office in Silicon Valley, and we have several collaborators that are in Silicon Valley. We’ll have 2,000 scientists and engineers in the San Francisco area, and then equity stakes in (a few) smaller businesses. But the value we bring, again, is this deep domain. It’s health care, aviation, transportation, energy, electricity, oil and gas. That in the end is the value that GE has, both the software piece, the analytics piece, but also the deep industrial piece, which is quite valuable.


Talk about how GE imagines drones in industry?

I’d say one of the most exciting things about drones right now is in the service industry, using drones to go through places like rail yards, where locomotives are parked. And you can do visualization of the wheels. One of the biggest issues for locomotives is actually cracked wheels. So you’re able to take a drone through a rail yard, and get preventative maintenance on early defects. So in these remote settings, whether it’s a utility, or a wind farm, or a rail yard, drones are going to be a seminal technology as it pertains to the Industrial Internet because of the time savings, and marrying it with visualization technology that’s as good or better as inspections. It allows you to get to places that are hard to get to in a factory or industrial setting, or something like a rail yard.

Is GE developing that technology in-house?

It’s a combination. At our Global Research Center, we’ve got work on drones. We’ve taken one or two equity stakes through our Ventures group, so we’re trying to co-develop with some of our partners. But along with the Industrial Internet is going to be a set of enabling technologies from sensors to robotics to drones to AI. This is all going to be developed alongside this software and analytics on top of the assets.

What about virtual reality?

We have 30,000 field engineers, so the ability to do both training and also activities around service and repairs in remote settings, things like Oculus or other virtual reality, can be quite helpful. They help you visualize a factory. It’s still nascent, but ones that we’re investing in and ones are going to be part of our service technology. The install base value of what we have is $1-2 trillion, jet engines and turbines and factories, so whatever we can do to bring more value to the people that are out with our customers, that’s going to have value, and I view VR as one of the technologies that’s going to be a part of that.

How do you imagine you’ll use robots in this ecosystem?

I come back to remote service settings. Any place where you have hard-to-get-to tasks that can be done in a repetitive way, robots can be quite beneficial. So again, we have a big installed base with valuable assets that need to be repaired. I think robotics can be helpful there as well.

Do you see robots as augmenting workforces?

I do. I think it’s fundamentally going to augment what goes on today. Industrial companies averaged 4% productivity from 1990 to 2010. From 2011 to 2015, it’s 1%. We have a tremendous productivity upside that’s out there, whether it comes from an investment in automation products, better training, better predictive maintenance, there’s a lot that we have to do as industrial companies to improve productivity before we worry about whether a factory has no employees. I view technology as a great thing, it creates more high-value jobs inside GE. So when you walk through one of our factories, the frontline people are also very facile with automation, and they get paid well.

About the author

Daniel Terdiman is a San Francisco-based technology journalist with nearly 20 years of experience. A veteran of CNET and VentureBeat, Daniel has also written for Wired, The New York Times, Time, and many other publications