On Nov. 29, Jeff Immelt pulled out the really big iron. General Electric’s (GE) chief executive climbed up to take the stage at a modified film studio in San Francisco and stood next to a 6.87-ton jet engine built by his company. Inside this mass of twisted metal—Immelt told the spectators at the company’s Minds and Machines event—were 20 sensors that monitor the engine’s performance, generating part of the roughly 1 terabyte of information produced on a one-way, cross-country flight. In the years ahead, GE plans to analyze this information as it’s never been analyzed before in a quest to build smarter machines and more lucrative services that it can sell to customers.
The event capped a yearlong effort to convince people in Silicon Valley and elsewhere of the merits of the “Industrial Internet.” That’s the grand wrapper that Fairfield (Conn.)-based GE has put around the idea of a new breed of connected equipment. Be it jet engines, generators, locomotives, or CT scanners, GE wants to extract data from the hardware and enlist teams of its analysts to find ways to make the products operate more efficiently. There’s big money to be had in tweaking giant machines; as Immelt puts it, even a 1 percent improvement in the operations of commercial aircraft would translate into $2 billion less per year in fuel costs for GE’s customers in the airline industry.
Based on company estimates
GE’s ideas here aren’t exactly novel, says David Linthicum, the founder of technology consultancy Blue Mountain Labs. “These sorts of concepts have been bandied about for 20 years,” he says. Immelt himself concedes GE could be seen as late to the party. Still, the company’s vision of the Industrial Internet has some people in the technology industry salivating. “You’re talking about taking consumer Internet skills and data science and applying it to industries that have been unaffected by that progress,” says Hilary Mason, chief scientist at Bitly, a firm that analyzes how people share links on the Web, who attended the GE event. The big question for GE is whether it can fire up a data analysis operation from scratch and still compete against technology giants such as IBM (IBM) and Oracle (ORCL) and any number of hungry startups that have already spent years preparing to get rich off these machines.
GE and Immelt certainly want to look the part of a Big Data hipster. The company recently set up shop on the outskirts of Silicon Valley, gutting an office building to make it look like a startup’s digs and hiring about 300 data gurus. GE might hire as many as 1,000 data specialists in total and has set aside about $100 million to invest in promising startups. These measures are part of the prerequisites for letting Silicon Valley know that GE has come to play. “We are trying to be the partner of choice in a foreign land,” Immelt says. “In order to do that, you need to be present.”
Immelt wants GE to produce 20 to 30 applications per year that its $45 billion services group can use as leverage in negotiations. GE has 250,000 machines installed with customers, and Immelt thinks it is in the best position to add intelligence to those machines. Without its own software and big data smarts, GE risks being disintermediated from its own gear. “If you are into these products, you ought to be into the analytics around these products,” Immelt says. “I think people give that up at their own peril.”
It’s already working with Mount Sinai Medical Center to use sensors on beds and diagnostic gear to help patient scheduling: A computerized system monitors patients as they’re admitted, alerts staff to prepare the right kind of room, and calculates the patient load the hospital can expect at a given time. Mount Sinai has used the system to run at near 95 percent capacity, vs. the accepted industry goal of 85 percent. Other efforts are underway to automate the flow of trains to let them run faster and to manage how things like electric vehicles from Tesla Motors (TSLA) can recharge at night when energy is in lower demand. GE has also put up $500,000 in prize money for a contest among developers aimed at helping airline pilots. It provided flight data and asked the contestants to create an algorithm that best weighs fuel usage, speed, and air traffic against each other, giving pilots guidance on how to reroute when unexpected delays hit.