Predictive analytics startup Uptake is acquiring a company that controls a huge trove of industrial data, a move that positions the tech unicorn to better compete in the race to digitize power plants, oil wells and factory floors.
Uptake, a Chicago firm started by two Groupon co-founders that boasts a $2.3 billion valuation, has agreed to buy Albuquerque-based Asset Performance Technologies, or APT, for an undisclosed amount. The deal will marry Uptake's predictive analytics software and machine learning capability with the world's largest database on why industrial equipment fails.
Both companies focus on serving a few sectors that operate expensive assets like factories, construction equipment and power plants. The deal illustrates how some firms like Uptake aim to set themselves apart in a crowded field of big data crunchers by narrowing their focus.
Uptake, a CNBC Disruptor, launched its service with construction giant Caterpillar and has since serviced other companies like the energy unit of Berkshire Hathaway. It deploys self-learning technology to help companies make sense of a dizzying amount of data gathered from sensors embedded on industrial equipment around the world.
The APT database, Asset Strategy Library, was developed over 20 years and helps clients prevent costly downtime and repairs by identifying when equipment is about to fail.
Uptake's machine learning technology essentially gets smarter as it consumes more data, and APT's library amounts to tens of thousands of history lessons on why equipment fails. By pairing the two, Uptake believes it will be able to spot potential problems sooner and with greater confidence, Uptake President Ganesh Bell said.
"The data is the real asset here, as well as the people that have built this curated database," he told CNBC.
Uptake will retain APT's staff in Albuquerque and plans to build out the team to expand the database.
Founded to help nuclear power plants run better, APT has culled information on nearly 800 pieces of equipment across power generation, mining, steel-making and other industries. It has catalogued some 55,000 conditions that can identify when equipment is about to fail.
"That's 55,000 ways the world will stop working," said Bell.
Some big data companies that work across many sectors struggle to deal with the size and complexity of information collected from these industrial customers, said Mike Guilfoyle, director of research at ARC Advisory Group.
"It's pretty simple conceptually. Without error codes, which is a lot of what APT can supply, you can't really do the predictive maintenance that people want to do," he told CNBC.
Acquiring APT potentially makes Uptake more trustworthy among plant managers who want partners with deep knowledge about their operations, according to Guilfoyle.
"That's big. That to me is an absolute step in the right direction if you're getting into industrial operations," he said.
"That is an indication that they are very serious about getting trust from industrial operators."
APT's library is currently used by the likes of General Electric, which sells predictive analytics services to buyers of its equipment. Uptake and APT will continue offering their own predictive analytics platforms to their respective clients but intend to integrate the two.