Swim.ai, a startup developing advanced analytics software that can process data in devices at the edge of a network—such as industrial turbines or self-driving cars—announced today it raised $10 million in a Series B funding round.
The San Jose, CA-based company is part of the emerging movement toward “edge computing,” or adapting devices to handle a share of the computational burden of a computer network. Edge computing tackles problems such as the need for instant calculations based on the street view data collected by cameras in autonomous vehicles, so they can avoid a crash. Sending the raw data via the Internet to Web-based computers for analysis could be fatally slow.
Swim.ai is one of the companies figuring out ways to use artificial intelligence software to deliver high-order insights from data at the point where it’s collected—for example, in driverless cars or in ships at sea that might lose their Internet connection.
In addition to safety benefits, edge computing can also relieve bandwidth bottlenecks and reduce costs for moving masses of data to Web-based processors. Swim.ai has some hefty competitors in the edge software arena, including tech giants General Electric (NYSE: [[ticker:GE]]) and SAP.
Swim.ai’s Series B fundraising round was led by Cambridge, U.K.-based Cambridge Innovation Capital, joined by earlier investors Silver Creek Ventures and Harris Barton Asset Management. The round also included a strategic investment from Arm, a leading chipmaker that provides software to manage Internet of Things devices. Arm, with a Cambridge, U.K. headquarters and a North American headquarters in San Jose, is a wholly owned subsidiary of SoftBank Group. The new capital brings Swim.ai’s fundraising total to $17 million.
Chipmakers are joining software developers to work out the kinks that can stand in the way of edge computing. Artificial intelligence software usually demands powerful semiconductors such as Nvidia’s (NASDAQ: [[ticker:NVDA]]) Graphics Processing Units (GPU’s) to carry out its sophisticated computations. But chipmakers have been racing each other to design small, specialized A.I. chips that can fit into edge devices such as smartphones.
As a software developer, Swim.ai comes at the problem from a different direction, by designing its Swim EDX software to work on existing edge devices not necessarily fitted out with next-generation semiconductors.
As a company spokesperson explained in an e-mail exchange with Xconomy, Swim.ai’s software divvies up the various tasks involved in a data analysis and assigns them to the hardware in the customer’s network that can best carry out each particular chore. Some tasks are simple enough to be handled by a rudimentary computer called the Raspberry Pi Zero, which was launched in 2015 at a price of $5 by the Raspberry Pi Foundation, a U.K.-based charity that aims to make tinkering with computers accessible to anyone. (The foundation also gave away free “Pi Zeros,” which were about the size of a mini-Hershey bar, by attaching them to the covers of a tech magazine’s December 2015 issue. Versions of Pi Zero are still for sale for $5, or a bit more, from various distributors.)
More complex calculations that are also part of a Swim.ai data analysis, however, might need to be shipped to a Web-based computer with a GPU, the company spokesperson says.
Swim.ai plans to use its new capital to open an artificial intelligence R&D center in Cambridge, UK; to speed up its product development; and to expand its marketing efforts. Cambridge Innovation Capital’s investment director Andrew Williamson and Arm’s principal data scientist Damon Civin will join Swim.ai’s board of directors.
Photo Credit: Depositphotos
