Cloudian Spins Off AI-focused Tech With $9M From NTT Docomo, Others
Joseph F. Kovar
Edgematrix was started a couple years ago when two non-AI engineers were assigned to develop AI solutions to help consume data stored by Cloudian, and now Cloudian is ready to make that technology available as an AI platform on which others can develop.
Cloud-based object storage technology developer Cloudian is taking advantage of some artificial intelligence capabilities it has developed over time to set up a new company, Edgematrix, with the help of a number of Japanese strategic investors.
Cloudian Tuesday said Edgematrix raised Series A funding of $9 million from three strategic investors: mobile device and services developer NTT Docomo, engineering and construction company Shimizu Corp., and financial powerhouse Japan Post Capital.
In addition, Cloudian has transferred intellectual property, business, and revenue to Edgematrix, giving it a majority stake in the new company, said Michael Tso, CEO and co-founder of San Mateo, Calif.-based Cloudian and member of Edgematrix's board of directors.
Cloudian's technology works often in connection with big data users who have multiple petabytes of data in hybrid clouds, and so the company two years ago assigned a couple of engineers to develop AI-based solutions to help consume some of that data, Tso told CRN.
"We built a couple of solutions, such as roadside digital billboards with cameras that can recognize an oncoming car within 100 milliseconds and bring up an advertisement targeting the driver of the car based on demographics," he said. "If an advertiser knows there's a specific target, they'll pay 10 times more for the ad. And they built a solution that can identify machine parts in the factory to route them to the right place."
Cloudian found actual customers for those solutions, and then realized that the more data a company has, the better the solution, Tso said.
"And think, if our two engineers without strong AI backgrounds can build solutions people are willing to buy, we must be on to something," he said.
However, the application is maybe only 20 percent of a total AI solution, with the rest coming from ruggedized computers, cameras, and related hardware and components taking the rest, Tso said. For instance, the digital billboard solution requires a computer which works rain or shine and which might need support at 3 a.m., he said.
"So we felt that instead of deploying solutions, we should focus on the AI platform and let others make the solutions," he said. "So we shifted to a platform and edge network platform, which is where real-time decisions are made."
Edgematrix's AI offering includes software and a ruggedized hardware compute device with Nvidia technology built in, Tso said. It was built with help from NTT Docomo, which in Japan has the posts on which the devices will be deployed as that company builds out its 5G network.
Cloudian decided to make Edgematrix a separate company because it was the best way to enter the market, Tso said.
"The AI world is now looking at global infrastructures," he said. "But it's really hard. If we can make the infrastructure easier, it will make AI deployment easier."
Cloudian, as the majority owner, is happy to support Edgematrix's growth, Tso said.
"But if the technology needs different partners and channels, it's better if it's separate," he said. "Selling AI to telcos id different from selling cloud storage to enterprises."
Edgematrix is starting first in Japan where the level of interest in such AI platforms is high, in addition to the fact that 12 companies in that country are already working with the company's technology, Tso said.
"Japan is a very willing early adopter of AI in part because of an aging population," he said. "In the next 12 months, if we can deploy more with Docomo and Shimizu, and get successful case studies and learnings, then we can look to North America and Europe. But for now, there's a lot of growing to do first."
About 50 percent of Cloudian's overall revenue currently comes from the U.S., he said.