This Hot Database Startup Just Snagged $50 Million in Funding

June 29, 2017, 11:30 AM UTC

Kinetica, a startup that builds databases that harness fast graphical processing chips to perform everyday business applications, now has $50 million in new funding.

The San Francisco company is banking that increasing numbers of mainstream computing jobs will run on Nvidia (NVDA) graphical processing units (GPUs), which until recently were used primarily for video games and animation. Most business data-crunching jobs have thus far relied on more general-purpose Central Processing Units (CPUs), typically Intel (INTC) X86 chips.

“GPUs normally had been all about video gaming and shading pixels on TVs and monitors,” Amit Vij, co-founder and CEO of Kinetica told Fortune. That started to change when Nvidia, the leading provider of GPUs, released its Cuda language, which made the processing power of these chips more broadly applicable, he added.

Now a handful of startups including Kinetica, Sqream, and MapD are building databases that use GPUs to parse all sorts of data, including the sort of information normally processed by traditional relational databases from Oracle (ORCL) and Microsoft (MSFT), but also messy, free-form data thrown off by sensors in the field and social media streams.

Related: This Startup is Taking Nvidia Chips Mainstream

Jonathan DeMent, program director for IBM Power Systems & OpenPower Innovation group is a fan. Some large IBM (IBM) customers are using IBM Power servers, which pair that company’s Power processor with an Nvidia GPU, for large complex supply chain tasks.

One unnamed retailer, for example, uses artificial intelligence (AI) software and Kinetica running on Power servers to take in data from trucks, the packages inside the trucks, weather information, and personnel information to adjust deliveries based on real-time conditions.

Related: How Machine Learning Fueled Nvidia’s Stock Jump

“If it’s pouring rain in Atlanta and there are not enough umbrellas in the stores, there’s a disruption. The goal is to have a real-time view of all that’s happening,” DeMent said, adding that the system can be queried quickly and process changes fast to meet changing demands.

Kinetica grew out of work that Vij and his Kinetica co-founder Nima Negahban did at the United States Army Intelligence and Security Command and the National Security Agency creating a system to track and capture terrorists in real-time. The company claims some large customers, including the U.S. Postal Service—perhaps the largest logistics company in the world—which has used Kinetica since 2014 in a system that optimizes routes and tracks packages.

Get Data Sheet, Fortune’s technology newsletter.

The $50 million in Series A funding was led by Canvas Ventures and Meritech Capital Partners, with contributions from new investor Citi Ventures and existing backer Ray Lane of GreatPoint Ventures. That brings total funding, including seed money from Lane, to $63 million. Paul Madera, partner at Meritech Capital Partners and Gary Little, partner at Canvas Ventures will join Lane, a former president of Oracle, on Kinetica’s board.

Read More

Artificial IntelligenceCryptocurrencyMetaverseCybersecurityTech Forward