GPU databases are hot

Kinetica is our second meeting with The IT Press Tour in that rapid growth data segment following MapD, now OmniSci, a few months ago. It started around 2010 at GISFederal with the first ever GPU database named GPUdb designed by Nima Negahban and Amit Vij to help US Army track terrorist threats. Currently GISFederal web site redirects you to Kinetica site. Then USPS picked GPUdb to boost routes optimization and increase accuracy. We can called this period the research and proof of concept phase with the first adoption and validation.

The second phase, more commercial of course and worldwide, is associated to Kinetica. The structure was born early 2016 with seed founding from Ray Lane, former Oracle executives. One year later, the company raised a series A of $50M and Paul Appleby joined as CEO to immediately put senior DNA at the exec level. The entity was incorporated for the long run and has already more than 100 employees. Since 2018, market adoption has accelerated in many verticals.

Kinetica argues, we all agree, that data is the new oil fueling the business. Good examples are smart applications, IoT, social media, medical, AI driven insights... So having the right engine to compute a vast amount of data is critical. Key fundamentals elements are:

• GPU to process things in parallel,
• In-memory to boost response time,
• distributed to offer scalability,
• streaming, ingest and on-the-fly analytics,
• SQL-style query capability to facilitate adoption and applications integration with finance, retail, healthcare, telecom or energy vertical use cases.

The company is ready to address 2019 challenges, among them real-time use cases, IoT demanding environments and AI as a key business differentiator.