Be part of our day by day and weekly newsletters for the newest updates and unique content material on industry-leading AI protection. Be taught Extra
As synthetic intelligence drives unprecedented demand for information processing, Mountain View startup Voltron Knowledge is providing an answer to one in every of AI’s least mentioned however most crucial challenges: shifting and reworking large datasets shortly sufficient to maintain up.
Voltron Knowledge, which introduced a strategic partnership with Accenture at the moment, has developed a GPU-accelerated analytics engine that might assist enterprises overcome the info preparation bottleneck hampering AI initiatives. The corporate’s core product, Theseus, permits organizations to course of petabyte-scale information utilizing graphics processing items (GPUs) as an alternative of conventional laptop processors (CPUs).
“Everyone’s focused on the flashy new stuff that you can touch and feel, but it’s that dataset foundation underneath that is going to be key,” mentioned Michael Abbott, who leads Accenture’s banking and capital markets follow, in an unique interview with VentureBeat. “To make AI work, you’ve got to move data around at a speed and pace you just never had to before.”
Constructing for the AI tsunami: Why conventional information processing gained’t lower it
The partnership comes as corporations speeding to undertake generative AI are discovering their current information infrastructure isn’t outfitted to deal with the quantity and velocity of knowledge required. This problem is anticipated to accentuate as AI brokers turn out to be extra prevalent in enterprise operations.
“Agents will probably write more SQL queries than humans did in a very short time horizon,” mentioned Rodrigo Aramburu, Voltron Knowledge’s CTO and cofounder. “If CIOs and CTOs are already saying they spend way too much on data analytics and cloud infrastructure, and the demand is about to step function higher, then we need a step function down in the cost of running those queries.”
Not like conventional database distributors which have retrofitted GPU assist onto current methods, Voltron Knowledge constructed its engine from the bottom up for GPU acceleration. “What most companies have done when they’ve tried to do GPU acceleration is they’ll shoehorn GPUs onto an existing system,” Aramburu advised VentureBeat. “By building from the ground up…we’re able to get 10x, 20x, 100x depending on the performance profile of a particular workload.”
From 1,400 servers to 14: Early adopters see dramatic outcomes
The corporate positions Theseus as complementary to established platforms like Snowflake and Databricks, leveraging the Apache Arrow framework for environment friendly information motion. “It’s really an accelerator to all those databases, rather than competition,” Abbott mentioned. “It’s still using the same SQL that was written to get the same answer, but get there a lot faster and quicker in a parallel fashion.”
Early adoption has centered on data-intensive industries like monetary companies, the place use instances embody fraud detection, danger modeling and regulatory compliance. One massive retailer decreased its server depend from 1,400 CPU machines to simply 14 GPU servers after implementing Theseus, based on Aramburu.
Since launching at Nvidia’s GTC convention final March, Voltron Knowledge has secured about 14 enterprise clients, together with two massive authorities companies. The corporate plans to launch a “test drive” model that can enable potential clients to experiment with GPU-accelerated queries on terabyte-scale datasets.
Turning the GPU scarcity into a possibility
The present GPU scarcity sparked by AI demand has been each difficult and useful for Voltron Knowledge. Whereas new deployments face {hardware} constraints, many enterprises possess underutilized GPU infrastructure initially bought for AI workloads, belongings that might be repurposed for information processing throughout idle durations.
“We actually saw it as a boon in that there’s just so many GPUs out in the market that previously weren’t there,” Aramburu famous, including that Theseus can run successfully on older GPU generations which may in any other case be deprecated.
The expertise might be notably precious for banks coping with what Abbott calls “trapped data” — info locked in codecs like PDFs and paperwork that might be precious for AI coaching however is troublesome to entry and course of at scale. “You’ve seen some of the data that Voltron would show you is potentially 90% more effective and efficient to move data using this technology than standard CPUs,” Abbott mentioned. “That’s the power.”
As enterprises grapple with the info calls for of AI, options that may speed up information processing and cut back infrastructure prices are prone to turn out to be more and more vital. The partnership with Accenture might assist Voltron Knowledge attain extra organizations dealing with these challenges, whereas giving Accenture’s purchasers entry to expertise that might considerably enhance their AI initiatives’ efficiency and effectivity.
