Big-data cost optimization startup Yuki Technologies Ltd. wants to help enterprises keep a lid on the cloud infrastructure bills associated with their artificial intelligence projects after raising $6 million in funding today.

The seed round, which was led by Hyperwise Ventures and saw participation from VelocitX, Tal Ventures, Fresh.fund and Yakir Daniel, the co-founder of Spot.io, coincides with the official launch of Yuki’s platform.

Yuki says it wants to help companies manage the spiraling data costs associated with AI workloads, which eat up vast amounts of information. It’s no secret that AI can be extremely expensive, and most people are aware of the insane computational costs associated with high-end chips like Nvidia Corp.’s graphics processing units. However, far fewer people realize just how pricey the associated data infrastructure can be.

One of the challenges, according to Yuki, is that enterprises tend to stick with the same rigid data infrastructure for their AI projects as they use with traditional workloads. But this one-size-fits-all model forces teams that have different budgets and performance requirements to use identical compute resources, leading to massive cost inefficiencies that compound as their AI projects scale and data volumes expand.

Yuki Chief Executive Ido Arieli Noga told SiliconANGLE his company aims to fix this by applying intelligent automation and real-time optimization to businesses’ data infrastructures. At the core of its enterprise platform is Yuki Fabric, a specialized AI optimization model that acts as a unified control and automation layer for data platforms including Snowflake, Google BigQuery and other Apache Iceberg-based data lakes.

Its job is to govern how AI workloads consume storage and compute resources. It does this by monitoring them in real time to learn about their behavior, service-level agreements and cost-performance tradeoffs, in order to strike the right balance between performance and cost.

“Data is the only resource in an organization that no one truly manages,” Noga said. “We know how to store it, but not how to govern it. There are budgets, cloud infrastructure and teams, but the data itself has no control system.”

Yuki changes that. As a data control layer, it sits above the data platform, preventing infrastructure and process duplication in order to optimize compute and storage and reduce the operational costs of these resources. It can be deployed quickly, with no need for any code changes to the underlying infrastructure or workloads, and it has the ability to distinguish between business-critical tasks and lower-priority internal processes. Based on this assessment, it routes each query to the most efficient compute resources available, enabling companies to reduce their data costs even as their data volumes grow.

“Yuki was built to become the control layer that makes data infrastructure workload-aware and governed in real time, and in the age of AI this becomes even more critical,” Noga added. “Teams are constantly running experiments, models and new workloads, and many organizations are discovering they’re paying enormous sums for data and compute resources that no one is using anymore, simply because no one took the time to manage and clean them up.”

The startup is so confident in its ability to help customers reduce their data infrastructure costs that it won’t charge them if it fails to do so. It has created a novel business model that’s based on customer success, charging a percentage of whatever savings it helps businesses to generate. So, if it fails to save them any money, customers won’t have to pay anything for it.

Nonetheless, Noga believes that most customers will be only too happy to pay for its services, because he reckons that it saves them an average of around 42% on their data costs. For large enterprises, such savings can add up to millions of dollars per month. Despite only just launching, Yuki already serves a number of high-profile businesses, including the edge cloud infrastructure firm Qwilt Inc., the cybersecurity provider Tenable Holdings Inc. and the independent film media company Angle Studios Inc.

Looking forward, Yuki plans to expand its 15-person team and its Israeli research and development center in order to enhance its product’s capabilities and enable support for more data platforms.

Angel investor Daniel said he was quick to recognize the pain point Yuki is trying to solve, thanks to his history at Spot.io, the cloud cost control startup that was acquired by NetApp Inc. before being sold again to Flexera Software LLC. He believes it’s launching at just the right time, as enterprises look to scale up their AI initiatives. “It’s building the control plane for data cost optimization just as AI is turning data spend into a board-level issue,” Daniel said.

Photo: Yuki Technologies