Yuki secures $6m seed to cut data costs across AI and analytics workloads

Yuki Inc., a New York-headquartered data infrastructure startup, has raised $6 million in seed funding to build what it calls the AI-native control layer for modern data environments. The round was led by Hyperwise Ventures, with participation from VelocitX, Tal Ventures, Fresh.fund, and Yakir Daniel, the co-founder of Spot.io, now part of Flexera following its acquisition by NetApp. The funding will support Yuki’s research and development expansion in Israel, deepen product capabilities across platforms such as Snowflake, Google BigQuery, and Iceberg, and accelerate its sales push into the United States.

With AI workloads placing increasingly dynamic demands on compute resources, Yuki aims to reduce costs and inefficiencies associated with today’s one-size-fits-all data infrastructure models. The company’s platform, called Yuki Fabric, offers real-time optimization across hybrid and cloud-native environments by treating infrastructure governance as an intelligent automation layer rather than a manual or policy-based function.

Why enterprise data infrastructure is breaking under the weight of AI workloads and static provisioning models

Yuki is entering the market with a provocative thesis: most organizations have no real control system for data itself. While storage, compute, and security frameworks have matured over the past decade, the governance of how workloads consume infrastructure resources remains largely unmanaged. Enterprises often rely on the same underlying compute for workloads with vastly different service level agreements, urgency, and business value, leading to overprovisioning, cost sprawl, and low utilization.

These inefficiencies are compounded by the rising footprint of AI. Large language model inference, fine-tuning pipelines, and AI-driven analytics introduce volatile, non-linear compute patterns that traditional infrastructure models were never designed to handle. As engineering and data science teams continue to experiment and scale models, finance leaders are seeing escalating data and compute costs with little visibility into return on investment.

Yuki aims to fill this gap by providing a dynamic control layer that continuously learns system behavior, optimizes query execution in real time, and routes workloads to the most cost-effective compute resource available based on current priorities and constraints. This approach transforms data cost management from a retrospective budgeting exercise into a proactive, intelligent system embedded in infrastructure operations.

How Yuki Fabric works as a metadata-only, real-time AI optimization layer for data workloads

At the heart of Yuki’s platform is Yuki Fabric, an AI model that operates at the control plane level across a company’s data infrastructure. The system interfaces directly with platforms like Snowflake, BigQuery, and Iceberg-based data lakes, but does not require changes to code or query logic. Instead, it observes query patterns, workload behavior, and SLA requirements using metadata and telemetry, then orchestrates execution decisions in real time to enforce performance and cost objectives.

Yuki distinguishes between business-critical operations and lower-priority background tasks, allocating compute accordingly to reduce waste and ensure high-value tasks are served first. According to the company, this allows organizations to reduce infrastructure and operational costs while maintaining service guarantees across teams and products. The system also prevents duplication of infrastructure by centralizing governance and eliminating the need for siloed compute instances per team or workload type.

The company claims that customers using the platform in 2025 saved an average of 42.6 percent on data infrastructure costs. In large-scale environments, such as those operated by cybersecurity vendors or media platforms with real-time analytics and streaming workloads, these savings can amount to several million dollars annually.

Unlike cloud cost dashboards that offer reporting and analytics, Yuki operates as an executional control system that influences behavior at runtime. This positions it in a distinct category from traditional cloud spend management tools, aligning it more closely with observability-driven platforms that emphasize automation and intelligence.

Why Yuki’s timing may be ideal as enterprises adopt Iceberg and decoupled compute-storage architectures

The platform’s early support for Apache Iceberg-based architectures gives it a strategic advantage as enterprises increasingly move toward decoupled storage and compute paradigms. Iceberg allows organizations to scale storage independently from compute and is rapidly becoming the default format for data lake analytics. However, this architectural flexibility creates a new challenge: without a unified control layer, teams can inadvertently overspend on compute while underutilizing storage or triggering unnecessary query reruns.

Yuki’s founders, Ido Arieli Noga and Amir Peres, argue that this fragmentation creates blind spots that traditional IT governance tools cannot address. Their background in the data space, reinforced by prior joint ventures and a shared frustration with resource sprawl, shaped the decision to focus on building a native intelligence layer that governs infrastructure usage based on real-time behavior.

As companies migrate toward modern data lakehouses and hybrid data platforms, the need for context-aware optimization becomes more pressing. Data-heavy organizations that frequently adjust workflows, launch new products, or respond to changing customer usage patterns are especially vulnerable to cost spikes and performance bottlenecks.

By sitting above the vendor layer and operating through a metadata-only model, Yuki positions itself as an enabler of governance and observability without introducing latency or security risks. The platform requires no access to raw data, which lowers the compliance burden in sensitive environments such as finance, healthcare, or defense.

How Yuki plans to scale its go-to-market in a crowded cloud management landscape

The cloud infrastructure cost management space is estimated to be worth approximately $9.8 billion, with more than 200 vendors competing across various subcategories including billing analytics, forecasting, budgeting, and cloud-native cost visibility. However, most existing players focus on historical reporting or static alerts rather than on execution-level optimization.

Yuki’s differentiation lies in real-time orchestration that adapts to live workload behavior. While tools like Flexera (which acquired Spot.io), VMware’s CloudHealth, and Apptio have established strong enterprise footholds, they do not operate as workload-aware control planes. Yuki’s challenge will be to prove that its automation-first, savings-linked model can generate consistent returns across heterogeneous environments and that it can scale into enterprise-grade requirements without introducing operational risk.

The startup’s business model charges customers only if savings are realized, based on a percentage of cost reductions achieved. This success-based pricing approach is designed to align incentives and reduce upfront commitment barriers for buyers who may be skeptical of another layer in their data stack.

Initial adoption from companies such as Tenable, a cybersecurity firm, and Angel Studios, a media company, suggests that Yuki’s use case resonates with organizations managing high-volume, variable-demand environments. These are typically the kinds of companies where cost volatility has direct consequences on margins and where experimentation with AI workloads has surged in the past two years.

What this signals for enterprise IT teams under pressure to enforce cost governance in the AI era

Yuki’s emergence from stealth is emblematic of a broader shift in how enterprises view infrastructure governance. As AI transitions from research to production, the economic consequences of infrastructure inefficiency are becoming harder to ignore. The proliferation of GPU-intensive workloads, experimental pipelines, and multiple team environments requires a system of accountability that can operate without slowing down innovation.

Yuki’s narrative—that data is the last major organizational resource without a dedicated control system—is increasingly hard to refute. In many large companies, financial discipline is reactive and fragmented, with individual teams optimizing locally without a global view. This can lead to duplicated systems, persistent zombie workloads, and infrastructure that grows faster than business value.

By offering a unified control layer that governs compute behavior based on real-time data, Yuki is positioning itself not just as a cost optimization tool, but as a foundational component of AI-era infrastructure design. Its long-term success will depend on its ability to integrate deeply into heterogeneous enterprise environments, deliver measurable savings, and maintain trust with both IT and finance stakeholders.

What Yuki’s $6 million seed round signals about the future of AI infrastructure control

• Yuki Inc. has raised $6 million to build an AI-native control layer for real-time data infrastructure optimization.

• The platform acts as a dynamic execution router across Snowflake, BigQuery, and Iceberg data lakes, targeting compute cost waste.

• Its value proposition is strongest in high-query, AI-intensive environments where static provisioning models are breaking down.

• Yuki customers saved an average of 42.6 percent on data infrastructure costs in 2025, based on early deployments.

• The company only charges a fee if measurable cost savings are achieved, aligning economic incentives with customer success.

• Yuki competes in the $9.8 billion cloud management tools sector but differentiates through real-time, metadata-only optimization.

• Early traction includes cybersecurity and media firms, indicating cross-sector relevance in AI-heavy operations.

• Funding will support R&D in Israel and commercial expansion in the U.S., with a focus on enterprise GTM execution.

• Yuki is betting that infrastructure governance will become AI-driven, moving beyond dashboards to automated control.

• If it succeeds, Yuki could define a new category of real-time AI cost governance platforms for hybrid cloud environments.