Seemoredata Launches Databricks Cost Optimization to Help Enterprises Fund AI

Tel Aviv startup extends its data cloud cost platform to Databricks as enterprises look for ways to fund a projected $2.5 trillion in AI spending

Seemore Data, the context-aware control plane for data and AI infrastructure, today announced early access to Seemore for Databricks. The new offering gives data platform and FinOps teams a single view of Databricks spend across SQL warehouses, Lakeflow jobs, all-purpose clusters and AI services, and automatically optimizes SQL warehouses within guardrails customers set.

Every dollar moving into AI has to come from somewhere, and for most companies the biggest line nobody fully controls is the data cloud”

— Yaniv Leven, CEO, Seemore Data

The launch comes as enterprises look for ways to pay for AI. Gartner forecasts worldwide AI spending will reach $2.5 trillion in 2026, and 45% of companies say their AI budgets are coming out of existing software budgets, according to IDC. For many organizations, the data and AI workloads driving that spend now run on Databricks.

Marketing Technology News: MarTech Interview with Jema Birkmeyer, VP @ Attentive

Databricks spend is hard to see and harder to control. The same workload can run on several compute types with different list prices, from $0.15 per DBU for data engineering compute to $0.40 for interactive workloads. Compute costs land on one bill and cloud infrastructure on another. And the teams running these workloads are measured on reliability, not cost, so overprovisioning is often the safe default.

“Every dollar moving into AI has to come from somewhere, and for most companies the biggest line nobody fully controls is the data cloud,” said Yaniv Leven, CEO of Seemore Data. “Unlike the application layer, the data cloud can’t be optimized without context. Every query has a structure. To right-size its compute, you need to know which tables it reads, where it writes and who relies on the results downstream. A data asset means little until you understand where it’s used, how it’s used and how that usage changes over time. Capturing that context is what Seemore was built to do, first on Snowflake and now on Databricks.”

Seemore for Databricks includes:

– Full cost attribution. Spend is read from Databricks system tables and priced with the customer’s own rates, then rolled up by workspace, job, warehouse, cluster and SKU.
– One view across every workload. SQL warehouses, Lakeflow jobs and pipelines, all-purpose compute, and AI services such as Model Serving and AI gateway.
– SmartPulse for SQL warehouses. The engine Seemore customers use to tune Snowflake warehouses now adjusts Databricks SQL auto-stop and scaling in real time, inside customer-defined guardrails, with every change logged.
– Idle-time optimization for serverless. Auto-stop is tuned to each warehouse’s actual query pattern, cutting paid idle time without adding noticeable cold starts.
– Right-sizing for classic compute (closed beta). Predicts the right cluster size for each workload before it runs, reducing both waste and the SLO risk of under-sizing.

Marketing Technology News: How MarTech Is Enabling Autonomous Brand Engagement Across Channels?

Seemore connects with metadata-only access. It reads Databricks system tables through a customer-created service principal and never accesses customer catalogs, schemas or table data. Write access is optional and limited to starting, stopping and resizing approved warehouses.

On Snowflake, Seemore customers see an average 33% bill reduction and 6.1x ROI with no code changes. Seemore expects Databricks customers to see comparable results, as the same waste patterns appear on both platforms.