Treasure Data, the Intelligent Customer Data Platform (CDP) built for enterprise scale and powered by AI, unveiled a groundbreaking “No Compute” pricing model that completely decouples cost from processing resources. Enabled by Treasure Data’s Hybrid CDP architecture, the new model primarily charges for the number of unified customer profiles stored and the behavioral events associated with them, allowing brands to run unlimited queries, segmentations and activations without fear of runaway bills.
“Predictable costs should be a right, not a luxury,” said Kaz Ohta, co-founder and CEO of Treasure Data. “Our ‘No Compute’ pricing removes the biggest barrier to full CDP adoption – cost anxiety – so every team can activate data and AI agents with confidence.”
Solving the industry’s cost dilemma
Enterprises have long been forced to choose between two imperfect options:
- Packaged CDPs that primarily meter the number of profiles and compute.
- Composable-only CDPs that primarily meter the number of profiles and shift all processing to a cloud data warehouse (CDW), creating unpredictable query charges.
Also Read: Evertune Raises $15M Series A to Scale AI Marketing Platform
Treasure Data’s pricing ends this trade-off by separating economics from compute location. Customers enjoy transparent pricing based primarily on the number of real-time, resolved customer profiles managed and the volume of associated behavioral events (e.g., website visits, mobile app usage, email activities, etc.).
Whether workloads run in Treasure Data’s high-performance database engine or inside the customer’s CDW environments, your teams enjoy the same consistent experience – without having to think about infrastructure or cost complexities.
“The true value is no longer just in data storage and movement, but in the intelligence you apply to it,” said Ohta. “Treasure Data is infused with AI everywhere, and now our customers pay for the value they get from AI-driven outcomes like better personalization, predictive insights, and campaign optimization — not for the underlying compute cycles.”
SOURCE: Businesswire