Tecton.ai raises $35M Series B co-led by a16z and Sequoia Capital as it releases its ML feature store, just a week after AWS debuted its Sagemaker Feature store
Context & Ripple Effects
Tecton's $35M Series B, co-led by a16z and Sequoia, lands one week after AWS debuted SageMaker Feature Store — the timing frames the raise as a bet that feature management is a product category worth owning before the hyperscalers bundle it away. The round coincides with Tecton's general release of its ML feature store, converting the company from a tooling vendor into a platform competitor against AWS's in-house offering.
The corpus shows the arc this round set up: Tecton followed with a $100M Series C led by Kleiner Perkins in 2022, and by 2025 Databricks moved to acquire the company outright — a trajectory that starts with this Series B and its head-to-head positioning against AWS.
First-order effects
- Tecton now competes directly with AWS's SageMaker Feature Store, and the $35M funds a general release aimed at enterprises deciding whether to buy feature infrastructure from a hyperscaler or an independent.
- a16z and Sequoia's co-leadership signals top-tier VC conviction that the feature store layer — the data plumbing between raw data and deployed ML models — is a defensible standalone market rather than a cloud checkbox.
Second-order effects
- AWS's bundling of a feature store into SageMaker forces independent ML-data vendors to differentiate on cross-cloud portability and real-time performance, since the default cloud-native option is now free-adjacent for AWS customers.
- The raise validates the ML data-tooling category for other investors: a year later Tonic.ai raised a comparable $35M Series B for synthetic data, showing capital flowing to adjacent layers of the ML data stack.
Third-order effects
- The 2025 Databricks acquisition confirms the structural endpoint this round pointed toward: standalone ML data infrastructure vendors get absorbed by broader data platforms, leaving enterprises to choose between hyperscaler bundles and consolidated data-platform suites.
- If the pattern holds, the ML stack consolidates the way the analytics stack did — independent layers survive only as acquisition targets, and platform distribution (AWS, Databricks) beats best-of-breed point tools.
The trend: ML data infrastructure is consolidating from independent feature-store vendors into the portfolios of hyperscalers and data platforms, with Tecton's 2020-2025 arc from Series B to Databricks acquisition as the template.