October 2026 Summaries

Fivetran announced lower pricing for high-volume Activations usage alongside product updates intended to make reverse ETL workflows easier to scale and manage. Activations moves governed data from warehouses or data lakes into operational tools such as Salesforce, HubSpot, Braze, Google Ads, and LinkedIn, helping organizations reduce reliance on manual exports and fragmented data pipelines while preserving data-team control over definitions, governance, and lineage. New capabilities include source-based activation management, deployment across Fivetran Cloud Regions with configurable storage, reuse of PrivateLink connections, unified permissions, and redesigned APIs and webhooks for automation. The company positions these changes as a way to support fresher audiences, personalization, lead scoring, forecasting, customer context, and AI-related workflows, citing customer examples including CaliberMind’s reported $400,000 annual engineering-capacity savings and Canva’s reported savings of more than $200,000 annually, while ClickUp reportedly reduced customer acquisition cost by 50%.

Fivetran describes how it centralized fragmented AI usage and spending data from providers such as OpenAI and Anthropic to help leadership assess costs, adoption, productivity, and return on investment. Previously, finance and analytics teams relied on provider portals, CSV exports, and manual processes to map user-level consumption to business cost centers, a task that could take at least a day per reporting cycle. Using Fivetran’s AI connectors, combined with HR, employee, financial, engineering, and operational data, the company created a unified cost model and near-real-time dashboard that can analyze usage by vendor, team, user, model, and activity type. Fivetran also released Quickstart models and an AI Reporting package to help customers turn raw provider data into analysis-ready metrics covering cost, user activity, coding productivity, and per-user summaries. The company says centralized visibility supports budgeting, vendor negotiations, employee AI allowances through its internal AI wallet, and efforts to connect AI consumption with outcomes from systems such as GitHub, Jira, CRM, and support platforms, framing AI management as an ongoing business investment rather than a standalone expense.

Fivetran CEO George Fraser describes benchmarking TPC-H analytical queries on an iPhone 17 Pro running DuckDB against Databricks Serverless SQL clusters at 25 GB, 50 GB, 100 GB, and 200 GB scales, chosen to represent the upper range of many practical business analytics workloads. The iPhone outperformed the Databricks clusters at all but the largest scale, where it remained competitive, despite having fewer CPU cores and less memory. Fraser attributes the result to DuckDB’s low overhead for relatively small queries and to the costs that distributed systems incur for planning, compilation, and data shuffling during joins and aggregations. He argues that many organizations may be paying substantial premiums for distributed cloud data platforms when individual queries could run efficiently on single-node engines such as DuckDB or Polars. The post proposes a lakehouse architecture centered on open storage formats such as Iceberg, with ingestion and transformations writing directly to storage and selected workloads using local or single-node compute, while distributed pools can still serve many users simultaneously. It also suggests examining production query-size distributions to determine whether workloads are suitable for this approach and contends that cheaper compute could make data access more available to users and AI agents.

Fivetran, dbt, and Databricks Lakeflow are presented as complementary components of a modern data platform rather than direct competitors: Fivetran manages ingestion from more than 770 SaaS applications and databases, dbt creates tested, documented SQL-based business models, and Lakeflow handles custom batch and streaming pipelines plus orchestration. In an account-health use case, Fivetran loads Salesforce, billing, support, and application data into Databricks, while Lakeflow processes Kafka product events such as logins and feature usage; dbt then standardizes commercial and support data, and Lakeflow combines these models with usage signals into a continuously updated account-health table. The approach is positioned for organizations with diverse enterprise sources, analytics teams reliant on SQL modeling, and engineering needs involving real-time data, AI, machine learning, or unstructured processing, allowing teams to preserve existing dbt assets while reducing connector maintenance and building custom data products.