October 2025 Summaries

dbt Labs has announced the open-sourcing of MetricFlow under the Apache 2.0 license at its Coalesce 2025 conference, marking a significant move towards fostering trustworthy AI and aligning with the Open Semantic Interchange (OSI) initiative. This strategic shift aims to address the challenges posed by inconsistent metrics and fragmented definitions in AI deployment by offering a transparent and extensible engine for governed metrics. MetricFlow, which powers the dbt Semantic Layer following dbt Labs' acquisition of Transform in 2023, compiles metric definitions into code that computes them reliably, thus enhancing data governance and consistency across tools and platforms. This initiative, in collaboration with industry leaders like Snowflake and Salesforce, seeks to standardize semantic metadata exchange and mitigate the impact of nonstandardized data definitions on AI adoption. By providing a unified source of truth, MetricFlow empowers organizations to streamline analytics processes, reduce re-work, and improve trust in data, positioning itself as a cornerstone of the OSI's mission to propel data forward in the AI era.

Coalesce 2025, hosted in Las Vegas and attended by over 12,000 participants both in-person and online, unveils significant advancements in data, analytics, and AI through the introduction of the dbt Fusion engine and other innovations. The event highlighted a new era of open data infrastructure with the merger of dbt Labs and Fivetran, aiming to set a standard for open data practices. Central to the presentations were the new features of the Fusion engine, which promises faster development cycles and reduced cloud costs through state-aware orchestration and smarter testing. The engine also enhances the developer experience with tools like Intellisense and instant refactoring, while maintaining compatibility with existing dbt Core projects. Additionally, dbt Labs introduced AI-driven tools like dbt Insights and the dbt MCP server, aiming to provide governed, auditable AI experiences for data management and analytics. These innovations are designed to empower data practitioners to work more efficiently and cost-effectively, with a strong emphasis on trust and reliability in AI-driven data workflows.

Tristan Handy discusses the merger between dbt Labs and Fivetran to create an open data infrastructure, distinguishing it from the modern data stack (MDS) which has been transformative yet flawed due to integration challenges and limited scalability. The modern data stack enabled fast, scalable data manipulation by integrating software engineering practices but led to tool fragmentation and complex integration issues. In response, "all-in-one" data platforms emerged, offering integrated solutions but at the cost of user choice and potential vendor lock-in. The proposed open data infrastructure aims to overcome these limitations by emphasizing pluggability, standards-based integration, and flexibility in compute engine choice, allowing data teams to maintain control over their data environments while adapting to new AI capabilities. By leveraging open standards and maintaining vendor neutrality, this infrastructure seeks to provide reliable, high-quality data and metadata management, reducing costs and improving collaboration across data teams without sacrificing flexibility or performance.

The rise of open data infrastructure marks a significant shift in how enterprises manage and utilize data, addressing longstanding challenges of data underutilization and lack of integration across platforms. With the increasing necessity of AI in decision-making processes, transparency and reliability in data handling have become crucial, prompting the adoption of technologies like Apache Iceberg, which decouples storage from compute and supports schema evolution and ACID compliance across various engines and catalogs. This transition is exemplified by the merger of dbt Labs and Fivetran, aimed at creating a unified, open data infrastructure that enhances data movement, transformation, and activation with a focus on governance and portability. This infrastructure supports seamless, governed data access and sharing across diverse platforms without the need for re-platforming, thereby improving data utilization and enabling AI to operate on a robust, reliable foundation. The Model Context Protocol (MCP) is highlighted as an emerging standard that facilitates clean, universal connectivity for AI applications, ensuring that data workflows are both transparent and efficient. This open approach empowers organizations to effectively manage data across varied environments, enhancing their ability to derive actionable insights and scale AI solutions.

Fivetran and dbt Labs have announced a merger in an all-stock deal to form a unified company aimed at setting the standard for open data infrastructure, combining Fivetran's expertise in automated data movement with dbt Labs' strengths in data transformation. The merger, which is set to generate nearly $600 million in annual recurring revenue, is positioned as a response to the increasing need for scalable and interoperable data solutions in the AI era. George Fraser will lead the new company as CEO, while Tristan Handy will serve as co-founder and President, with both companies maintaining their commitment to open and community-driven development. This strategic partnership aims to simplify data management by unifying data movement, transformation, metadata, and activation, ensuring flexibility and avoiding vendor lock-in, while maintaining dbt Core's open license. The merger has received approval from both companies' boards and shareholders, with finalization pending customary closing conditions and regulatory approvals.

dbt Labs and Fivetran have announced a merger, creating a combined company with approximately $600 million in annual recurring revenue and over 10,000 customers, significantly expanding their reach in the data infrastructure space. This merger aims to leverage the complementary strengths of both companies—dbt's transformational capabilities and Fivetran's data ingestion expertise—while maintaining their commitment to simplicity and openness. The merger is not expected to disrupt current products or practices, as both companies are focused on enhancing their offerings without altering their core identities. The collaboration is poised to develop what is described as "open data infrastructure," which integrates seamlessly with various platforms and offers more flexibility compared to traditional data stacks. This strategic move is seen as a natural evolution given the long-standing partnership and shared vision between the two companies, and it is anticipated to foster greater innovation and open-source contributions within the analytics and AI sectors.