July 2026 Summaries

The Modern Data Stack, as discussed by five founders at a San Francisco event, represents a shift in data management from vertically integrated systems to a horizontally integrated model using cloud-based SaaS tools. This transition involves breaking down the traditional end-to-end data tools into distinct layers for ingestion, storage, transformation, monitoring, and operational analytics, which can be independently managed and integrated. This approach enables data teams to function more like product teams, incorporating DevOps practices to manage data as a product across various business domains. As data becomes more complex and abundant, the need for robust tooling to support smaller teams in managing vast datasets has become crucial, with cloud services and distributed systems playing vital roles. The democratization of data processes, allowing non-technical teams to engage with data at various stages, is a key characteristic of the Modern Data Stack, which aims to empower organizations to handle data at scale more efficiently.

Simon Späti's article outlines a comprehensive approach to managing multi-cloud cost analytics by integrating and visualizing expenses from various cloud service providers such as AWS and GCP, alongside revenue data from platforms like Stripe. The project aims to create a unified dashboard that offers a single view of cloud costs and revenue, enabling companies to better understand their financial performance across different platforms. The solution involves setting up data exports from AWS and GCP, integrating them using dlt, and visualizing the results with Rill. The article details the technical stack, including using DuckDB for local data storage and ClickHouse for cloud storage, and emphasizes the importance of composability, allowing for easy addition of new data sources. Späti also discusses the initial challenges of setting up cost exports and the benefits of achieving a clear, comprehensive view of multi-cloud expenses, encouraging others to replicate the project using the provided GitHub resources. Additionally, the article highlights the use of AI tools like Claude for prompt engineering, demonstrating the potential for further automation and refinement in managing cloud costs.

In the discussion on building authentic data communities, several founders from notable data-focused companies emphasize the importance of creating genuine, purposeful communities rather than just audiences. Alana Anderson from base case capital highlights the community-driven nature of the data ecosystem, where professionals often participate in numerous Slack channels to share insights. Benn Stancil of Mode suggests that being provocative can attract attention and help build an audience, while Boris Jabes of Census differentiates between audiences and communities by emphasizing mutual support among community members. Mark Grover from Stemma shares his experience with community-building at Lyft and emphasizes the importance of allowing communities to form organically without over-commercialization. Overall, the conversation underscores the evolving strategies and challenges in fostering meaningful connections within the data ecosystem.

Data Talks on the Rocks featured five founders from leading data companies discussing the biggest unresolved challenges in the modern data stack. Boris Jabes of Census identified the tension between fast and slow data processing and urged using available data before prioritizing speed. Mike Driscoll from Rill Data likened the current state of big data to pre-Google Maps navigation, emphasizing the need for more interactive data exploration tools. Egor Gryaznov of BigEye criticized the overcomplexity in data tools integration, suggesting a need for seamless interconnectivity akin to what Fivetran offers. Benn Stancil from Mode highlighted the discrepancy between data collection efforts and the actual value derived from data-driven decisions, pointing out the lack of mechanisms to track the impact of analytics beyond initial dashboards. The discussions also touched on measuring the ROI of data-driven decisions, with Gryaznov noting the challenge of recognizing the contributions of data teams in decision-making processes.