Scale AI’s Former CTO Launches AI Agent That Could Solve Big Data’s Biggest Problem
Roundup #6
Hello data folks 👋
Scale AI's former CTO Arun Murthy launched Isotopes AI with $20M seed funding to bridge the gap between data infrastructure teams and business users who need the insights. The goal is to reduce the bottleneck between data engineering resources and business intelligence requests.
The startup's AI agent, Aidnn, handles complex multi-step data operations. It extracts from sources like Salesforce and Snowflake, cleans, normalizes, joins datasets, and generates answers and documents through natural language queries. The team has deep experience with enterprise data infrastructure challenges, with Murthy having co-founded Hadoop at Yahoo and later Hortonworks.
Source: Techcrunch
Snowflake’s Quarterly Financial Report
Snowflake | September 5, 2025 | 5 minute read
Snowflake's Q2 filing confirms its strategic push beyond its core warehouse with the $164.5M cash acquisition of PostgreSQL provider Crunchy Data. This move signals a deliberate expansion into operational and transactional workloads.
The filing also highlights continued heavy investment in AI/ML capabilities, noting increased third-party cloud costs related to GPUs for technology development. Additionally, the report highlights the adoption of open data formats like Apache Iceberg as key to reducing customer lock-in and competing more broadly.
The platform's direction is clear. With the combination of native operational database support (Postgres), a push for open formats, and deep AI/ML investment Snowflake wants to become a more vertically integrated platform.
Domo’s Quarterly Financial Report
Domo | September 5, 2025 | 5 minute read
Domo's latest 10-Q filing shows a deep focus on AI capabilities while highlighting the growing stakes of data governance. The company is investing in ML algorithms and predictive analytics but also identifies new risks, including those from potentially "deficient, inaccurate, biased" AI outputs.
The filing specifically notes the need to comply with emerging regulations like the EU AI Act and Utah's AI Policy Act. There’s an increasing operational burden for platforms handling AI-driven analytics and the potential for new liabilities that are tied to model performance and data usage.
This is an indicator that vendor risk assessment must now include a deep dive into a platform’s regulatory and AI compliance roadmap.
How Ready Is Agentic AI for a Full Analytics Engineering Workflow?
Analytics Engineering Roundup | September 7, 2025 | 44 minute podcast
On the Analytics Engineering Podcast, Mikkel Dengsøe shares results from his experiment using Cursor, dbt, & Omni. The biggest surprise? A massive productivity boost in generating dbt tests by feeding the AI a 'testing philosophy'. But the real challenge emerged at the BI layer. Mikkel explains why AI's reasoning needs expert supervision before it's ready for business users.
Two More Things
How did you get really good with SQL? (Reddit)
Things people need to stop doing in data immediately (LinkedIn)
That’s the brief.