One Governed Layer, Now Live Across Excel, Power BI, and Horizon Context

Power BI and Excel users can soon query governed Snowflake Semantic Views directly. No mirrored data. No business logic was rebuilt outside the warehouse. That’s the news customers have been asking for and AtScale formally announced at Snowflake Summit 2026.

The new Snowflake Semantic Views XMLA Endpoint, Powered by AtScale, is heading into private preview. It gives Power BI and Excel access to the same certified metrics that already power Snowflake Intelligence and Cortex Agents. One definition. Every tool. No drift between what the AI agent sees and what the analyst sees.

AtScale showed up to make that point in person: a Green Circle sponsor at Booth 1106, running a home base at The Clancy for three days of coffee, beer, and co-working. CTO and co-founder Dave Mariani presented a session on “How a Universal Semantic Layer Provides Context for AI” (his takeaways from the week here). But the bigger story was the alignment between AtScale’s product news and Snowflake’s own Horizon Context announcement.

Agents Don’t Get a Margin of Error

Mariani made the case for why governance matters now, in a theCUBE interview with Snowflake’s Carl Perry.

“A definition that lives in only one dashboard is invisible to an agent acting somewhere else.“

That’s the risk with headless AI agents: they act on data without a human checking the math first. A metric defined once, in one tool, doesn’t exist anywhere else an agent might look. Mariani’s fix is to pair a creative, probabilistic LLM with a deterministic semantic query engine, so the agent gets the same governed answer every single time it asks.

Snowflake Built the Pipes. AtScale Built the Translator.

Snowflake moved in parallel, introducing Horizon Context, a new capability inside Horizon Catalog that collects, enriches, and activates governed business context for AI, BI, and apps. It runs on a three-stage model: Collect pulls metadata from systems beyond Snowflake. Enrich adds lineage and AI-generated documentation. Activate surfaces it all through Universal Search and automatic semantic view discovery.

Mariani is quoted in Snowflake’s announcement, framing the two efforts as one strategy:

“AtScale and Snowflake share a simple belief: business definitions should be governed once and used everywhere analysts and AI applications work. With Horizon Context, trusted semantic context becomes part of Snowflake’s broader governance framework. Whether teams define semantics in AtScale or Snowflake, they can invest with confidence.“

– Dave Mariani, CTO & Co-Founder, AtScale

Snowflake built the governance framework. AtScale extends those governed definitions into the tools Snowflake doesn’t natively reach, like Excel and Power BI.

Where Semantic Views Ends and AtScale Enterprise Begins

AtScale’s CPO Luis Maldonado addressed the question that came up repeatedly on the show floor in his interview on The Ravit Show: where does the free, embedded capability stop, and the fuller platform start?

His answer: the XMLA endpoint is fully embedded within Snowflake Semantic Views. No added infrastructure required. Customers who need a universal semantic layer across multiple data platforms, MCP server connectivity, and more complex modeling move up to AtScale Enterprise. And that move doesn’t cost them anything they’ve already built. AtScale Enterprise runs natively in Snowflake Container Services, so customers carry their work forward instead of starting over.

“There are millions of Power BI and Excel users. Right now, many of them feel left out if they can’t connect into a semantic layer. This solves that.“

– Luis Maldonado, Chief Product Officer, AtScale

Govern the Definition Once, or Let the Agent Guess

Every conversation on the show floor circled back to the same choice: govern the definition once, or let the agent guess. There’s no analyst left in the loop to catch the difference before the agent acts on it.