AtScale

/ Startups / By

Founder & Key Exec: David Mariani (Co-Founder & CTO), Chris Lynch (CEO)
Founding: 2013
Mission: Helping customers create value from their data and AI/ML investments
Employees: 80 & 40% Local
Workplace: Hybrid
Stage & Capital Raised: Series D & $120M raised
Investors: Morgan Stanley, Storm Ventures, Atlantic Bridge, Wells Fargo,
Key Customers: The Home Depot, Vodafone, Trek, Corning, Telus

AtScale is building the universal semantic layer to power Enterprise AI. This team has spent more than a decade building through the nuances of cloud data, metric dissemination, and traditional BI analysis that’s now evolving to AI analysis through the new MCP, model context (open) protocol.

A platform shift ago, Dave Mariani arrived at Yahoo after his startup was acquired by the purple giant near the height of its powers. As the Chief Data Officer, he was tasked with unifying analytics systems across Yahoo’s disparate web properties and ad businesses on the technical side of the house. Since everyone at Yahoo (obviously) had their own stack, from Front Page to Finance to Search, Display, and even Yahoo Fantasy…there were a lot of unique variables to wrangle.

Big Data was all the rage and the Hadoop open source framework was even conceived inside of Yahoo. Dave and his team built an internal solution that could give each team complete data fidelity through a multi-modal service with full database access for business units to operate independently. It was really effective. There wasn’t anything on the market like it! 

But Yahoo wasn’t in the business of building and selling software products, so Dave left to build a universal semantic layer for enterprises everywhere who needed better business intelligence and data access from their databases & data lakes. 

The public cloud services market today has surpassed $700B (src). Databricks and Snowflake are both near or beyond $4B in annual revenue run rates (src / src). Having a single source of truth to access and analyze data is corporate technology table stakes.

AtScale has built the universal semantic layer to make enterprise data analytics consistent, scalable, and increasingly AI-ready. Ok, cool. But why is having a universal semantic layer solution so important if you already have a data lake? 

A semantic layer sits in the middle of a company’s data consumers and its data lake to help deliver a universal source of truth, better performance (i.e. speed) and meaningful cost management reduction. The cost of querying data has risen sharply and, if all of your cloud data lives in a data lake, paying “per drink” can be expensive. 

AtScale supports MDX, DAX, SQL, and a new MCP integration, which means Excel users can run live queries directly on data in Snowflake, Power BI users get full functionality without needing move data into a Fabric data mart, and Chatbots & agents can run queries like humans.

With the arrival of Generative AI, enterprises see new ways to interact with their data. AI platforms like ChatGPT and Claude can write queries, but they don’t understand your business logic which results in incorrect results. Semantic layers help act as the filter and context engine to deliver accurate, enterprise ready results without relying on a cloud partner or needing to rewire an analytics system tied to any vendor. AtScale acts as the brain that translates prompts into enterprise-specific queries using clean, consistent semantics. Think of it as a translation layer between messy corporate data and intelligent agents.

One of AtScale’s under-the-radar innovations is SML, its open-source semantic modeling language. It’s object-oriented, modular, and designed for real-world enterprise scale which enables teams to build semantic models from reusable components, not just raw tables. This is part of AtScale’s bigger vision: making semantics not just machine-readable, but human-composable.

After a decade of infrastructure-first building, AtScale is leaning hard into Agentic AI. Their team is building agents to run analytics, make decisions, and trigger workflows. AtScale is investing in its own MCP server to power these new, autonomous use cases. They’re also building automated semantic model creation. Today, building semantic models still takes human effort. AtScale wants to get to “90% done with AI,” letting teams guide the last 10%. That unlocks semantic layers for everyone, not just data lake elite!

With hundreds of enterprise customers (think Fortune 1000 logos) and AI driven tailwinds, AtScale is emerging as a critical player for Enterprise AI readiness, a platform-agnostic service that could define metrics once and serve them anywhere from Tableau to Excel or even AI agents through the MCP (model context protocol).

AtScale is built for the enterprise: SSO passthrough, row/column-level security, Active Directory and Okta support for hundreds of Fortune 1000 customers. Most notable in 2025 has been the shift from “semantic layer is nice-to-have” to “AI needs semantic layers to work” after the release of the model context protocol by Anthropic in Q4 2024. 

With the AI tailwind firmly at its back after 9 years of evangelizing why a semantic layer is needed, AtScale is now at the center of enterprise AI conversations, and enterprise SQL transactions. This team is growing in the enterprise through both new logos and deeper expansion into existing customers, building a bright AI enabled future right here in Boston.

Operators to Know (Locally):

My investigative powers continue to need work so apologies to the AtScale team I know I missed many up & coming operators internally

If I were interviewing here are some questions I’d ask:

  • What are the biggest challenges as you scale the company into 2026?
  • What is the long term vision for the company? How has that changed in 2025?
  • What are the most important resources you’ll be looking to add in 2026 // teams that need the most help?
  • What are the biggest enterprise opportunities and risks with Fortune 1000 buyers?

We’re optimizing for readability here so to learn more about AtScale you’ll have to D.Y.O.R. I’m excited to watch this team bring more enterprise teams into the age of AI. All analysts and corporate titans applaud your efforts. See you around town!

Subscribe
Subscribe