Meridian raises $17 million to rebuild proxy spreadsheet

The battle to tame spreadsheets with AI is far from over. A new company called Meridian has emerged from obscurity with a more comprehensive IDE-based approach to proxy financial modeling — and a lot of funding to build it. The company on Wednesday announced $17 million in seed funding at a post-money valuation of $100 million.

“Our goal is to make financial models and spreadsheets more predictable and auditable,” CEO and co-founder John Ling told TechCrunch. “How do you take a process that would normally take many hours and compress it into about 10 minutes?”

The round was led by Andreessen Horowitz and the General Partnership, with participation from QED Investors, FPV Ventures and Litquidity Ventures. The company says it is currently working with teams at Decagon and OffDeal, signing contracts worth $5 million in December alone.

Excel agents have been a popular target for AI startups, in part because of the high cost of human-led financial analysis. But while previous Excel agents, like Shortcut AI, built agents into Excel, Meridian operates as a standalone workspace, closer to Cursor. This allows the application to act like an IDE, integrating data sources and other external references that might otherwise occur.

The Meridian team, based in New York, includes graduates from AI companies such as Scale AI and Anthropic as well as financial experts from companies such as Goldman Sachs.

As Ling describes, the biggest challenge facing Meridian is the stringent requirements of financial clients, which often conflict with the non-deterministic nature of AI models.

“If you go to 10 different software engineers at Google and want to add some new feature to an app, you’ll probably get 10 completely different apps,” Ling says. “And that’s perfectly fine.” “But if you went to 10 banking analysts at Goldman Sachs and asked for 10 valuation models for a company, you would probably get 10 almost identical business plays.”

As a result, the Meridian team has done significant work to make their outputs more auditable and quantifiable, while maintaining the flexibility of LLM-based tools. The result is a combination of agentic AI and more traditional tools, reducing the hallucinations that slow down many enterprise deployments.

“Our goal is to remove the layer of uncertainty right out of the LLM process,” Lee says. “You know exactly how the logic flows, and all these assumptions or whatever else goes into the model, you can see exactly where they’re coming from.”