Multiplier Raises $6M Seed Financing to Build AI for Asset Managers
NEW YORK, August 26, 2026--Multiplier, an AI company making agent harnesses for asset managers, has raised a $6 million seed financing round. The round was led by Lux Capital with participation from Y Combinator, GoAhead Ventures, Rebel Fund, General Advance, Unpopular Ventures, and Amino Capital. Additional investors include Greg Jensen and Karen Karniol-Tambour (co-CIOs, Bridgewater Associates), Pete Briger (Chairman, Fortress Investment Group), Jasjeet Sekhon (CSO, DeepMind), Sandeep Nailwal (CEO, Polygon), Kaz Nejatian (CEO, OpenDoor), and Multiplier customers (especially Mercator Partners).
"We tried nearly every tool on the market and ended up with an unwieldy collection of fragmented point solutions," said Michael Siliciano, co-founder of California hedge fund Verso Partners. "We've also attempted to build our own tools, but maintaining them became a full-time job. That's why we brought in Multiplier."
Multiplier is the via media between hobbled software and duct-taped DIY AI efforts for leading asset management firms.
"Early vertical tools have grown fast," said Ian McInnis, Multiplier's co-founder and CEO. "'ChatGPT, but for finance, and it has some data access, and it's compliant.' Okay, great." Architectural and product decisions that helped others move faster have left investment management firms unprepared for the agentic era.
"You still need domain specialization, but you also need firm specialization," said McInnis. "Your data access demands are now 'all the data, information, and context that I as a human can access and then some,' whether or not the providers have built nice MCPs. Now that the AI's getting woven so deeply into your firm, your security demands go way beyond compliance box-ticking. Investors are legitimately concerned about their edge being stolen. All these demands are difficult to satisfy on webapps."
Multiplier meets those demands with repeatable infrastructure: it deploys the same substrate inside each customer's own systems, then builds firm-specific customizations on top. Every firm needs that substrate, but frameworks, info sources, processes, and tastes necessarily differ, so Multiplier builds bespoke evals for each firm it works with. "For the time being, that's what it takes," said McInnis. "Without evals, your AI drifts whenever you try to do any kind of continual learning, gets kind of crufty and dumb. It slips back into pretraining biases. Those biases are also why you need investor oversight when building evals."