AI
AI automation startup Model ML lands $250M Series B to handle investment banking busywork
London-based Model ML, which uses AI agents to handle financial modeling and report-writing for bankers, closes a fresh funding round at a $1.5 billion valuation.
TechCrunch — AI
Model ML, a London startup automating routine analytical work for financial professionals, has secured $250 million in Series B funding at a $1.5 billion valuation. The funding round was co-led by General Atlantic and QED Investors, arriving roughly a year after the company's $75 million Series A.
The startup, now three years old, builds AI agents designed to handle the grunt work that typically falls to junior bankers and consultants—tasks like constructing financial models, assembling presentation decks, and assembling due diligence documentation. Its co-founders, brothers Chaz and Arnie Englander, previously built Fat Llama, a peer-to-peer rental platform acquired by Swedish firm Hygglo for $41.5 million in 2022.
How the technology works#
Model ML's AI agents plug directly into the tools and systems that financial teams already use, including Excel and PowerPoint, as well as clients' internal data infrastructure. The agents maintain awareness of the broader context surrounding deals and transactions, enabling them to refresh financial models and update presentation materials automatically when market conditions or other factors shift.
The competitive landscape#
Model ML is part of a broader wave of startups targeting the high-volume, repetitive work in financial services. Rivals like Hebbia and Rogo—the latter of which pulled in $160 million at a $2 billion valuation in April—are pursuing similar automation strategies. Meanwhile, heavyweight AI labs including Anthropic and OpenAI have rolled out their own tools aimed at finance professionals.
Cost advantage#
Model ML distinguishes itself partly through its technology choices. Rather than relying on the expensive frontier models from major AI labs, the company runs on open-source models, allowing it to operate at roughly 2% of the cost while achieving faster execution.

