Model ML just raised $250 million to automate the work junior bankers have done for decades. TechCrunch AI reports that the London-based startup closed a Series B at a $1.5 billion valuation. General Atlantic and earlier backer QED Investors co-led the round. It comes just a year after Model ML announced its $75 million Series A.
The Short Version
- What: A $250M Series B for Model ML at a $1.5B valuation
- Who: Co-led by General Atlantic and QED Investors
- What it does: AI agents that build financial models, prepare pitch decks and write due diligence reports
- The angle: It runs on open source models and says that costs about 2% of what frontier models cost
- Why it matters: Investors are betting big that entry-level finance and consulting work can be automated, and that this can be done cheaply
What Model ML Actually Does
Ask any first-year analyst what eats their week and you’ll hear the same list: building financial models, formatting pitch decks and grinding through due diligence reports. That’s the work Model ML is going after.
Its AI agents plug into the systems, data and tools clients already use, including Excel and PowerPoint. According to the company, the agents understand the full context of the deal a team is working on. When something changes, like new numbers or a shift in the deal, the agents can update the financial models and presentations on their own.
That last part is what stands out. Producing a first draft of a deck is one thing. Keeping a 40-tab model and a pitch book in sync while a deal keeps changing is where junior bankers lose their nights and weekends.
The People Behind It
Brothers Chaz and Arnie Englander co-founded Model ML three years ago. It’s not their first company. They also co-founded Fat Llama, a peer-to-peer rental marketplace that Swedish rival Hygglo bought for $41.5 million in 2022.
Going from a $41.5 million exit to a $1.5 billion valuation in about three years is a big jump. It tells you how much investors want a piece of vertical AI aimed at expensive professional services.
A Crowded Field
Model ML isn’t alone. Several well-funded companies are going after the same work:
- Hebbia is automating research and analysis for finance professionals
- Rogo raised $160 million at a $2 billion valuation in April
- Anthropic and OpenAI have both built tools for financial professionals
So the startups face pressure from two directions. They’re competing with each other, and the frontier labs are moving into their market.
The Open Source Bet
This is where Model ML stands apart. Rogo and the big labs lean on frontier models. Model ML says it runs on open source models, and that lets it work at about 2% of the cost of frontier models while also delivering results faster.
If that claim holds up, it matters a lot. Banks and consulting firms run huge volumes of repetitive work. When you’re updating models and decks thousands of times across hundreds of deals, inference costs pile up quickly. A 50x cost advantage could decide who wins the big enterprise contracts.
It also fits a wider trend. More vertical AI companies are finding that well-tuned open models, built around one specific workflow, can match frontier performance on narrow tasks. The deep integration and domain context become the product, and the model behind it becomes something you can swap out.
What This Means for Practitioners
- For banks and consultancies: Expect more vendors pitching agents that work inside Excel and PowerPoint, not just chatbots on the side. Look closely at cost per task, not just output quality.
- For junior professionals: The grunt work that defined the analyst years is shrinking. The skills that will matter more are judgment, client relationships and spotting when the AI’s output is wrong.
- For AI builders: Model ML is a strong example of open source models running a high-stakes enterprise workflow. Expect more startups to compete on margin this way.
What Comes Next
With fresh capital and a clear cost story, Model ML is likely to push hard into bigger banks and consulting firms. The real test is trust. Finance runs on precision, and one wrong number in a model can sink a deal. If Model ML can show its cheaper, faster agents are also reliable, the analyst pyramid at investment banks could look very different within a few years.
You can find more details in the original TechCrunch AI report.