London Startup Model ML Valued at $1.5 Billion as AI Agents for Bankers Grow Revenue 12x
London startup Model ML, which builds AI agents for bankers, raised $250 million at a $1.5 billion valuation. The agent runs on the firm's own data: when news breaks about a client's company, it assesses how important the news is and prepares a briefing in Excel and PowerPoint. Annual recurring revenue (ARR) grew 12x over the year, and the number of daily users grew 26x. Investors are paying for narrow, industry-specific agents built into everyday work tools.
AI-processed from TNW; edited by Hamidun News
Bankers, consultants and investors spend a huge share of their week on repetitive work: tracking news about companies, compiling spreadsheets and building presentations from the firm's templates. London startup Model ML is automating this routine. It closed a $250 million Series B round, and investors valued the company at $1.5 billion. According to the startup itself, in the third quarter its annual recurring revenue (ARR, meaning the sum of regular customer payments converted to an annual figure) grew 12x compared with the previous year, and its number of daily users grew 26x.
Who Provided the Money and How the Company Is Growing
The round was led by General Atlantic and QED Investors. Other participants include Evolution Equity Partners, Airtree, Forestay Capital, Y Combinator, FT Partners, Latitude, Inception Fund and 13Books Capital. In total, the company has raised $350 million: a $75 million Series A in November 2025 was led by FT Partners.
The new funding will go toward expanding research and engineering teams in London and New York. The growth figures come from the company itself and have not been verified independently, and absolute revenue and user numbers have not been disclosed.
Model ML was founded in 2023 by brothers Chaz and Arni Englander. Before that, they created Fat Llama, a marketplace for renting out personal items, which Sweden's Hygglo acquired in 2022 for $41.5 million.
How the Agent Works in a Banker's Place
Each employee gets their own agent, which runs on the firm's own systems and data. Here is an example: a banker covers a certain company, and news about it breaks. The agent assesses on its own whether the news matters and prepares the analysis in advance, in the formats the firm uses. It works inside familiar tools, Excel and PowerPoint, where it assembles client presentations and reports for due diligence (the review of a company before a deal).
The key engineering detail is routing. Model ML sends each task to the model best suited for it. A growing share of the work runs on the company's own fine-tuned open models. According to the company, on the tasks they handle, these models cost 2% of the price of top models and in most cases run five times faster. The startup claims that client data is not used to train the models.
For developers of agentic products, this is a practical pattern: call the expensive top model only where it is needed, and move routine operations to small specialized models. The economics of this approach are especially visible in services with a large number of repetitive requests.
Job Cuts, Competitors and Caveats
The topic is a delicate one: banks are already linking job cuts to AI. Barclays spoke about this in October, and Standard Chartered plans back-office cuts by 2030. CEO Chaz Englander, in a conversation with Bloomberg, rejected the image of a "machine for saving on people." According to him, Model ML's average user is 33 years old, not 23, meaning they are experienced professionals rather than junior analysts. He added that such tools no longer count as productivity and efficiency aids: they generate revenue.
The startup's revenue is spread fairly evenly across banks, consulting firms and asset managers. Competition is intense: New York-based Rogo raised $160 million in April at a $2 billion valuation, and OpenAI and Anthropic also sell tools for financial professionals.
What to keep in mind. The claim about 2% of the cost and five-times speed applies only to the tasks covered by the company's own models, not to all of the agent's work. The financial industry is heavily regulated, and the question of who is liable for an agent's errors remains open for clients.
For founders and product managers, the takeaway is this. Investors are willing to pay for vertical agents built into the industry's everyday tools, working on client data and delivering measurable results. A general-purpose chatbot loses in this niche to whoever speaks the language of Excel and a given firm's templates.
- Nov 21, 2026FDE job postings grew 4.6x in a year: what a forward deployed engineer actually does at the client's site
- Nov 21, 2026A swimming team captain handed ChatGPT three scheduled tasks and cut her routine work by a multiple
- Nov 21, 2026OpenAI GPT-6 Cyber Found About 35 Flaws in iPhone, Mac and Other Apple Devices
- Nov 21, 202622 Claude Agents Spent Six Weeks on Black Hole Problems — and Solved Two
Key Questions Answered
Who Provided the Money and How the Company Is Growing?
The round was led by General Atlantic and QED Investors. Other participants include Evolution Equity Partners, Airtree, Forestay Capital, Y Combinator, FT Partners, Latitude, Inception Fund and 13Books Capital. In total, the company has raised $350 million: a $75 million Series A in November 2025 was led by FT Partners. The new funding will go toward expanding research and engineering teams in London and New York.
How the Agent Works in a Banker's Place?
Each employee gets their own agent, which runs on the firm's own systems and data. Here is an example: a banker covers a certain company, and news about it breaks. The agent assesses on its own whether the news matters and prepares the analysis in advance, in the formats the firm uses. It works inside familiar tools, Excel and PowerPoint, where it assembles client presentations and reports for due diligence (the review of a company before a deal). The key engineering detail is routing.
What are the key details on Job Cuts, Competitors and Caveats?
The topic is a delicate one: banks are already linking job cuts to AI. Barclays spoke about this in October, and Standard Chartered plans back-office cuts by 2030. CEO Chaz Englander, in a conversation with Bloomberg, rejected the image of a "machine for saving on people." According to him, Model ML's average user is 33 years old, not 23, meaning they are experienced professionals rather than junior analysts. He added that such tools no longer count as productivity and efficiency aids: they generate revenue.
Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews
Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.