VC & Startups/Nov 17 edition/Reporting & analysis
FundingValuations

FundingRounds, leads & terms

Model ML raises $250M Series B at $1.5B valuation, co-led by General Atlantic and QED

London-based Model ML, which builds AI agents for bankers, consultants and investors, has raised a $250M Series B at a $1.5B valuation. The round takes total funding to $350M less than two years after its $12M seed, as it shifts toward proactive agents.

From the Tech.eu article of November 17, 2026 and cited sources: General Atlantic and QED Investors co-led the $250M Series B, which values Model ML at $1.5B. [1] [6] [2] [4]

SOURCES[1] Tech.eu: Model ML hits $1.5B valuation (Nov 17, 2026) tech.eu[6] Bloomberg: AI Startup for Bankers Raises Funds at $1.5… bloomberg.com[2] Tech.eu: Model ML raises $75M (Nov 24, 2025) tech.eu

AI-assisted, not human-edited. Methodology Corrections Who owns this site

KEY FACTS
What happened
General Atlantic and QED Investors co-led the $250M Series B, which values Model ML at $1.5B. [1] [6] [2] [4]
Original source
Tech.eu article, independent perspective
Brief by
AT Pulse research desk AI-assisted; publisher and contact on the desk page
When
November 17, 2026 article published
Evidence
7 sources across 5 sites

Reach: not measured yet

Illustration from Tech.eu: Model ML raises $250M Series B at $1.5B valuation, co-led by General Atlantic and QED
Image: Tech.eu — Original article ↗
THE CORE IDEAS4 TAKEAWAYS
01

General Atlantic and QED Investors co-led the $250M Series B, which values Model ML at $1.5B. Bloomberg also reported that valuation. The round follows a $12M seed led by LocalGlobe in February 2025 and a $75M Series A led by FT Partners in November 2025, bringing total funding to $350M. [1] [6] [2] [4]

02

The product has moved beyond automating Word, PowerPoint and Excel deliverables toward proactive agents. When news breaks about a company a banker follows, the agent judges whether it matters, analyses it and drafts a report in the firm's own format. [1] [5]

03

Model ML is not tied to one AI provider. It picks models for each job based on performance, speed and cost, and relies more and more on fine-tuned open models. It says these cost about 2% as much as frontier models, run about 5x faster and match their accuracy on some tasks. It has not named the base models or published the task set or any benchmark. [1] [3] [7]

04

The company says annual recurring revenue grew 12x year on year in Q3 2026 and daily active users grew 26x. These multiples are self-reported and come with no disclosed revenue base, so they show momentum but not scale. [3]

WHY IT MATTERS

Investors priced a finance-specific AI agent company at $1.5B within two years of its seed round. If cheap open-model routing holds up, margins improve and dependence on frontier vendors falls.

Read the full assessment

Buyers still lack accuracy evidence, and rivals Rogo and Hebbia add pricing pressure.

Executive brief

Model ML, a London-based company that builds AI agents for bankers, consultants and investors, says it has raised a $250M Series B at a $1.5B valuation. General Atlantic and QED Investors co-led the round (Tech.eu). Its only earlier priced rounds were a $12M seed in February 2025 and a $75M Series A in November 2025 (Tech.eu, 2025). Total funding is now $350M. The company's growth figures are self-reported: it says annual recurring revenue (ARR) grew 12x year on year in Q3 2026 and daily active users grew 26x. It has not disclosed any absolute revenue figure (Tech Funding News).

What changed and event timeline

  1. $12M seed

    Coverage describes the company as already serving banks, asset managers and two Big Four firms ().

  2. $75M Series A

    FT Partners led, with Y Combinator, QED, 13Books, Latitude and LocalGlobe joining. At the time the product focused on automating Word, PowerPoint and Excel deliverables (;).

  3. International expansion

    Model ML says it opened offices in Paris, Madrid, Tokyo, Singapore and Dubai during the year ().

  4. $250M Series B at $1.5B

    Bloomberg independently reported the $1.5B valuation the same day (;).

Capabilities and access

  • The pitch has moved from workflow automation to proactive "personal AI agents."
  • Product list. Earlier materials name Custom Agents, Agent Task Manager, Document Review, Notetaker, Grids, Chat, Connectors and Sites (Model ML).
  • Firms can run it inside their own infrastructure, and the company says customer data is not used for training.
Read the full section
  • Product shift. The pitch has moved from workflow automation to proactive "personal AI agents." For example, when news breaks about a company a banker follows, the agent judges its relevance, analyses it and drafts a report in the firm's own format (Tech.eu).
  • Product list. Earlier materials name Custom Agents, Agent Task Manager, Document Review, Notetaker, Grids, Chat, Connectors and Sites (Model ML).
  • Deployment. Firms can run it inside their own infrastructure, and the company says customer data is not used for training.
  • Models. No specific model names or versions are disclosed.
  • Pricing. Not public.

Technical analysis for researchers and developers

  • Model ML says it is not tied to any single AI provider.
  • The company says it is using more internally fine-tuned open-source models.
  • No documentation. None of these claims can be reproduced from public material.
Read the full section
  • Model routing. Model ML says it is not tied to any single AI provider. It tests models on financial tasks and picks among them by performance, speed and cost (Tech.eu).
  • Fine-tuned open models. The company says it is using more internally fine-tuned open-source models. It claims they cost about 2% as much as frontier models, run about 5x faster, and match their accuracy on certain tasks (UltraThink; Tech Funding News).
  • No documentation. None of the reviewed sources describes the base models, the task set, the evaluation method or any public benchmark. None of these claims can be reproduced from public material.

Claims and evidence

  • $250M round, $1.5B valuation, General Atlantic/QED co-leads
  • $350M total raised — Company-announced ()
  • 12x ARR growth and 26x daily-user growth in Q3 2026
Read the full section
ClaimStatus
$250M round, $1.5B valuation, General Atlantic/QED co-leadsAnnounced by the company. Bloomberg independently reported the $1.5B figure (Bloomberg). No regulatory filing confirms it. The reviewed sources don't say whether $1.5B is pre- or post-money (UltraThink).
$350M total raisedCompany-announced (Tech.eu)
12x ARR growth and 26x daily-user growth in Q3 2026Self-reported. No absolute revenue figure or comparison period is given.
Open-model cost, speed and accuracy figuresSelf-reported. No independent evaluation was found.
ISO/IEC 27001:2022 and SOC 2 Type II certificationsCompany-stated (UltraThink)

Context and prior work

  • Brothers Chaz and Arnie Englander are repeat founders. Their earlier companies, Fancy and Fat Llama, were both acquired (Tech Funding News).
  • The same article lists Rogo ($160M Series D, April 2026), Hebbia ($130M Series B, July 2024), Unique, Eunice and Gradient Labs.
  • Comments on the company have come from its own investors.
Read the full section
  • Founders. Brothers Chaz and Arnie Englander are repeat founders. Their earlier companies, Fancy and Fat Llama, were both acquired (Tech Funding News).
  • Competitors. The same article lists Rogo ($160M Series D, April 2026), Hebbia ($130M Series B, July 2024), Unique, Eunice and Gradient Labs.
  • Investor commentary. Comments on the company have come from its own investors. At the Series A, FT Partners founder and CEO Steve McLaughlin, whose firm led that round, praised Model ML (Model ML).

Limitations, safety and contested findings

  • The Series B headline says the raise came 18 months after Model ML left stealth, which points to around May 2025.
  • No published safety data. There is no public information on error rates or human review of outputs from agents that act on their own.
  • Growth multiples without a baseline say little about the company's actual scale.
Read the full section
  • Stealth timing conflicts. The Series B headline says the raise came 18 months after Model ML left stealth, which points to around May 2025. Series A coverage instead dates the launch to about twelve months before November 2025, i.e. late 2024 (Tech.eu, 2025).
  • No published safety data. There is no public information on error rates or human review of outputs from agents that act on their own.
  • Hard to evaluate. Growth multiples without a baseline say little about the company's actual scale.
  • Prior concern. At the Series A, the CEO himself said that errors slip through manually checked deliverables of 100 pages or more (Tech.eu, 2025). That raises the question of how outputs from proactive agents get verified.

Business and practitioner implications

  • Pricing a company at $1.5B within roughly two years of its seed round shows growth investors will pay up for vertical AI agents in finance.
  • Cost strategy. Routing between models and leaning on fine-tuned open models is a margin play.
  • Ask for task-level accuracy evidence, verification workflows, and proof that on-premises deployment is actually available.
Read the full section
  • Investor appetite. Pricing a company at $1.5B within roughly two years of its seed round shows growth investors will pay up for vertical AI agents in finance.
  • Cost strategy. Routing between models and leaning on fine-tuned open models is a margin play. If it holds up, it lowers the company's dependence on frontier-model vendors.
  • Buyer checklist. Ask for task-level accuracy evidence, verification workflows, and proof that on-premises deployment is actually available.
  • Competition. The field already includes heavily funded rivals such as Rogo and Hebbia. Expect pricing pressure and a fight over integrations.
FOLLOW THE EVIDENCE

The source trail.

Sources (7)
A LITTLE LESS NOISE. A LOT MORE CONTEXT.

Stay curious.
Follow the evidence.

Independent perspectives, the original sources, and room for the questions that don't have easy answers.

How we build the brief
Connect with us

Find us where you already read.