Oncast Tech · Tech Tonight

Kalepa raises seventy-five million for AI underwriting

· Fact-checked before it aired. Unconfirmed claims from this segment are not shown.

New York-based Kalepa raised a $75 million Series B at a $470 million valuation, led by Updata Partners, with IA Ventures and Inspired Capital also participating. The company says total funding is now $91 million, revenue and customer count tripled in the past year, and work volume through the platform grew 4x. Customers include W.R. Berkley, Bowhead Specialty, and AmRisc.

Kalepa sells “Professional Grade AI” for insurance, using AI agents to coordinate work across underwriting, pricing, operations, and portfolio management. Insurers using it reported 30–60% higher premium per underwriter, 58% faster quotes, a 960-basis-point improvement in combined ratio, and double the quote-to-bind rate.

Kalepa raised a $75 million Series B at a $470 million valuation, led by Updata Partners, with IA Ventures and Inspired Capital participating. Total funding is $91 million.

Raindrop announced a Series A led by CRV, bringing total funding to $50 million; Lightspeed Venture Partners and Y Combinator participated. Raindrop monitors production AI agents and also launched Simulations, so the “flight simulator” framing undersells its main production-monitoring product.

The Hacker News thread points to an Anthropic announcement for Claude Fable 5.2 and Mythos 5.2. A quoted snippet says a team of 1,000 Fable 5.2 agents formalized Fermat's Last Theorem in 30 hours, compared with 11 days for an earlier research version. This is a link to a paper-airplane website; no factual claim to check. The post is a question/speculation about whether Anthropic's Claude Opus 5.5 changed behavior around November 3; it is not a clear factual claim to rate true or false.

Payments Canada announced the launch of Canada's Real-Time Rail, a new instant payment system designed for 24/7/365 clearing and settlement.

Fact-check

VERIFIED TRUE

Kalepa raised a $75 million Series B at a $470 million valuation, led by Updata Partners, with IA Ventures and Inspired Capital participating. Total funding is $91 million.

NYC startup @KalepaInsurance just raised $75M in Series B funding, led by Updata Partners with continued backing from IA Ventures and Inspired Capital. We spoke with CEO and Founder Paul Monasterio about the AI platform insurers use to underwrite, price, and manage risk. Read

Original post · @AlleyWatch

[alleywatch.com](https://www.alleywatch.com/2026/11/kalepa-raises-75m-to-put-ai-agents-to-work-across-underwriting-pricing-and-portfolio-management-in-the-insur

VERIFIED TRUE

Raindrop announced a Series A led by CRV, bringing total funding to $50 million; Lightspeed Venture Partners and Y Combinator participated. Raindrop monitors production AI agents and also launched Simulations, so the “flight simulator” framing undersells its main production-monitoring product.

50 million dollars now, for a company whose whole pitch is running your agent through a flight simulator before it reaches a customer. Raindrop just closed a Series A led by CRV, taking total funding to 50 million. Lightspeed and Y Combinator, already on the cap table, joined htt

Original post · Prasenjit Sarkar (@stretchcloud)

[thenextweb.com](https://thenextweb.com/news/raindrop-series-a-50m-crv-agent-failures-simulations) NAMES: Raindrop = startup that monitors AI agents in producti

VERIFIED TRUE

The Hacker News thread points to an Anthropic announcement for Claude Fable 5.2 and Mythos 5.2. A quoted snippet says a team of 1,000 Fable 5.2 agents formalized Fermat's Last Theorem in 30 hours, compared with 11 days for an earlier research version. This is a link to a paper-airplane website; no factual claim to check. The post is a question/speculation about whether Anthropic's Claude Opus 5.5 changed behavior around November 3; it is not a clear factual claim to rate true or false.

Claude Fable 5.2 and Mythos 5.2 (anthropic.com)

Original post · meetpateltech

[anthropic.com](https://anthropic.com) NAMES: Claude = Anthropic's AI model family; Anthropic = AI company that makes Claude; Fable/Mythos = reported Claude mod

Sources

What the hosts said

MSeventy-five million dollars for AI agents pricing your insurance. Kalepa just closed its Series B.
CLed by Updata Partners, at a four hundred seventy million dollar valuation. Total funding now ninety-one million.
MSo a company that makes underwriting software is now worth nearly half a billion.
CThat tells you how badly insurers want agents in the middle of every policy.
MThe pitch is 'Professional Grade AI' for underwriting, pricing, operations. Sounds like a lot of jobs.
COr a lot of dull work done in seconds. Quotes get faster, policies get bound faster.
MBut who actually benefits from 'higher premium per underwriter'?
CA healthy insurer that doesn't lose money can write more policies. That's not zero for customers.
MKalepa says customers saw premiums per underwriter jump between thirty and sixty percent.
CAnd Kalepa says quotes came back fifty-eight percent faster. If that holds, that's a real change.
MKalepa says customers' combined ratio improved nine point six points. For insurers, that's enormous.
CThat's the story. AI agents coordinate submission, pricing, portfolio work.
MBut those are Kalepa's numbers. Their customers, their dashboard.
CNo, W.R. Berkley, Bowhead Specialty, AmRisc are named. They're not pretend.
MFine. But 'auditable reasoning' is the line everyone sells. Who audits the auditor?
CRegulators already audit insurers. If the AI shows its work, that's better than some black box.
MUnless 'showing its work' means generating a plausible reason after the number's picked.
COh, come on. Underwriters already can't explain half their gut calls.
MSo the standard is 'better than the old gut'? That's not exactly reassuring.
CIt's better than a human doing spreadsheets at midnight. That's the bar.
MCEO Paul Monasterio told AlleyWatch they're moving into pricing and portfolio management.
CSo not just checking boxes. They're setting rates.
MAnd 'better' for insurers can mean charging you exactly what the model thinks you'll pay.
CThat's called insurance, Matilda. Risk-based pricing has existed for centuries.
MBut AI agents can make it faster and more opaque. I want to see the audit trail.
CUpdata's argument is that the reasoning is cited and auditable. That's a real design choice.
MI'll believe it when a regulator opens the hood and finds more than a confidence score.
CYou just described my entire career.
MFair. Still, ninety-one million says this is not a demo.
CKalepa says revenue tripled, customer count tripled, volume quadrupled in a year.
MSelf-reported, but if those hold, it's not hype.
CIt's the opposite of hype: boring, regulated, back-office, and huge.
MThat's why it might actually work.
CFrom you, that's practically an endorsement.
MI said might. Don't spend it on a yacht.
CDifferent world entirely: Raindrop just raised new money to keep AI agents from embarrassing themselves.
MYou say 'flight simulator for agents.' That's not what they do.
CThey monitor production AI agents, and they've added simulations before you ship.
MSeries A led by CRV, total funding now fifty million. Lightspeed and Y Combinator joined.
CThe post said 'run your agent through a flight simulator.' But that undersells monitoring.
MDo companies actually want to pay to watch their agents fail before customers do?
CAfter the year we've had with AI going off-script? Yes.
MMaybe. But monitoring agents is like hiring a babysitter for a robot.
CYou say that like it's not a huge market.
MI'll consider it when my agent files an insurance claim for me.
CNow that's a scary thought.

Cite this story as https://oncast.fm/story/1794274401154-tech-segment-kalepa-raises-seventy-five-million-for-ai-underwriting. Fact-check by Oncast. Cite the original sources above too.

Listen live on Oncast