Nico Laqua has given some version of the same speech in every public appearance over the last eighteen months. It goes like this: a hundred-year-old insurance carrier needs tens of thousands of people shuffling paper to operate. Corgi will run the same institution with a hundred AI engineers. The incumbents are dinosaurs run by 90-year-olds and fax machines. Insurance is the biggest words-based industry, and computers finally got good at language.
That claim holds up the whole pitch. It justified the $108M raise at a $630M valuation in January 2026, and the rounds since: a $160M Series B at $1.3B in May, a $106M B1 at $2.6B three weeks later, and a B2 reported in July at $4.0B. About $378M raised in total. It also carries the Chubb comparisons, the 3,000-year framing, and the "we are the carrier" posture.
Here is what he has actually said, word for word, on the record. Every link jumps to the moment in the video.
"The way that insurance companies have worked for the past 3,000 or so years is that normally as a function of revenue, as you get more revenue, more customers, you hire more people in order to interpret these contracts."
AI-Native Insurance: Reimagining Insurance Foundations, 3:00
"A normal insurance carrier might employ 50, 60, 70,000 people that are kind of doing the task to run the company. In our case, we have engineers doing it. So it's kind of a fixed headcount regardless of how much of the product we're selling."
AI-Native Insurance: Reimagining Insurance Foundations, 23:00
"Think like 80s back office fax machine, phone calls, all the bad things. They tend to run without a lot of technology."
Building with Corgi: Kanyi Maqubela with Nico Laqua and Emily Yuan, 2:26
"These kind of highly repetitive, highly regulated tasks that normally require tens of thousands of people at scale can instead be done with practically no people, maybe just a couple really smart engineers."
"Probably around 50,000 people run Chubb. And almost none of those people are on go to market. They're like, you know, sending faxes to each other and shuffling paperwork around. It's our vision that this type of financial institution can be run instead of with 50,000 people with like maybe a hundred AI engineers."
Nico Laqua and Emily Yuan, Corgi Founders on AI Native Insurance, 39:25
"AI agents talk to other AI agents instead of people talk to other people."
"I'm sure that an AI native mortgage originator would be a great company. Because probably the way they work is with like 80s back office and like thousands of people doing kind of repetitive work."
AI-Native Insurance: Reimagining Insurance Foundations, 18:01
Next to the labor argument sits a second move he repeats almost as often: naming the incumbents out loud, a stance you can only take if you have no fear of the org chart.
"There's two companies mostly that service a good portion of technology companies. What brokers do is they take policies normally from a carrier called State National, which is this multi-billion dollar insurance carrier. I promise you they don't care very much about startups, so I don't mind naming them."
"I don't think that there's a lot of incentive for these hundred-year-old behemoths to say 'sorry 80,000 people that work here in all of our call centers, you're out of a job.' I don't think the 90-year-olds that are running these companies or the shareholders would very much appreciate those sort of approaches."
"We don't sit on top of an insurance carrier. We don't sit on top of a reinsurance company. We are the reinsurance company. We are the carrier."
"We look at the most regulated type of financial entities in the United States and we actually become those. We don't sell to them. We don't build on top of them. We don't talk to them. We actually become them."
When an investor hesitates, Nico offers an alternative. He delivers it as a flex, a closing line meant to make the safer option feel embarrassing.
"If you don't think that those things are true or you think AI is just a fad or worried about like I don't know hallucination or all that sort of stuff, then you can buy stock in Berkshire Hathaway or any of our competitors. It's very easy to do instead of investing in us."
So let's take him up on it.
Berkshire Hathaway is a conglomerate. Insurance is one segment next to railroads (BNSF), utilities and energy, manufacturing, and retail. Nico's thesis is about insurance carriers, so the comparison below uses only the insurance arm: GEICO, Berkshire Hathaway Reinsurance Group, and Berkshire Hathaway Primary.
For context, the parent company does about $371B in revenue with about 390,000 employees across all segments, or roughly $950,000 per employee. That is not the number we compare against.
The insurance segment alone brings in over $100B in premium with an effective workforce of 100,000 to 150,000. That works out to $700,000 to $1,000,000 in revenue per employee.
Corgi, in January 2026, reported about $40M in annual revenue with 137 employees. That was about $292,000 per employee.
By September 2026 the picture had moved. The company says it hit about $100M in annualized revenue in May, and headcount is now around 250. That is about $400,000 per employee. Neither number is audited, and the revenue figure is from May while the headcount is from July to September, so treat it as approximate.
Formula: revenue per employee = annual revenue รท employees.
Berkshire insurance segment Revenue ~ $100,000,000,000 Employees ~ 125,000 (midpoint of 100k to 150k) Per head ~ $800,000 (range $667,000 to $1,000,000) Corgi, January 2026 Revenue = $40,000,000 (~$40M ARR) Employees = 137 Per head ~ $292,000 Gap 2.3x to 3.4x (2.7x at midpoint) Corgi, September 2026 Revenue ~ $100,000,000 (annualized, May 2026) Employees ~ 250 Per head ~ $400,000 Gap 1.7x to 2.5x (2.0x at midpoint)
| Revenue per employee | |
|---|---|
| Berkshire insurance segment | $700k to $1M |
| Corgi, Jan 2026 | $292k |
| Corgi, Sep 2026 | ~$400k |
On the one metric that lines up most directly with the labor-replacement thesis, insurance revenue per person hired, Berkshire's insurance segment is still about 2x ahead of the AI-native startup that keeps mocking it for being run by 90-year-olds and fax machines. In January the gap was 2.7x. It is closing, but it is still there.
Corgi went from 137 people to about 250 while revenue grew from about $40M to about $100M. Revenue per head rose, but headcount nearly doubled along the way. That is not the "fixed headcount regardless of how much of the product we're selling" the pitch describes. It looks more like the hire-as-revenue-grows curve Nico says has defined insurance for 3,000 years, just a flatter one.
"A hundred AI engineers replace 50,000 people" does not describe the present. It is a bet. The company's year-end target is a $450M run rate. At roughly today's headcount that would be about $1.8M per employee, well past Berkshire, and the thesis would look right. Until that shows up in real numbers, the dinosaur is still ahead on the exact metric the pitch is built around.
"We want a hundred billion dollar company. We want a company that actually matters and does important things."
Where the quotes come from. All quotes are from eight archived YouTube appearances by Nico Laqua between 2024 and 2026. Speaker attribution was checked against the speaker markers in the transcripts and against YouTube's auto-captions, pulled with yt-dlp. Quotes from multi-person panels were left out unless the context made clear who was speaking.
Timestamps. Each link jumps to the second the quote starts in the video's auto-captions. Three videos had timed captions (9JL2sJ6ao0A, AzNfFI6zFHE, JY8cnW2lZg0). One did not (JnM78mIN18k), so quotes from it were swapped for similar statements from captioned videos, so that every quote can be checked.
How revenue per employee was calculated. Stated annual revenue divided by stated full-time headcount. Berkshire figures are approximate annual report figures (~$371B revenue, ~390,000 employees). The insurance-segment range covers GEICO, Berkshire Hathaway Reinsurance Group, and Berkshire Hathaway Primary, with 100,000 to 150,000 staff against ~$100B in revenue. Corgi's January figures come from founder interviews and the January 2026 fundraise announcement (~$40M ARR, ~137 employees). The September figures come from press coverage of the 2026 rounds (~$100M annualized revenue in May) and reported headcount (~250, July to September 2026). All Corgi figures are company-reported and unaudited.
Limits. Revenue per employee is not a clean way to compare productivity across industries. Insurance revenue is inflated by premium pass-through and float compared with software revenue. Berkshire's insurance arm sits on a sixty-year-old balance sheet, and Corgi's own risk-bearing entities only took effect on January 1, 2026, so on almost any maturity-adjusted measure the comparison is stacked against Corgi. The claim here is narrow: on the one metric closest to the labor-replacement thesis, the gap favored the incumbent by 2.3x to 3.4x in January 2026 and by about 1.7x to 2.5x in September.
What would prove this wrong. If Corgi's revenue per employee reaches $700k to $1M within five years while headcount grows more slowly than revenue, the thesis holds and this note is out of date. The thesis can be tested. The Series B, B1, and B2 have now happened and the gap narrowed from 2.7x to about 2.0x. Rerun the numbers at year-end against the $450M target, and at the next round.
Not covered. Corgi's regulatory setup (an Arizona risk retention group and reinsurance captive effective January 2026, a new admitted carrier announced August 2026, and a claimed $35M carrier acquisition that no public filing names), distribution, reinsurance, loss ratios, and the broker markets it sells into. Each of these affects whether the bet works, separate from the efficiency point. This is one measurement, not a verdict.
Sources. Eight archived YouTube transcripts and auto-caption tracks, pulled with yt-dlp on 2026-04-24. The Kindred Ventures announcement post (January 2026), the Berkshire Hathaway 2025 annual report, and reported headcount figures, and press coverage of Corgi's 2026 funding rounds. Only public information is used.
Written April 2026. Corgi figures updated September 2026.