> ## Content Index
> Fetch the complete content index at: https://www.signalful.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Buy the Incumbent, Then Refound It
- URL: https://www.signalful.com/buy-the-incumbent-then-refound-it/
- Published: 2026-10-03T14:29:00.000Z
- Updated: 2026-10-09T19:40:12.000Z
- Description: Michael Lee left Lone Pine to buy established companies and rebuild them around AI. He does about one deal a year.
- Author: Signalful Editorial
- Tags: Investors, Sequence Holdings, Michael Lee, Private Equity, AI Strategy, #sf-0016, #yt-TCpRwJBQvW0, #Import 2026-10-06 15:50

Michael Lee is co-founder and CEO of Sequence Holdings, a permanent holding company that buys stakes in established businesses and rebuilds them around AI with its own engineering team. He started Sequence about 20 months ago with co-founder Alex, after a career at Goldman Sachs, Apollo, and Lone Pine Capital, where he joined in 2017, covered AI early, and built out the firm's private investing business. Sequence's first investment was a minority stake in Bank South, a family-owned Georgia bank. Its second is a $7.7 billion take-private of insurance broker Baldwin, co-controlled with Michael Dell's family office. These signals are from Michael Lee's interview with Sarah Guo on No Priors.

1. **Uneven AI Impact.** Michael Lee sorts the economy into three buckets. AI barely touches some industries, like restaurants and golf courses. Startups will win others: buying an outsourced coding shop "feels like a terrible idea" next to funding Cognition or Anthropic. Sequence targets the third group, where incumbents hold the advantages (brand, scale, network effects, regulation), on the bet that buying the right incumbent lets you inherit those advantages and build the market leader.
2. **Small Machines on the Line.** Most enterprise AI agents today hand a small machine to every human on the existing assembly line so they work faster. Lee calls that better than nothing. Machines that run 24/7 and scale with electricity call for redesigning how the organization works, and he argues that redesign only happens when someone owns the company and can change it.
3. **The Celebrated Persona.** Every company has one. At Blackstone it's the investor, which is why Blackstone collects the best investors in the world. If you believe returns now come from engineering and AI, the engineer has to hold that spot, and a Fortune 500 CEO trying to refound the company in-house can't recruit or retain that talent. Lee says a typical PE firm would never let a 25-year-old engineer weigh in on an investment, while his engineers expect a say in whether a business is good or bad.
4. **Services Incrementalism.** Accenture, McKinsey, and Palantir do celebrate engineers, but Lee says the incentives break the model. Services firms optimize for getting into your wallet, staying there, and growing their share of it, which he calls "a path towards incrementalism." They also can't change what refounding requires: how people are organized, who works there, and how they're paid.
5. **Software Sells Today's Workflow.** Off-the-shelf software is available to every competitor, so it can't differentiate you. The subtler problem is how software companies get built: they look for a common workflow they can install quickly and deeply enough to stick. That means they always sell to the workflow as humans run it today, never to an organization that doesn't exist yet.
6. **One Deal a Year.** PE fund structures reward deploying capital, and Lee says a typical fund investment is planned around selling in about 3 years. Sequence has no deployment cadence and no LPs telling it it's underinvested. It aims for one deal a year, only in companies it thinks can grow past $100 billion, and Lee says a year with zero deals would be fine.
7. **The Cold Start.** A holding company has a circular problem: no money means no deal, no deal means no engineers, and no engineers means no money. Sequence broke it by taking Bank South on as a customer on August 4 (Lee skipped his anniversary) and working in a services motion through November. The family then asked Sequence to become a permanent partner, and the minority stake closed in March after Fed and OCC approval. Lee says ownership changed everything: engineers took on deeper work they'd have to maintain, and employees pushed for more ambitious projects.
8. **Organizational Physics.** Lee calls the bank's regulation "a feature, not a bug." A regulated institution has well-defined rules and excellent data hygiene, which suits agents. Banks are also dense and centralized: branches spread out, but underwriting runs from headquarters, so anything Sequence builds there applies everywhere. He contrasts that with rollups, where you integrate different systems, send engineers to many sites, and standardize procedures and cultures.
9. **Loans Without Headcount.** At a typical bank, loan volume grows in step with middle and back office headcount. Since March, Sequence has cut average consumer underwriting time by 94% and taken the average loan from 30 days end to end to 11\. When loan volume doubled from Q1 to Q2 (Lee credits luck for the demand), a bank that used to turn business away handled all of it with unchanged standards and a smaller underwriting team, after one person retired and another moved to the front office.
10. **The Atlas Platform.** Sequence's engineers, many from Scale AI and Palantir, found that about 80% of a business breaks down into common units and 20% is specific to its industry. Atlas, built at the bank, has four layers: a data ontology to "make the business legible to the models," an agent builder, an orchestration engine called Lattice, and an application builder called Artifacts. Lee says all of that core infrastructure carries over to Baldwin and future companies.
11. **Brokerage Economics.** Insurance carriers pay brokers a cut of more than $2 trillion in annual premiums. Carriers historically make almost nothing on underwriting and earn their money investing, so they prize high-quality gross written premium. Brokers get paid by the carrier rather than the customer, don't compete on price, and keep about 90% of clients each year. Lee says that relationship business is huge and nearly immune to startups, which makes it ideal for Sequence.
12. **Management Teams Already Moving.** Lee says Sequence has no expertise running a broker or a bank and depends on management teams that already win in their industry. He looks for teams that started changing before it was obvious: moving to the cloud, centralizing data, rolling out OpenAI or Anthropic. Trevor Baldwin pushed Anthropic across the firm early and runs Baldwin on a single instance of Applied Epic, its core system. Lee calls the "human engineering" of employee anxiety and change harder than the software, so he wants partners who have already started it.

Watch the full video at [https://www.youtube.com/watch?v=TCpRwJBQvW0](https://www.youtube.com/watch?v=TCpRwJBQvW0&ref=signalful.com).