Re-Founding Incumbents for the AI Era with Sequence Holdings Co-Founder and CEO Michael Lee

Summary

Michael Lee explains Sequence Holdings’ thesis: use a permanent holding-company structure, frontier engineering, and long-term ownership to transform incumbents into AI-era market leaders. He contrasts this with traditional PE and software vendors, arguing that most enterprise AI efforts are too incremental and that real change requires reorganizing the business around engineers and agents, not just giving workers better tools. He details how Sequence’s first operating test in a Georgia bank validated the model, citing centralized operations, good data hygiene, and measurable productivity gains. Lee also describes the company’s broader investing framework, its partnership with the Dell family office on Baldwin, and his belief that exceptional people and durable ownership matter more than ideas alone.

Notes

• Sequence is structured as a permanent holding company, not a standard fund.
• Core goal: refound incumbents into AI-era market leaders.
• Sequence was founded around frontier engineering applied to partner companies.
• Lee says the firm expects to do only one deal per year.
• He frames engineering as the celebrated persona inside the company.
• He traces his AI interest to Lone Pine in 2017, around AlphaGo and the first transformer paper.
• He says early architectures degraded as they scaled.
• ChatGPT convinced him AI was a durable scaling shift.
• He expects AI’s economic impact to be uneven across industries.
• Sequence prefers incumbents with structural advantages like brand, scale, network effects, or regulation.
• The idea emerged in early 2023 and became actionable in 2024.
• He argues AI should reorganize organizations, not just speed up individual workers.
• He says the holding-company form best aligns capital duration with operational transformation.
• He wants the engineer to be the celebrated persona, like at Palantir.
• He says services firms create incentives problems for deep refounding.
• Sequence’s model is built around ownership, engineering, and long-term alignment.
• Engineers are attracted by working on businesses that touch everyday economic life.
• Application engineers like the idea that better models expand Sequence’s toolkit rather than threaten the business.
• Sequence initially solved its cold start by first becoming a customer.
• A Georgia bank became Sequence’s first customer and then first investment.
• He says regulated banks are a good agentic pilot because their rules and data are legible.
• Banking is centralized, so tooling can be amortized across branches.
• He describes brokerage as a large, relationship-driven market with strong retention.
• Atlas is Sequence’s reusable platform intended to generalize across industries.
• Atlas has four layers: ontology, agent builder, orchestration engine, and app builder.
• He says Atlas speeds deployment, improves agent performance, and improves engineer throughput.
• The bank taught Sequence that employee anxiety must be managed explicitly.
• Traditional PE is optimized for capital deployment and exit timing.
• He says PE firms struggle to recruit engineers and support long AI timelines.
• Sequence deliberately avoids a deployment cadence and transacts sparsely.
• Good management teams are already using cloud, centralized data, and internal AI tools.
• Baldwin is an example of an early AI-adopting CEO and a single-platform operating model.
• At the bank, consumer underwriting time fell by 94%.
• Commercial loan processing dropped from 30 days to 11 days.
• The bank handled doubled Q2 loan volume without loosening standards.
• Lee says the practical effect was better role allocation across staff.
• He says founder life is lonely, hard to finance, and emotionally volatile.
• Take-private deals require certainty of capital and a committed partner.
• He says his career combined deal mechanics, public-company study, and frontier-tech investing.
• He says ideas are cheap and execution is hard.
• He argues exceptional people on hard problems in large markets are the key variable.
• He cites Jensen Huang as a model for repeated business refounding.