All podcasts / Lenny's Podcast / Summary

How Anthropic, Costco, and Patagonia all build incorruptible companies | Eric Ries

2026-05-10 - 99 min - source - Read full transcript
Lenny Rachitsky (host)Eric Ries

Key insights

The thing that destroys great companies is almost never competition - it's their own success becoming a liability that invites extraction.
Eric argues that the more valuable a company becomes ('the more gold in the goose'), the greater the temptation for investors, boards, or acquirers to butcher it for short-term value. He illustrates this with restaurants degraded after private-equity buyouts and consumer brands (Vital Farms) whose quality reportedly dropped after ownership changes pushed for higher growth and margins.
corporate-governance
Mission protections must be put in place before a company needs them, because success itself removes the leverage to negotiate for them later.
Eric describes watching founders get told at every stage - incorporation, seed, growth round, IPO prep - that mission-protective provisions can wait until 'later,' until it becomes provably too late during IPO filing. His framing: 'it is always too early until it's too late,' and success makes you a target rather than protecting you.
founder-protection
Standard venture-backed governance overwhelmingly ends with founders losing control of their own companies.
Citing Harvard Law School data, Eric says only about 20% of founders at venture-backed companies with standard best-practice governance documents are still CEO three years after going public. He recounts a founder who was ousted five months after a successful IPO despite the company's business fundamentals not meaningfully changing in that window.
founder-protection
Under 'any lawful purpose' charters and shareholder-primacy law, boards can have a legal fiduciary duty to accept the highest acquisition bid regardless of mission fit.
Eric traces shareholder primacy to roughly the last 40 years, contrasting it with 19th-century corporate law that required companies to declare a publicly beneficial purpose and let boards fight off unwanted buyers. The Vectura Corporation case - an inhaler-therapeutics maker whose board felt legally bound to sell to Philip Morris despite public and medical-society outrage, followed by a $900M write-down and dissolution within three years - is his central proof point.
corporate-governance
Public Benefit Corporation status is a near-costless legal filing that gives a company legal cover to reject profit-maximizing pressure, and it's already the norm among major AI labs.
Eric describes PBC status as a two-page Delaware filing that replaces the default 'any lawful purpose' language with a specific stated purpose, so a board can point to an actual legal duty to something other than shareholder return. He says nearly all major AI labs, including Anthropic, are incorporated this way, and calls it 'the one thing that has no tradeoffs at all.'
ai-company-structure
'Harder is easier': absorbing the upfront cost of a principled decision compounds into trust that becomes a durable competitive advantage.
Cloudflare's decision to give away SSL encryption for free - despite it being their most profitable upgrade driver - required rebuilding cryptographic infrastructure in-house and initially hurt premium conversion, but Eric credits it with driving an order-of-magnitude increase in top-of-funnel growth and underpinning the trust behind Cloudflare's $70B valuation.
trust-and-culture
Companies governed by a dedicated 'mission guardian' structure - separate from any single founder's control - are measurably more durable and profitable than conventional peers.
Eric cites academic research showing that companies with nonprofit-foundation or purpose-trust ownership structures (like Novo Nordisk and Zeiss) are six times more likely to survive to year 50 and have superior return on invested capital compared to conventionally governed companies.
corporate-governance
Anthropic's Long-Term Benefit Trust - AI-safety experts with board-appointment power but no equity - is the structural reason it can make costly mission-aligned decisions without being overruled.
Eric, who advised Anthropic's founders early on to write these rights into the charter from inception (implemented fully at their Series C), argues this is why Anthropic could refuse to ship a model it judged too dangerous or turn down a $200M Pentagon contract: the trustees overseeing mission alignment have no financial incentive tied to growth, unlike a standard board that can be pressured or replaced.
ai-company-structure
The 'culture bank' rule - only make deposits, never intentional withdrawals - is what turns principled behavior into internalized culture rather than a rule that has to be enforced top-down.
Drawing on Todd Park (Devoted Health) and Howard Schultz (Starbucks), Eric contrasts this with organizations that treat doing-the-right-thing as an occasional act of individual courage; in a strong culture, employees like an HEB store manager giving away groceries during a blackout don't see it as heroic, just as what the company trains everyone to do by default.
trust-and-culture
Founder control is a viable early-stage bridge but a fragile long-term structure, because it makes the mission dependent on one person's endurance.
Eric notes founders with concentrated control frequently experience public mental health breakdowns from carrying the entire mission alone, and recommends transitioning to a durable, renewable structure (nonprofit foundation, perpetual purpose trust, employee ownership trust) rather than relying indefinitely on personal control, which has no mechanism to 'grow back' once eroded.
founder-protection
AI alignment and organizational alignment share the same unsolved root problem: whose values get encoded, and Conway's Law shows those values become structurally embedded either way.
Eric argues that because organizational values flow from parent to child the way an org chart shows up in software architecture, technical AI alignment is downstream of a harder, unaddressed human-alignment problem - 'who aligns the aligners.' He uses an ant-colony 'piano movers puzzle' study as physical evidence that organizations exhibit genuine emergent intelligence, meaning ungoverned org structure produces unintended emergent behavior the same way an unaligned model would.
ai-company-structure
Founders can take three low-cost, no-tradeoff actions immediately: file as a PBC, write and adversarially stress-test a mission statement, and reserve future mission-protective rights in the charter even before formal structures exist.
Eric's concrete homework for early-stage founders: file Public Benefit Corporation status; workshop a mission statement by brainstorming ways to profit while violating it and rewriting until none survive; and write reserved rights into the charter now (e.g., 'pledge 10% of equity to a future foundation') so the option exists later without having to build the full structure immediately - describing this as 'fire and forget.'
founder-protection

Books referenced

Media referenced

Companies

Techniques and frameworks

Summary

Eric Ries, author of The Lean Startup, returns to Lenny's Podcast to discuss his new book, Incorruptible: Why Good Companies Go Bad and How Great Companies Stay Great. Where Lean Startup was about building a successful company, Incorruptible is about protecting one after it's built. Eric's core claim is that the force that ruins great companies is rarely competition - it's the temptation created by their own success, which invites investors, boards, and acquirers to extract value at the expense of the mission. He traces this through vivid case studies: a private-equity-owned restaurant he could "taste the ownership structure" in, Vital Farms' reported quality decline after a change in ownership, and most centrally, the UK inhaler company Vectura, whose board felt legally bound by fiduciary duty to sell to Philip Morris despite public outrage, only for the company to be written down and dissolved within three years.

The conversation's structural core is Eric's "ethos plus integrity" formula. Ethos is the specific, adversarially-tested purpose a company commits to (Cloudflare's "make a better internet," Devoted Health's "treat every customer like your own parents"). Integrity is the structural apparatus that makes betraying that purpose actually costly - not a slogan, but an enforced mechanism. He walks through the toolkit for building that apparatus: Public Benefit Corporation status (a near-costless Delaware filing nearly all major AI labs use), a "mission guardian" entity that owns responsibility for keeping the company mission-locked, and what he calls "spiritual holding companies" - nonprofit foundations (Novo Nordisk, founded on this structure in the 1920s and still governed by it), perpetual purpose trusts with a dedicated "purpose protector" (Patagonia), employee ownership trusts (John Lewis Partnership), or employee voting trusts (Alibaba). Academic research he cites shows companies with these structures are six times more likely to survive to year 50 and post superior returns on invested capital than conventionally governed peers.

Anthropic is the episode's throughline for how this plays out in practice. Eric played a small early advisory role, encouraging Dario Amodei and the founding team - then an unproven, non-hot startup with no major VC backing - to write mission-protective rights into their charter from day one, which they later implemented fully as the Long-Term Benefit Trust at their Series C. Because that trust's AI-safety-expert trustees hold no equity, Eric argues, Anthropic can refuse to ship a model it judges unsafe or turn down a $200M Pentagon contract without a standard investor-controlled board being able to force the decision. He extends this into a broader "harder is easier" principle, illustrated at length through Cloudflare's decision to give away SSL encryption for free despite it being their top revenue driver - a costly, mission-consistent choice CEO Matthew Prince backed with "let's figure it out," which Eric credits with the trust behind Cloudflare's later $70B valuation.

The back half of the conversation moves from governance to culture. Eric introduces the "culture bank" (via Todd Park and Howard Schultz): organizations build trust through deposits (principled sacrifices) and lose it through withdrawals (self-interested shortcuts), and in strong cultures doing the right thing - like an H-E-B manager giving away groceries during a blackout - isn't treated as individual heroism but as the trained default. He revives Mary Parker Follett's 1920s concept of the "invisible leader": the internalized common purpose that governs decisions made when no manager is present, which he argues is the actual mechanism by which mission survives at scale. He closes by connecting this to AI alignment directly - organizational values flow into architecture the same way Conway's Law predicts, and an ant-colony "piano movers puzzle" study demonstrates that organizations, like transformer models, exhibit genuine emergent intelligence that goes uncontrolled without deliberate structure.

Eric ends with concrete, low-cost homework for early founders: file for Public Benefit Corporation status, adversarially stress-test a mission statement by trying to find profitable ways to violate it, adopt a "director's oath" for board members, and write mission-protective rights (like a future equity pledge to a foundation) into the charter now even without building the full structure immediately - framing all of it as "fire and forget" optionality that costs nothing to reserve and everything to lack later.

Notable Quotes

"Their very success became a liability, because the more gold in the goose, the greater the temptation to butcher." - Eric Ries

"It is always too early until it's too late." - Eric Ries

"If you don't get this right, no other decision you make about your company will matter for the long term, because you're not going to be the one making it." - Eric Ries

"Trustworthiness is the most underrated asset in all of business. And the things that create trustworthiness, by definition, stack rank to the bottom if we do it by ROI. Because doing the right thing has intangible rewards, but tangible costs." - Eric Ries

"Only make deposits. Never make withdrawals." - Eric Ries