holocron
中/en
← Business
Business

Founder Succession

In one sentence: Every great company eventually has to hand the keys over, and there's no single way to do it well — some founders retire and get pulled back in against their will (Morris Chang, Sam Walton), some hand off through decades of apprenticeship that barely register as drama (Costco), some turn succession itself into a decade-long governance project that's still unresolved (Berkshire), and one hands off with a single parting instruction that a successor ends up defending with his life (Nintendo). Six real handoffs, laid side by side, refuse to collapse into one tidy lesson.

The Six Handoffs

TSMC: Morris Chang's Two Exits, Two Returns

Morris Chang is the only person on this page who completed a full retire–return–retire-again cycle, twice over. In 2005, at 74, he retired for the first time, handing the CEO title to longtime deputy Rick Tsai and staying on only as chairman; operations were split into two groups, advanced (under Mark Liu) and mainstream (under C.C. Wei). Morris later gave his own account of what "letting go" was supposed to mean: "Sometimes you have to let the CEO make his own mistakes and learn from them. Of course, not if the whole company is going down the drain. You have to interfere then, but only then." (Morris Chang 访谈：张忠谋亲述 TSMC)

The 2008–09 financial crisis put that principle to a real test. Under Rick Tsai, TSMC laid off 600–700 employees through disguised "worst performance rating" cuts, triggering two rounds of protests outside Morris's own house — 50–60 police against over 100 protesters, roughly 25 of whom camped overnight in a nearby park, prompting his wife Sophie to bring soy milk and youtiao at dawn. Morris's own summary was blunt: "This ... promoted me to take back the CEO position." He returned as CEO at 78 in 2009, spending nearly all of his first four to five weeks resolving the NVIDIA 40nm dispute (his predecessor had offered zero compensation), then merged the two operations groups, fixed R&D spending at 8% of revenue, tripled capex to bet on 28nm, and four years later landed the Apple foundry deal (Morris Chang 访谈：张忠谋亲述 TSMC TSMC：纯代工模式发明者).

He didn't handle the deposed CEO as a casualty. Rick Tsai kept his rank, salary, and bonus, was moved into a new solar/LED business unit, and was told to his face he remained a candidate for the top job — even though that new venture eventually failed to take off and Tsai resigned, later becoming vice chairman and CEO of MediaTek (Morris Chang 访谈：张忠谋亲述 TSMC). In 2013, once Apple was locked in, Morris stepped down as CEO a second time while remaining chairman, then retired fully in June 2018 at 86 — no third act this time. The groundwork for that had been laid years earlier: after the 2009 merger, Morris offered the BD role to Mark Liu, who managed 10,000 people, and was refused ("You want me to go run a department of sixty or seventy people?"); he turned instead to C.C. Wei, who "accepted it even delightfully" — the same person the source material names as TSMC's current chairman and CEO (Morris Chang 访谈：张忠谋亲述 TSMC).

Walmart: Sam Walton Handed Off, Then Took It Back

Walmart's succession story splits into one failure and one that worked. In 1966, Sam Walton personally attended an IBM seminar on business computing and recruited Ron Mayer there (Walmart：小镇折扣店如何长成世界最大公司). According to the entity page (Tier-0-sourced detail not present in the episode extract itself, with no precise date given), Sam later promoted Ron Mayer to CEO but soon realized he "simply wasn't ready to retire," and took the power back; in Made in America he's generous toward Ron, blaming himself rather than his protégé. On air, David and Ben draw an explicit parallel to Morris Chang's "retired and got called back" arc at TSMC (Sam Walton（山姆·沃尔顿）).

The second handoff stuck. Before his final illness, Sam handed the CEO title cleanly to David Glass; by 1992, when he died of cancer, Glass had already been running the company. The show's verdict on this post-founder era is telling — the early-90s Supercenter breakthrough got graded A+ by David before Ben corrected it with data to "a textbook decade of decline," but both agree on one thing: it wasn't Sam's move, and the single best play in Walmart's history landed after the founder had already let go. Ben draws the same comparison to Ted and Todd buying Apple at Berkshire (Walmart：小镇折扣店如何长成世界最大公司).

More than any individual, what Sam actually left behind was a structural insurance policy: the 1953 family partnership (Walton Enterprises), set up on his father-in-law's advice, which locked in that no family member could sell shares individually — 60 years later the Waltons still hold a majority stake. Sam's own words on the matter read almost like a curse: "If you start any of that foolishness like changing the structure, selling off stock, going off and doing fancy things — I will come back and haunt you." Ben's on-air aside — "like buying NBA and NFL teams" — landed because, by the time of recording, a third generation of Waltons had in fact already started doing exactly that (Walmart：小镇折扣店如何长成世界最大公司).

Costco: Three Generations, One 22-Year Apprenticeship

There's no retire-and-return drama in Costco's line — just a nearly straight line of transmission. Jim Sinegal started as a part-time bagger at FedMart around 1955 and worked for Sol Price for 22 years, rising to run the entire distribution and central-warehousing operation. Asked whether he'd "learned a lot" from Sol, his answer corrected the premise of the question: "No, that's inaccurate. I didn't learn a lot, I learned everything." In 1983 he and Jeff Brotman founded Costco in Seattle, with a business plan that was "basically a straight clone of Price Club" (Costco：会员费买来的信任机器).

The 1993 merger with Price Club forced a real choice on Sol Price about who should run the combined company — the material is explicit that he'd rather hand it to his protégé Sinegal than sell to Walmart. Sinegal became CEO of the merged company. The next handoff went to Craig Jelinek, another FedMart lifer who started as an hourly employee as a teenager; the exact year of transition isn't given in the mirror, and working backward puts it around 2012–13 — imprecise, flagged rather than asserted. Three CEOs — Sol Price, Jim Sinegal, Craig Jelinek — all trace back to FedMart, a chain the show calls "stacked learning," "the same people, the same story" (Costco：会员费买来的信任机器).

The evidence for cultural transmission is concrete. The habit of visiting every store every year passed unchanged from Sinegal to Jelinek. The hot-dog vow is the best-known artifact: when Jelinek, as the new CEO, reported that the hot-dog-and-drink combo's margin was breaking even and suggested a price increase, Sinegal's answer was "If you raise the price of the hot dog and drink combo, I will effing kill you." Costco eventually built its own production line to cut costs instead of raising the price — using manufacturing capability to make good on what started as a joke. One small, precise detail: a decade into the CEO job, Jelinek's LinkedIn title still reads EVP (Costco：会员费买来的信任机器).

Nintendo: One Parting Instruction, One Successor Who Died Defending It

The protagonist of Nintendo's story isn't the founder who retired — it's the successor who took the instruction literally and held the line at real personal cost. In 2002, Hiroshi Yamauchi handed the company to Satoru Iwata, leaving behind a single line that defined the next two decades: "Nintendo give birth to wholly new ideas and create hardware which reflects that ideal and make software that adheres to that same standard." (Nintendo 之二：主机战争（1989-2023）)

The real test came in the early 2010s: Wii U had flopped, the company was posting multi-year losses, and shareholders were pushing Nintendo to exit hardware the way Sega had and put its IP straight onto phones — which would almost certainly have been profitable in the short run. Iwata refused, in the line that carries the most weight in this whole theme: "Making smartphone games is absolutely not under consideration. If we did this, Nintendo would cease to be Nintendo... It probably would be the correct decision in the sense that the moment we started to release games on smartphones, we'd make profits. However, I believe my responsibility is not to short term profits, but to Nintendo's mid- and long-term competitive strength." He and Shigeru Miyamoto secretly drew up a three-point plan in 2013 — embrace mobile on Nintendo's own limited terms, extend the IP everywhere, and use the time bought by both to rebuild the hardware strategy from scratch. The eventual product was the Switch (Nintendo 之二：主机战争（1989-2023）).

Iwata died of cancer in 2015 at 55 — he never lived to see the Switch's 2017 turnaround. The episode pins its whole thesis to this arc: "the seeds of success are sown in a fall, and the seeds of a fall are sown in success." The source material doesn't name whoever succeeded Iwata himself; the narrative closes at his death (Nintendo 之二：主机战争（1989-2023）).

Berkshire: Succession as a Decade-Long Project, Still Unfinished

Berkshire is the one story on this page with no founder-retirement beat at all — Buffett's own definition of retirement is a joke that's also true: "(I'll retire) about five years after I'm dead." (Berkshire 之三：现代帝国与接班（1992-2021）) Because of that, succession at Berkshire was never a single handoff moment; it was a piece of institutional engineering built over more than a decade. In the 2006 annual letter he publicly advertised for an investment successor: "I intend to hire a younger man or woman with the potential to manage a very large portfolio, who we hope will succeed me as Berkshire's chief investment officer." Todd Combs, who ran the $100M fund Castle Point, joined in 2010; Ted Weschler, who won his way in by bidding a combined $5.2M across two charity lunches, joined in 2011. Both started managing around $1B and were given more capital as performance justified it, reaching close to $20B each by the time of recording (Berkshire 之三：现代帝国与接班（1992-2021）).

The mechanism also had a real failure it had to correct. In 2010, the Wall Street Journal ran a front-page story calling Li Lu's succession "a foregone conclusion" — Charlie Munger himself used those words — but Li Lu never joined; the story is that the economics of giving up his own fund's 2-and-20 didn't pencil out, an echo of Buffett's own 1956 refusal to take over Graham-Newman for the same reason. The CEO-side "envelope" mechanism was tested for real in 2011: David Sokol, the presumed heir apparent, was caught front-running his own recommendation that Berkshire acquire Lubrizol, left the company once it came out, and the envelope was swapped seamlessly for his former deputy, Greg Abel (Berkshire 之三：现代帝国与接班（1992-2021）).

In January 2018, Greg Abel (non-insurance) and Ajit Jain (insurance) were promoted to vice chairmen together, turning succession from a secret in an envelope into a semi-public fact; at the 2021 shareholder meeting, Charlie Munger's slip — "Greg will preserve the culture" — outed the successor in front of the whole room. But the episode is explicit that this thread is still open: what used to be one brain switching between acquisitions, insurance, and public markets is now four fiefdoms — Greg, Ajit, Ted, and Todd — each compensated mostly against their own portfolio. David's question is whether a non-founder CEO without Warren's authority can make a snap billion-dollar call the way Warren could; his closing jab is that "the number one qualification for the next CEO is 'large-company operating experience' — by that standard, Warren himself wouldn't qualify." (Recorded 2021-06-06; time-capsule note: as of that date, the question remains open.) (Berkshire 之三：现代帝国与接班（1992-2021）)

Patterns Across the Six

What triggers a retire-then-return. Morris Chang and Sam Walton are the only two "retired and got pulled back" cases here, and the triggers aren't the same shape. Morris's return was compound and external — a financial crisis, a succession widely seen as a failure (margin warnings plus mishandled layoffs that sparked two rounds of protest), and the mobile-and-cloud wave crystallizing at once — layered on his own read that "golden opportunities" had appeared. Sam's clawback of power was almost purely psychological: in his own telling, the fault was entirely his, not Ron Mayer's — he simply wasn't ready. What Costco, Nintendo, and Berkshire have in common by contrast is that none of them has a "return" chapter at all: the first two never took power back once it was handed off, and Berkshire has no return because Buffett never actually left. That absence of a road back home is itself worth noticing — it tends to correlate with a handoff whose timing and choice of successor were tested more thoroughly the first time.

How successors get chosen. The most durable pattern here is extreme-duration apprenticeship paired with real performance testing: Sinegal spent 22 years under Sol Price, Jelinek started as an hourly FedMart teenager, Todd and Ted started with roughly $1B portfolios at Berkshire and were scaled up toward $20B as results justified it, and TSMC rotated Mark Liu and C.C. Wei through their own operations groups before making a call. The fragile pattern is the outside hire promoted too fast: Ron Mayer, recruited straight out of an IBM seminar and pushed into the CEO seat, turned out not to earn Sam's real confidence. Character veto shows up as a genuinely load-bearing part of these mechanisms, not decoration — Berkshire's "envelope" was swapped out for cause when David Sokol's front-running surfaced, proving the check actually functions; TSMC's handling of Rick Tsai shows the more forgiving flip side of the same principle — a lapse in judgment isn't a character failure, and gets a soft landing, a kept rank, and an honest word that the door isn't fully closed.

Culture versus strategic continuity. This is where the six stories diverge the most. Costco's transmission is the softest and also the deepest — the hot-dog pricing vow, the annual store-walk habit, the written order of the Code of Ethics, all copied nearly verbatim through decades of master-apprentice cohabitation. Nintendo is the clearest case of culture outweighing short-term profit: Iwata treated "Nintendo would cease to be Nintendo" as a real guardrail, not a slogan, and gave up certain mobile-gaming profit to protect it, at the cost of never living to see the bet vindicated. TSMC's continuity is more institutionalized cognition than soft culture — R&D fixed at 8% of revenue, the pure-play foundry model, "customer trust as a third competitive axis" — rules that can be written down and don't depend on any one person's temperament. Walmart's continuity runs mostly through governance structure rather than culture itself: the family partnership preserved concentrated ownership, but Sam's near-curse of a warning was already showing cracks by the third generation buying sports franchises. Berkshire goes furthest in the opposite direction, explicitly rejecting the myth of a "sacred culture" — what can actually be codified and passed down comes to just three head-office rules (never risk the company's reputation, minimize and defer taxes relentlessly, sweep all cash to headquarters for reallocation), with everything else fully decentralized. That may be the least inspiring answer of the six, but it might also be the most honest one about how much culture can actually survive a handoff.

Tensions and Counter-Examples

None of these successions was painless, and this page isn't going to pretend otherwise. Sam Walton's clawback from Ron Mayer is the only failure explicitly owned by the founder himself in his own words, and even with all the blame absorbed personally, the outcome was still a real reversal — it took handing off to David Glass, years later, before it actually worked. Morris Chang's own 2005 transition was likewise widely judged a failure by outsiders: a man who built his entire career on "designing a company to fit its constraints" tripped over the same problem when designing his own company's succession, and had to come back and clean it up himself in 2009. The two deposed executives fared differently afterward — Rick Tsai got a dignified soft landing and an explicit statement that he remained a candidate; Ron Mayer's aftermath isn't detailed in the source material beyond the fact that the power was taken back.

Nintendo's case carries the heaviest cost and is the hardest to turn into a reusable lesson: Iwata protected the company's soul, but the verdict that he was right only arrived after he died of cancer in 2015, never seeing the Switch succeed in 2017. That's a real caution against treating "culture over short-term profit" as an obviously correct move in the moment — it's often only provable after the fact, whether it was foresight or stubbornness. The source material itself supplies a sober footnote here, in Ben's words: "if history didn't prove him right... we also wouldn't be doing this episode." Berkshire remains the one genuinely open case: recorded in June 2021, just after Munger's slip let the successor's name out of the envelope, the show's stated worry — four fiefdoms instead of one brain — is left unresolved on air, not answered.

Costco looks, on the surface, like the one frictionless exception among the six, but it's worth being precise about why. Its succession was never blood inheritance — Sinegal wasn't Sol Price's son, and Jelinek wasn't Sinegal's son. It's a "spiritual kinship" path built on 22 years of shared work, distinct from Walmart's family-blood governance and just as distinct from Ron Mayer's fast-tracked outside hire. Put the six stories side by side and what actually stands out isn't a single correct path so much as a handful of recurring failure modes (premature promotion, psychological unreadiness, judgment lapses under crisis) and a handful of recurring insurance mechanisms (extreme-duration apprenticeship, character veto, structural governance design). Which combination survives the next crisis is exactly the question Berkshire's still-unfinished story leaves honestly open.