AcquiredComplexity Investing & Semiconductors with NZS Capital (Extended Cut)
CHAPTERS
Why NZS Capital (and this episode) happened: from TSMC research to a live meetup
Ben and David explain how an NZS interview became one of their key sources for the Acquired TSMC episode—and then how they serendipitously met Brinton Johns in person at Capital Camp. That chance meeting sparked this extended conversation that starts with investing frameworks before semiconductors.
Sponsor segment: Banco Inter and SoftBank Latin America Fund
João Menin shares Banco Inter’s transformation from a local wholesale bank into a digital-first retail bank and then a Brazilian “super app.” The segment highlights competing with dominant incumbents via a free, full-service mobile account and expanding into non-banking products.
Complexity theory as an investing operating system (Santa Fe Institute origins)
NZS describes how repeated investor “pain” led them to complexity theory and the Santa Fe Institute’s work on complex adaptive systems. The key takeaway: emergent behavior makes precise prediction unreliable, so investors should prioritize adaptability and heuristics over forecasts.
The ant colony lesson: resilience beats productivity (and why hypergrowth can be a trap)
Deborah Gordon’s ant research becomes a metaphor for organizations: many ants appear “idle,” but the colony is optimized for longevity and resilience, not short-term output. NZS connects this to founder-led long-term thinking and argues durable compounding often outperforms sprint-style hypergrowth.
NZS’s barbell: defining resilience vs. optionality (and the “root MOS” hybrid)
NZS lays out its portfolio construction: a concentrated resilience “head” plus a diversified optionality “tail,” with a deliberate gap between them. They also describe “root MOS” companies—resilient cores with out-of-the-money options that justify larger positions.
Non-zero-sum investing: beyond ‘moats’ and toward shared value
NZS challenges the common interpretation of moats as customer lock-in for aggressive value extraction. They prefer companies that create more value than they take (non-zero-sumness), arguing it builds goodwill, reduces backlash/regulatory risk, and increases long-term durability.
Power laws, non-ergodicity, and why traditional risk models break
The discussion shifts from Gaussian assumptions to power-law realities in markets and nature. Using Ole Peters’ coin-flip example, they show how ensemble averages can be misleading in non-ergodic systems—many go bankrupt while a few become massive winners—undermining standard risk/utility frameworks.
Culture and process: avoiding ‘conviction,’ documenting theses, and trimming risk
NZS reframes “conviction” as a synonym for overconfidence and emphasizes team-based bias detection. They describe writing down theses and “key leverage points,” expecting failure in optionality bets, and actively managing position sizes to prevent large drawdowns from torpedoing the portfolio.
Valuation in a prediction-forcing market (and why semiconductors feel like ‘10/10 multiverse’)
In an environment of elevated asset prices, NZS explains that expensive valuations force narrow predictions, which they try to avoid. They lean toward areas where the future feels highly probable—like the semiconductor “building blocks” of the information age—shifting the portfolio toward resilience when appropriate.
Semiconductor portfolio map: catalog semis, leading-edge, tools, and ‘silicon carbide’ optionality
NZS outlines how they invest across the semiconductor value chain, with roughly a third of the portfolio in semis. They distinguish catalog/analog/microcontroller businesses (resilient, long-lived parts) from higher-valuation digital names (more optional) and highlight silicon carbide as an asymmetric growth bet.
Case study: why NZS owns Cadence (EDA duopoly, management quality, and industry-channel research)
Jon explains how trade-show research led them to EDA and why they favored Cadence over Synopsys. The decision centered on leadership (Lip-Bu Tan), cultural turnaround, perceived share gains, and the advantages of learning from industry conferences rather than investor roadshows.
Advanced semis: Samsung vs TSMC, memory vs logic, and Taiwan geopolitical tail risk
The conversation clarifies Samsung’s dominance in memory (DRAM/NAND) versus TSMC’s leadership in logic manufacturing and its trusted neutral foundry model. They discuss how Taiwan risk is real but not isolated—TSMC disruption would reverberate globally—and note potential second-order effects (equipment rebuild boom).
Moore’s Law today: EUV, gate-all-around, chiplets, packaging, and system-level scaling
Jon argues Moore’s Law remains alive in spirit through the mid-2030s, driven by EUV scaling plus innovations beyond transistor shrink. The future is increasingly about advanced packaging, chiplets, new transistor architectures, and system-level compute (clusters/interconnects) rather than a single monolithic die.
The semiconductor ‘superorganism’: ecosystem dependence, ASML’s origin story, and the virtue of growth governors
They describe semiconductors as an interdependent ecosystem where many layers capture high margins, indicating a healthy non-zero-sum “superorganism.” Stories include ASML’s customers funding EUV development and the idea that extreme complexity creates natural growth governors—often enabling long-duration compounding rather than hypergrowth.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome