The Twenty Minute VCInsights from Coatue's Growth Investor Lucas Swisher
CHAPTERS
- 0:00 – 2:32
AI breaks the SaaS “annuity” narrative and compresses public multiples
Lucas explains why public SaaS is selling off: investors are newly questioning terminal value as AI threatens workflows and moats. Uncertainty about which SaaS names are disrupted causes broad de-risking and multiple compression.
- •AI (especially coding models) makes SaaS durability feel less certain
- •Accounting optics (SBC, GAAP vs non-GAAP) matter more once terminal value is questioned
- •Investors can’t confidently pick winners/losers, so they exit the whole sector
- •Near-term fundamentals are hard to read because earnings are backward-looking
- 2:32 – 4:03
How to judge “thrown out with the bathwater” SaaS names
The conversation moves to practical indicators to separate resilient SaaS from truly threatened businesses. Lucas emphasizes sequential growth and retention signals, while acknowledging the data will lag reality for months.
- •Look for sequential revenue growth and rising net-new ARR
- •Watch retention dynamics as the best near-term signal
- •Recognize bull/bear cases exist for almost every SaaS company now
- •Expect a 3–9 month window where clarity remains limited
- 4:03 – 7:36
Public vs private: owning the future vs owning liquidity
Harry challenges whether beaten-down public comps are better risk-adjusted than mega-priced private rounds. Lucas argues publics offer liquidity, but privates provide access to the fastest-growing ‘future’ companies that stay private longer.
- •Publics can look cheap ‘for a reason’—disruption risk is real
- •The most expensive private deals can be best if growth is explosive
- •To be “levered long” frontier themes (OpenAI/Anthropic/SpaceX), you need privates
- •Rise of massive private ‘platform companies’ limits public investor access
- 7:36 – 9:23
Revenue durability in an architecture shift: bet on reinvention, not just growth
Lucas reframes durability: in major platform transitions, prior winners can evaporate unless they repeatedly reinvent. He uses Databricks as the archetype of hopping multiple S-curves and expanding its role in the enterprise.
- •Architecture shifts (on-prem→SaaS, internet→mobile, now AI) reset moats
- •Key trait: talent density and willingness to reinvent repeatedly
- •Databricks example: multiple reinventions across product waves
- •The goal is to find companies that can create Act 2/3/4, not just early growth
- 9:23 – 11:16
Valuing exponential AI companies: “price matters least” and valuation comes last
They discuss how traditional valuation anchors break when revenue is compounding extremely fast. Lucas says Coatue prioritizes identifying the curve and treats valuation as the last question—within limits.
- •Explosive growth can make an ‘insane’ entry price look cheap quickly
- •Framework: evaluate valuation last when growth is exponential
- •There is a price boundary, but preemptive rounds help set a rational price
- •Litmus test: if it executes, would you invest more at a higher price?
- 11:16 – 19:22
TAM vs founder quality: big markets first, founders enable multi-act expansion
Harry presses on whether market size beats founder quality. Lucas argues both matter, but market size is the first gate—founders are crucial for expanding into multiple TAMs and surviving successive waves.
- •Market size is the first principle for high-valuation growth investing
- •Great founders in small, constrained TAMs struggle to reach mega outcomes
- •Platform companies ‘skip TAMs’ via multi-product expansion over time
- •Enduring companies require Act 2/3/4 expansion, not one wedge forever
- 19:22 – 22:51
Concentration strategy: few bets, big checks, and the power of double-down rounds
Lucas explains Coatue’s preference for concentrated ownership and aggressive scaling into winners. He cites the skew of value creation (a small number of companies drive most outcomes) and why ‘spray and pray’ fails at large scale.
- •Private market value creation is extremely concentrated (few companies drive most EV)
- •Strategy: avoid being in the wrong horse; make fewer, higher-conviction bets
- •Double-down rounds are often the best rounds if the company keeps compounding
- •Competitive overlap is harder early; later-stage platform exposure can resemble public-style portfolios
- 22:51 – 27:01
Mega-fund math: when $5B+ growth funds can still work (and when venture can’t)
Lucas separates venture from growth economics. He argues large venture funds struggle to get enough ownership in the few breakout outcomes, while large growth funds can work because companies stay private longer and rounds can absorb huge checks.
- •A multi-billion venture fund faces difficult ownership and outcome capture math
- •Growth funds can deploy very large amounts into late private rounds
- •If you can put $1B into a winner and 10x it, it can move a $5B fund
- •AI-era outcomes may be larger than SaaS-era outcomes, improving scalability
- 27:01 – 29:54
What returns are “enough” at growth: why a 3x isn’t exciting
Harry asks what upside is compelling at high entry prices. Lucas explains fund math: misses require larger winners, and you need a believable path where future buyers can also underwrite upside—often via public market demand.
- •To deliver ~3x net fund return, losses imply needing 5–6x+ winners
- •A 3x outcome is insufficient unless there’s a path to re-rate and compound again
- •You must believe a public-market buyer will want the stock later
- •Big idea + durable path to liquidity are essential underwriting criteria
- 29:54 – 32:37
When double-downs go wrong: overestimating TAM and multi-product ability
Lucas identifies the primary failure mode: believing a company can expand into multiple markets and products when it can’t. Metrics and team quality are less often the issue than incorrectly sizing the long-term market and expansion potential.
- •Common mistake: TAM overestimation
- •Second mistake: overestimating multi-product expansion success
- •Coatue’s style tends to reduce loss ratio but still fails on expansion theses
- •The bar has risen for what qualifies as a ‘platform company’ bet
- 32:37 – 37:11
Margin in AI: misleading early, critical at scale; focus on retention and cost curves
They tackle whether gross margin still matters for AI-native businesses with heavy inference costs. Lucas argues margin matters at scale, early gross margin can mislead during architecture shifts, and retention becomes essential when margins are thin.
- •Gross margin can be ugly early in new infrastructure/platform waves (hyperscalers, Snowflake/Databricks)
- •AI inference costs should decline; mix of frontier and smaller models can improve margins
- •Lower gross margins may be offset by lower opex and higher operating margins
- •If margins are low early, retention must be extremely strong to avoid fragility
- 37:11 – 40:11
Why seed is harder now: mega-fund behavior, bigger checks, and ballooning valuations
Harry argues mega funds can ‘destroy’ seed economics; Lucas agrees seed has become harder. Larger early rounds, higher valuations, and greater capital intensity raise risk and reduce early-stage ownership for traditional seed funds.
- •Mega funds entering early with founder-friendly terms changes competition
- •Bigger early checks and valuations reduce seed ownership opportunities
- •AI-era companies can be more capital intensive than classic SaaS
- •Flexible mandates help investors avoid being forced into overheated segments
- 40:11 – 45:23
Kingmaking, capital advantage, and why platforms still go public
Lucas rejects ‘kingmaking’ as deterministic, while acknowledging capital can help in specific conditions (PMF + scaling). He also outlines why even dominant private platforms may still choose to go public: liquidity at scale, feedback, and defensibility.
- •Capital is an advantage with strong PMF; can be a disadvantage without it
- •Kingmaking isn’t ‘game over’—great businesses can emerge despite crowded incumbents
- •Public markets provide true liquidity and cleaner cap tables than layered SPVs
- •Public listing offers feedback loops and makes companies harder to ‘mess with’
- 45:23 – 53:40
Canva as a platform case study + lessons from Mary Meeker and Mamoon Hamid
Harry challenges whether Canva is truly a platform given AI image generation and competitive pressure. Lucas argues Canva’s multi-product expansion and early AI adoption prove platform behavior, then shares key lessons from Mary Meeker (data storytelling) and Mamoon Hamid (spotting inflection points).
- •Canva’s strength: hopping S-curves and building a suite of fast-growing products
- •Early AI integration as a strategic advantage
- •Mary Meeker: tell stories with data; analytical rigor matters
- •Mamoon Hamid: identify company ‘kinks’/inflection points early (e.g., Figma retention/usage at top customers)
- 53:40 – 1:06:36
OpenAI vs Anthropic, memorable founder meeting, and career reflections
In quick-fire, Lucas compares OpenAI’s consumer/enterprise vectors and ‘unknown unknowns’ (hardware, new surfaces) with Anthropic’s coding beachhead and multi-cloud/multi-chip optionality. He closes with his most memorable founder meeting (Harvey), a key career decision, a major miss (Anduril), and excitement about new AI products.
- •OpenAI: consumer franchise, enterprise expansion, and potential new device surface area
- •Anthropic: coding focus as wedge; multi-cloud/chip strategy enables capacity and partner support
- •Most memorable meeting: Winston at Harvey—clear founder-market fit and thesis early
- •Career lesson: get off the linear path; biggest miss: not seeing Anduril’s trend/founding strength; excitement: new AI products and devices
