The Twenty Minute VCElon’s Empire: SpaceX, Tesla, Neuralink After the Storm & Anduril’s $2.6BN Power Move
CHAPTERS
- 0:37 – 4:30
Circle’s blockbuster IPO and the “money left on the table” problem
The group dissects Circle’s outsized first-day performance and why celebratory IPO headlines quickly turn into frustration about underpricing. They emphasize how secondary-heavy IPOs make the “pop” far more painful for selling shareholders than for companies raising mostly primary capital.
- •Circle’s IPO jumped ~2x+, skipping straight from “IPO window is shut” to “we underpriced it”
- •Secondary sellers can lose billions in upside when pricing is too low
- •A modest IPO pop is structurally necessary to compensate buyers for volatility
- •Meme-like demand can coexist with fundamentally strong businesses
- •The real issue: how to price when retail demand is hard to predict
- 4:30 – 10:15
Why IPO pricing is structurally messy (and why alternatives haven’t fixed it)
Rory explains the core incentives and information asymmetries in book-built IPOs, where bankers and anchor investors heavily shape outcomes. The hosts debate why oversubscription metrics are unreliable and why direct listings, SPACs, and auctions haven’t become universal replacements.
- •Oversubscription figures (5x/10x/30x) are game-theory artifacts, not true demand
- •ECM bankers have repeated-game advantage vs. one-time issuer decision-making
- •Anchors vs. flippers: issuers are told certain investors ‘won’t flip’—often incorrectly
- •Direct listings work best for exceptional companies not needing primary capital
- •SPACs largely failed; Dutch auctions worked in limited cases (e.g., Google)
- 10:15 – 12:57
Does the IPO rebound pull in juggernauts (Stripe, Databricks) and what Figma’s filing signals
They argue the ‘window’ is effectively always open for the very top tier; the limiting factor is whether leaders like Stripe and Databricks want the trade-offs of being public. Figma is framed as a compelling candidate because of its prior near-liquidity event and the desire to cement a definitive win.
- •‘Room at the top’: $50B+ IPO candidates face different constraints than $2B IPOs
- •Stripe/Databricks staying private is a choice, not a market-access issue
- •Figma’s aborted Adobe deal makes an IPO psychologically and strategically appealing
- •Timing is inherently random; companies should prepare and then wait for opportunity
- •Strong IPO tapes can create internal pressure to ‘go now’ even if not essential
- 12:57 – 15:55
CoreWeave as a case study: public markets de-risk companies fast
CoreWeave’s IPO is used to show how quickly sentiment can flip and how public markets can unlock large-scale financing. They highlight how a successful IPO can eliminate existential risks (like looming debt) and reinforce the U.S. market’s capital formation advantage.
- •CoreWeave cut its range pre-IPO, then traded ~2x+ soon after—valuation is sentiment-driven
- •IPO success enabled additional debt issuance to remove major risk factors
- •U.S. public markets can supply huge capital quickly and efficiently
- •Being public can be the most scalable way to fund capital-intensive growth
- •A humility check: even experts miss obvious upside opportunities
- 15:55 – 19:21
Why the U.S. stock market dominates—and why firms leave London for New York
Rory lays out a macro argument for U.S. market dominance using population/GDP/market-cap stats, then applies it to Wise and Deliveroo’s moves. The key takeaway is that liquidity, analyst coverage, and investor depth make the U.S. the default venue for global tech and fintech listings.
- •U.S.: ~4% of population, ~23–24% of GDP, ~67% of global market cap
- •Wise popped on U.S. listing news—implied valuation uplift from deeper liquidity
- •For international businesses, the U.S. is the most liquid, most buyer-rich market
- •Smaller tech IPOs can be thinly traded even in the U.S.; outside the U.S. can be worse
- •Analyst coverage and institutional participation drop sharply in smaller markets
- 19:21 – 20:33
Should $2B–$5B companies be public? Employee liquidity vs. private-market constraints
They debate whether smaller public companies make sense given thin liquidity, but note private liquidity is often worse and more controlled by management. The discussion shifts to employee equity: long holding periods and unreliable tender offers weaken the traditional startup value proposition.
- •Even thin public liquidity can beat private-market liquidity for many stakeholders
- •Employees prefer daily tradability over sporadic, management-approved tenders
- •Most unicorn equity becomes effectively ‘notional’ without IPO/tender access
- •Longer time-to-liquidity (4 years becoming 12) creates retention and morale strain
- •Only a small elite can run reliable, large-scale tender offer programs
- 20:33 – 25:04
Unicorn outcomes: a large share will fail, many will merge, few will IPO
Jason and Rory outline a sobering distribution for unicorns: meaningful failure risk, a mid-tier that won’t reach public markets, and a minority that can IPO. The implication is that the ecosystem must normalize mergers, PE outcomes, and write-offs rather than assuming IPOs for all.
- •~1,500 unicorns; estimates suggest ~20% may outright fail
- •Rory’s distribution: ~20% IPO-quality, ~50% mid-tier (M&A/PE), ~20–30% tail-risk
- •Tender offers at scale are available to only a tiny fraction of companies
- •The longer companies stay private, the more acute liquidity needs become
- •Equity compensation becomes less compelling without credible liquidity paths
- 25:04 – 29:44
Anduril’s $2.6B round and Founders Fund’s $1B bet: concentration, thesis, and mission
They unpack why Founders Fund repeatedly writes enormous checks into Anduril, tying it to a decades-long national security thesis and willingness to concentrate. Rory contrasts this with his firm’s concentration limits and explains why larger funds often force later-stage concentration by necessity.
- •Founders Fund’s largest checks ever have been to Anduril (again)
- •A 20-year national security investing thesis (Palantir roots) culminates in Anduril
- •Purpose/motivation: national defense mission aligns with capital allocation
- •Rory’s firm caps concentration at ~10%; typical portfolio aims for many smaller bets
- •Bigger funds often require more concentration because fewer late-stage outsized outcomes exist
- 29:44 – 38:13
‘If LPs let you go wild’: discipline vs. freedom and the myth of endless follow-ons
The hosts debate what investors would do without constraints—massive concentration in generational winners vs. maintaining strategy. Rory argues that ‘just double down on winners’ is mathematically harder than it sounds because true compounding outliers are rare and late-stage entry multiples compress.
- •Rory favors sticking to what the firm does best; widening aperture increases execution risk
- •Jason’s ideal: invest from seed through every round ‘up to $1B’ in true winners
- •Rory’s math: only a small fraction of deals become late-stage ‘stuffable’ outliers
- •Late-stage ‘winners’ from early rounds often become only 3–4x opportunities later
- •Many firms solve this by expanding the number of early bets to find rare mega-compounders
- 38:13 – 42:09
DocuSign hindsight and how to think honestly about follow-on decisions
They discuss the common regret of not investing more in eventual winners, using DocuSign as an example. Rory emphasizes intellectual honesty: you must compare the winner to other deals that looked equally promising at the time but did not pan out, and recognize the limits of foresight.
- •Regret is easy in hindsight; the real question is decision quality under uncertainty
- •You must evaluate what else looked ‘as good’ when you declined to double down
- •Late-stage valuation shifts can make ‘stuffing’ capital value-destructive
- •Harry’s lesson: be visionary but relentlessly honest about TAM expansion
- •‘21 inflated late-stage signals; the decade-long reality is harsher
- 42:09 – 48:28
Growth-rate nuance: hypergrowth isn’t the only path to huge outcomes
A tactical debate emerges around a vertical SaaS example growing from $1M to $7M over three years and whether it can be venture-backed. They explore evidence that top-quartile early growth is not perfectly correlated with the biggest outcomes, and that second-quartile growers can compound into giants.
- •Rory: slow growth can be a great business but not fit a VC return model
- •Jason: at the right price and with strong founders/TAM, slower ramps can still be worth it
- •Rory’s portfolio analysis: little correlation between 1st vs. 2nd quartile growth and outliers
- •Examples: Bill.com and HubSpot as compounding winners without being the fastest initially
- •Takeaway: clear minimum growth threshold matters more than being #1 in growth rate
- 48:28 – 55:14
The SaaS spending slowdown: maturity, consolidation, and AI’s budget pull
They analyze data suggesting H1’25 SaaS spending growth is slowing again and argue it’s consistent with a maturing category approaching saturation. AI both competes for attention and budgets and may (or may not) expand TAM by automating labor, creating a contentious debate about where the dollars come from.
- •SaaS maturity: as cloud/SaaS penetration rises, sustaining high compounding becomes impossible
- •Examples of TAM saturation dynamics: Zoom, Salesforce, DocuSign-era categories
- •AI creates CIO ‘brownie points,’ shifting projects away from legacy SaaS migrations
- •Key question: is AI additive (labor-dollar replacement) or mostly a knife fight for software spend?
- •Consolidation and bundling become dominant behaviors in mature SaaS markets
- 55:14 – 1:03:30
AI economics in practice: contact centers, pricing power, and the TAM expansion debate
Jason argues AI is rapidly displacing labor in contact centers but warns vendors may not capture proportional value—replacing a $50k worker doesn’t automatically yield $50k of software ACV. Rory counters with a ‘2-for-1 arbitrage’ model, suggesting meaningful but not total TAM expansion where savings are large and quantifiable.
- •Observed: significant headcount reduction in support/contact center functions
- •Concern: ACV uplift may lag labor savings, limiting TAM expansion for vendors
- •Rory’s framework: adoption happens when automation delivers ~2x cost advantage (robotics analogy)
- •Macro view: contact center labor spend dwarfs software spend; partial labor capture could 2–3x TAM
- •Pricing power likely higher in enterprise than SMB due to larger savings pools
- 1:03:30 – 1:11:17
Elon’s empire after the storm: reputational drag vs. business fundamentals
They assess Elon-linked companies strictly as businesses, arguing the recent public episode was predictably damaging but likely transient. SpaceX/Starlink are framed as ‘must-buy’ vendors for governments due to irreplaceable capability, while Tesla is more exposed to consumer sentiment and subsidy/regulatory shifts.
- •Best-case from a company perspective would have been avoiding the episode entirely
- •SpaceX/Starlink: monopoly-like strategic value—customers may dislike Elon but still must buy
- •Tesla: more vulnerable due to consumer brand effects and policy/subsidy risk
- •News cycles move fast; long-term impact may fade, though Tesla data already showed softness
- •They highlight Elon’s extraordinary track record across Tesla, SpaceX, Neuralink, and beyond
- 1:11:17 – 1:22:58
Kalshi quick-fire: Google CEO odds, NYT vs OpenAI, and X leadership questions
In a rapid prediction round, they debate whether Sundar Pichai will leave, the likely outcome of the New York Times lawsuit against OpenAI, and whether X’s CEO will be replaced. The discussion widens into how AI will force new definitions of fair use, content value, and licensing dynamics.
- •Pichai leaving: they lean ‘no’—boards resist CEO churn during existential transitions
- •NYT lawsuit: likely settlement-like ‘win’ for NYT; Supreme Court path is uncertain
- •Content licensing game theory: LLMs may not need every outlet—marginal publishers may have less leverage
- •Rory’s point: one or two major sources may suffice for ‘modern news’ needs
- •X CEO: mixed views; external performance suggests vulnerability, but internal value is unclear
