The Twenty Minute VCWill Cursor Kill Figma? Lightspeed Raises $9B & OpenAI’s $1B from Disney & #1 App in App Store
CHAPTERS
- 0:00 – 1:47
Lightspeed’s $9B raise: what it really signals about multi-stage power
The group breaks down Lightspeed’s headline $9B raise, arguing the number is split across vehicles but still reflects enormous multi-stage leverage. They discuss why recent early-stage exits and late-stage concentration (e.g., Anthropic) make the LP case work.
- •$9B is across multiple funds; the early-stage portion is much smaller than the headline
- •Why big multi-stage firms can treat seed as an “entry ticket” to later rounds
- •LP success criteria: long-duration early wins + aggressive late-stage concentration
- •How mega outcomes (Databricks/Coinbase-style) can make fund math work
- 1:47 – 6:38
Mega funds vs. seed economics: barbell venture and ‘marketing’ seed checks
They debate whether mega funds destroy seed returns and how competition changes when price sensitivity disappears. The conversation emphasizes barbell outcomes, where huge funds can overpay early because the goal is access to $100B companies.
- •At mega scale, seed price matters less; speed and access matter more
- •Multi-stage can ‘swamp’ seed and write it off as customer acquisition for later rounds
- •Seed manager fundraising appetite is debated—has it cooled since 2021?
- •Large check sizes turn modest multiples (6–7x) into meaningful dollars returned
- 6:38 – 11:17
Late-stage competition and the ‘growth supercycle’ fueled by delayed IPOs
The hosts argue that companies staying private longer has shifted compounding from public markets to private growth investors. They frame this as a generational opportunity for late-stage capital—and a structural change in who captures upside.
- •Fewer IPOs = private investors capture more compounding; retail/public miss the upside
- •Late-stage capital has more ‘places to go’ today than in 2021 (OpenAI/Anthropic soak capital)
- •Comparing Tesla’s early IPO to SpaceX potentially IPO-ing near $1T+
- •Cost of capital explains why founders stay private longer
- 11:17 – 15:08
Disney’s $1B into OpenAI: IP licensing templates and the ‘revenge of IP’
They downshift from the headline number to what the deal structure implies: cross-licensing, experimentation, and precedent-setting. The key debate is whether premium IP becomes more valuable in an AI era after initial “scrape everything” dynamics.
- •$1B is ‘neither here nor there’ relative to OpenAI scale; structure may be round-trippy
- •Disney tests licensing characters for generation while policing unlicensed use elsewhere
- •Deal as a template: content owners negotiating paid access to IP for model providers
- •Strategic lock-in: IP as a retention moat as models commoditize
- 15:08 – 19:36
ChatGPT as the #1 app: slowing user growth, monetization, and becoming a ‘meta app’
They discuss what it means for ChatGPT to top the App Store while growth rates decelerate. The focus shifts to whether OpenAI can expand monetization per user (Robinhood-style) or broaden adoption (Meta-style).
- •User growth slowing is normal at massive scale; key is sustaining engagement and ARPU
- •Constraints: limited ad model today and pricing in global markets
- •Enterprise/backend vs consumer: potential advantage for players like Anthropic
- •Risk framing: high-growth bets are terrifying when they slow because valuation basis flips
- 19:36 – 21:46
OpenAI drops the 1-year cliff: talent wars, exceptions, and easier mobility
The team explains why OpenAI ending cliff vesting is a sign of intense competition for top talent. They note that at today’s valuations, even partial-year equity is life-changing money—so candidates negotiate hard, and companies standardize exceptions.
- •Typical 4-year vesting with 1-year cliff is breaking under AI-lab hiring pressure
- •At high equity values, 12 months can represent $2M+ for individual contributors
- •Removing the cliff can make leaving easier, but also removes recruiting friction
- •Standardizing policy likely followed a ‘gazillion exceptions’ reality
- 21:46 – 27:23
Oracle’s AI-driven CapEx shock: sugar highs, RPO hype, and the CoreWeave parallel
Oracle’s stock drop becomes a case study in how AI infrastructure narratives can reverse once costs surface. Rory frames it as an unwinding of prior exuberance, while Jason argues these ‘weak’ infra players are high-beta and may rebound with AI demand.
- •RPO-driven optimism vs reality of capital intensity to deliver on contracts
- •Oracle building data centers for OpenAI changes margin profile vs legacy cash-flow business
- •CoreWeave/Oracle as ‘marginal providers’ with amplified upside/downside
- •Key question: is this a temporary jitter or a sign the AI CapEx cycle is peaking?
- 27:23 – 30:47
Broadcom’s $300B wipeout: custom AI chips, margin anxiety, and who gets a ‘GM pass’
They unpack why markets punished Broadcom despite large AI demand signals. The discussion distinguishes Nvidia-style branded margins from made-to-order chip economics, and zooms out to the broader market’s obsession with gross margins in AI transitions.
- •Anthropic orders imply custom silicon designed to avoid Nvidia’s premium margins
- •Custom chip business can be strong but structurally lower margin/defensibility than Nvidia
- •Market can re-rate trillion-dollar companies violently on small guidance shifts
- •Confusion over which companies get ‘gross margin passes’ during AI buildouts
- 30:47 – 34:29
Apollo’s ‘0% returns’ warning: entry valuations, 10-year math, and private-market mirages
Rory contextualizes Apollo’s claim as a long-horizon valuation model: high starting P/E correlates with weak 10-year returns, not necessarily short-term declines. He connects this to private-market pricing that may look cheap only relative to inflated public comps.
- •High entry P/E predicts lower 10-year returns; short-term is not reliably predictable
- •Dot-com lesson: markets can rise for years, but decade returns can still be flat
- •Cisco as a cautionary tale: 25 years to regain peak price
- •Private valuations ‘cheap vs public’ may still be high if public markets normalize
- 34:29 – 47:02
Cursor vs Figma: design-and-code convergence and the fight for the unified workflow
The episode’s central product theme: AI is collapsing the boundary between design, prototyping, and production. They explore whether Cursor’s design mode is the first credible threat to Figma and why the long-term winner may be whoever executes with the most intensity.
- •AI drives fast category convergence (design, coding, prototyping collapse into one flow)
- •‘Single agent’ thesis: teams want one shared agent across roles and departments
- •Momentum vs incumbency: Cursor’s speed vs Figma’s installed base and design DNA
- •Coding remains the epicenter of enterprise AI spend (~55% cited), raising the stakes
- 47:02 – 56:38
The biggest danger to incumbents: getting ‘maimed’ by AI rather than killed
Jason introduces ‘maiming’ as the dominant competitive outcome: incumbents keep customers but lose expansion and new-logo growth. They apply this to SaaS and infrastructure players, describing how budget shifts and new AI-native workflows quietly erode growth rates.
- •Maiming pattern: strong logo retention but declining NRR and slower new customer adoption
- •Examples discussed: Atlassian, GitLab, Mongo’s competitive pressure, UiPath’s displacement risk
- •AI shifts CIO priorities—products fall down the funding list even without churn
- •Incumbents must ‘make themselves cool’ again by restoring ~30% growth
- 56:38 – 1:02:23
Boom Supersonic’s surprise pivot: airplane engines as data-center power generation
They react to Boom raising money tied to selling engines/turbines to power data centers. Rory argues it’s less crazy than it sounds—engine tech overlaps with generators—and frames it as a high-risk hard-tech bet that could be rescued by adjacent markets.
- •Boom building both planes and engines is unusually ambitious (Boeing vs Rolls model)
- •Using turbine tech for power generation is a known playbook among engine makers
- •AI/data-center demand creates unexpected ‘budget attachment’ for hard-tech companies
- •Still speculative: no planes/engines sold yet; funding history implies volatility
- 1:02:23 – 1:17:17
SpaceX at $1.5T: the ‘Elon Option Value,’ narrative pricing, and IPO mechanics
They wrestle with how SpaceX could justify $1.5T, concluding traditional multiples don’t work and introducing ‘Elon Option Value’ as the premium for repeated, category-defining execution. The group explores demand drivers, potential strategic anchors (e.g., Google), and the difficulty of selling such a large IPO.
- •Traditional revenue multiples don’t explain $1.5T; ‘EOV’ explains the premium investors pay
- •Starlink as the key growth driver; speculation on future ‘space data centers’ narrative
- •IPO logistics: how much dilution/raise is feasible and who supplies $100B+ of demand
- •Narrative tactics: anchor investments (Google/Nvidia) could create scarcity and momentum
- 1:17:17 – 1:21:56
Would-you-rather: Figma vs Cursor and OpenAI/Anthropic vs Google bets
They close with a rapid-fire investing game that reveals their underlying theses on durability vs momentum. Rory leans into Cursor’s upside and Anthropic’s path to profitability, while Jason favors incumbency stability with Figma and weighs the volatility of early AI leaders.
- •Figma at $17B vs Cursor at $29B: Rory picks Cursor; Jason picks Figma on stability grounds
- •Debate: AI leaders may be less stable than they appear over a 24-month window
- •OpenAI vs Anthropic vs Google: Rory highlights Google’s renewed internal confidence but leans Anthropic
- •Risk-adjusted thinking: profitability trajectory and ability to meet commitments matter
