Will Cursor Kill Figma? Lightspeed Raises $9B & OpenAI’s $1B from Disney & #1 App in App Store

Summary

• Lightspeed’s $9B raise confirms seed pricing is now irrelevant to multi-stage firms. Harry’s backward math: only ~$2B is venture/early, ~$7B growth — and Jason’s read is that a firm at that scale “doesn’t care what you pay for seed,” can “swamp a seed business… and just write it off as marketing,” which is why $20-30M pre-seed rounds exist. Harry’s LP checklist explains the raise: early wins that went the distance (Rubrik, likely Netskope) plus ~$1B concentrated into two Anthropic rounds — “tick, tick, insert $9 billion.”
• The no-IPO era is “the greatest gift to venture in our lifetimes.” Tesla had to IPO at $1.7B; SpaceX is rumored to pursue a $1.5 trillion IPO — “a thousand-x more,” with all of that compounding kept private. Harry’s conclusion: “It’s not just being in OpenAI. It’s playing the growth super cycle bet today. That’s the winning play.” The mechanism: relative cost of capital decides where companies live, and right now public capital feels more expensive than private.
• Jason’s core warning on OpenAI: “There is nothing as terrifying as a high growth bet that slows down.” OpenAI’s mobile-app growth is down to single digits (“empirically true on any source” — Rory), and a private company that shifts from growth valuation to cash-flow valuation leaves someone holding the bag — the class of ‘21 paid 20-30x revenue at 100% growth and now grows 8%. Monetization is still working, “but if that were to taper off, it would be a world of pain.”
• Oracle (-45% from September) and CoreWeave (-60% from July) are the “weak guys” — and therefore the purest marginal AI bet. One guest called the RPO pop a “total sugar high on a huge contract with someone who may or may not be able to afford to pay for it”; the frame is they’re “the marginal provider of the commodity” — buy Google for 20%, buy CoreWeave for a possible double. Broadcom’s $300B two-day wipeout makes sense too: Anthropic’s $21B order exists precisely to avoid Nvidia’s 75% gross margins.
• Macro overhang: Apollo’s 10-year zero-return call is the chart Rory trusts most — entry PE barely predicts year one but strongly predicts year ten. “Cisco, the darling of ‘99, just got back this week to its ‘99 stock price” — 25 years. The realistic path isn’t a flat decade but a 30% drop in the next 2-3 years, which would leave private valuations comped to all-time-high public markets “high and dry.”
• Convergence is the 2026 theme, and coding is the epicenter: 55% of enterprise AI end-user spend. Jason: marketing/sales/support have already collapsed into one agent in e-commerce; design and code are next, and “Figma feels behind.” The incumbent risk isn’t death — it’s being maimed: retention holds, seats shrink, NRR drifts, new cohorts never buy. UiPath is the archetype: 98% GRR, 107% NRR (from ~140% at IPO), and 6-24 months to ship agentic products — “30% growth, that’s what’s cool.”
• SpaceX at $1.5T only works with the “EOV — the Elon option value.” Harry: ~$156B of revenue means 70-80x 2026 sales; “you can’t run the numbers on SpaceX and come up with the 1.5 trillion. You just can’t” — you’re buying the only founder since Jobs to do it three times. Jason thinks space-based data centers are the telegraphed next act and sketches how the IPO gets done: Google anchors $10B, Nvidia $5B, Fidelity $2B — “the round is sold out.” On pure plays, Rory takes Anthropic at $360B over OpenAI: “sensibly and boringly converging on profitability.”

Deep dive

1. Lightspeed’s $9B: seed is now a marketing line item

• Harry’s backward math on the headline: roughly $2B for venture and early, ~$7B across other vehicles, mostly growth — “the 9 billion is a bit misleading.” His provocation stands anyway: if you’re not playing the big game, do you really matter?
• Jason’s answer — “bad,” in air quotes, for seed VCs: “It really means you don’t care what you pay for seed. It just doesn’t matter… you just got to get into one hundred-billion outcome.” That’s why $20-30M pre-seed rounds exist, and it deepens venture’s barbell. Rory agrees he can’t quantify it but concedes a multi-stage firm “can swamp a seed business to some extent and just write it off as marketing” — an acquisition cost for A/B/C access.
• Harry’s LP framing of why they earned it: a 2025 multi-stage manager must show early funds that went the distance — Rubrik last year, likely Netskope this year — and concentrated late-stage picks, “by all accounts a billion in two rounds of Anthropic,” which “feels pretty smart right now.” “Tick, tick, insert $9 billion.” The nuance: one of those early wins returned only 6-7x aggregate — the real insight was they got $200M+ to work.
• Harry cites David George: Databricks 7x’d a $1B a16z growth fund and Coinbase 5x’d it — a 15x billion-dollar fund. But on his surprise at LP appetite for $50-100M SF seed funds, Rory pushes back hard: “I think that’s a myth… I see that appetite having faded since 2021. They want to find a 20VC or a NEO… I don’t think it’s easy for emerging managers.” Add Dragoneer’s (likely) $4.3B venture fund and it’s $13B+ of fresh late-stage firepower in a week.

2. Not IPOing is the greatest gift to venture — Tesla vs SpaceX is the proof

• Jason, crediting Rory’s earlier call: “All these leaders not IPOing is the greatest gift of venture in our lifetimes.” VCs keep the top-20 companies’ compounding for themselves, and unlike prior super cycles, growth is the big beneficiary — “It’s playing the growth super cycle bet today. That’s the winning play.”
• Rory’s dry gloss on the consumer-protection origin: the good news is retail was protected from losing 1x their money; the bad news is “they’ve left the entire compounding of Databricks, of SpaceX, of Anthropic, of OpenAI on the table” — while late-stage firms got “early stage venture economics on masses of money.”
• The cleanest natural experiment, same founder: Tesla IPO’d at $1.7B; SpaceX is rumored to pursue a $1.5 trillion IPO — “a thousand-x more.” Anyone who bought Tesla after 2010 got roughly a 70% compound return for 15 years; “that product was not available” for SpaceX or Databricks. Harry’s mechanism: the choice is relative cost of capital — Musk went public after what he recounts as predatory VC behavior made private money too expensive; today the public market’s cost of capital feels higher, “so everyone’s staying here.”
• On whether late-stage is dangerously competitive: 2021 “turned out to be treacherous” as growth attenuated, but one blessing today — OpenAI and Anthropic “can soak up 60 billion of your late-stage dollars and just keep on moving.”

3. OpenAI weekend: Disney’s IP template, single-digit growth, and the death of the cliff

• On Disney’s $1B investment, the two disagree. Jason: it’s a cross-licensing deal, “very round-trippy,” experimental — “interesting, but no up.” Harry: it’s the template for the next era beyond “just ripping everybody’s content off” — a 3-year deal where first movers get better terms and “we’re going to ratchet up the rates” at renewal. “It may be the revenge of IP.” Iger’s line, quoted approvingly: “creativity is the new productivity.”
• ChatGPT was 2025’s most-downloaded US app (the last decade’s winners: two years of TikTok, two of Temu, Zoom in 2020). Rory’s harder point: OpenAI’s mobile-app growth is “down to single digits” monthly — “we’ve run out of humans on planet earth” — so the question is whether it Robinhoods into a meta-app (Robinhood is on fire despite 8% new-customer growth) or stagnates around 1-1.2B users. Jason’s doubt: with no ads and only the free-to-$20/month conversion to cross-sell, “I’m not sure everyone on the planet wants to do complex AI lookups.”
• Jason’s warning, the episode’s sharpest line: “There is nothing as terrifying as a high growth bet that slows down” — you go from being valued on growth to being valued on cash flow, “and you really would not want that to happen while you’re still private.” The class of ‘21 paid 20-30x revenues at 100% growth and now grows 8%, barely clearing the last round price. He’s not sounding the alarm — monetization still works — “but if that were to taper off, it would be a world of pain.”
• The one-year vesting cliff ending is a hiring-war artifact: on a $10M four-year package, the 11/48ths at month twelve is ~$2M — real money worth pushing back on. [Speaker?]: it makes leaving easier — “perhaps you are” creating mercenaries — but “they probably had to make so many exceptions it stops mattering.”
• The pure-play would-you-rather lands here too: OpenAI at $500B, Anthropic at $360B, or Google at $2T. Jason takes Google — the Cloud team he works with is the most energized he’s seen in a decade, and it clocks north of $100B profit at ~20x EBITDA. Rory takes Anthropic: “way more sensible than OpenAI… sensibly and boringly converging on profitability, will go public and will be a very nice public company,” while OpenAI is “more likely to just get caught in the middle with commitments it can’t meet.”

4. Oracle’s sugar high — the weak guys are the high-octane bet

• Oracle fell 15% Friday, -45% from September highs, on $12B of quarterly capex vs $8.2B expected, mostly data centers for OpenAI. The told-you-so: the 30% RPO pop “was absurd and it’s just been unwound” — “a total sugar high on a huge contract with someone who may or may not be able to afford to pay for it,” in a capital-intensive business worse than the free-cash-flow core. “I should have bought those puts.”
• The rebound case won’t call the top: Oracle and CoreWeave (-60% from its July high) “are the weak guys — they don’t really own anything themselves,” at high risk of margin compression, so they should take the biggest hits from jitters. But “we’re still anti-gravity here” — no reason they can’t rebound with the trend. The synthesis: they’re “the highly amplified bet because they are the marginal provider of the commodity” — buy Google and get 20%; buy CoreWeave and maybe get a double. The whole call reduces to whether the capex cycle has two more strong years.
• Harry’s middle scenario, offered with self-awareness that nobody rewards nuance: the market is interrogating who has a plan — “Gemini, Google, you keep going… OpenAI, we’re good for it, we’ll give you $40 billion. Facebook, not so in love. CoreWeave and Oracle — it’s not clear to me why you’re doing it.” Plausible 2026: Google’s capex rises while marginal players retrench — “neither the end of the world nor a rebound.”

5. Broadcom’s $300B lesson: who gets the gross-margin pass?

• Broadcom lost $300B of market cap in 48 hours over the $21B Anthropic order, and the concern is framed as rational: Anthropic designs custom silicon precisely because “they don’t want to pay 75% gross margins to Nvidia” — “if I’m going to pay full retail, I might as well go buy the designer brand.” Broadcom is a made-to-order business — “this is meant to be Kohl’s here, dude” (likely) — good coin, but never Nvidia’s defensibility of imposing architecture on customers.
• Perspective on the “crash”: Broadcom is still worth $1.6T at a high-teens sales multiple — the first trillion-dollar company was Apple in 2018. “It’s not like everything went cheap, Harry. It’s just slightly less expensive.”
• Jason’s unresolved puzzle — worth keeping precisely because he admits he can’t solve it: “Who gets a pass on gross margins and who doesn’t? Oracle got this great pass until it didn’t. CoreWeave gets apparently an entire pass. OpenAI does, Meta doesn’t. I can’t keep up.” Palantir taught him some deserve the pass; the market will teach who doesn’t.

6. Apollo’s zero and Cisco’s 25 years — the entry-price warning

• Rory insists on precision: Apollo predicted a ten-year return of zero, not a crash — and it’s built on the chart he watches most (Vanguard sends it too): entry PE barely correlates with one-year returns, correlates at five, and is strongest at ten. Greenspan called irrational exuberance in ‘96 and stocks ran three more years; buy in ‘99 and you went zero-for-ten. The kicker: “Cisco, the darling of ‘99, just got back this week to its ‘99 stock price” — 25 years to earn it back.
• What to do with a warning that isn’t short-term predictive: don’t go binary, adjust allocation — “there’s a reason that nice Mr. Buffett has piled up $300 billion in cash, cuz he reads these data too.”
• The venture meta-point: private valuations “look attractive relative to a public market that’s at an all-time high.” And zero-for-ten never arrives smoothly — “what tends to happen is at one point in the next two or three years things drop 30%, and then you crawl back slowly over the rest of the decade.” If that happens, privates comped off today’s marks “could feel lofty.”

7. Convergence: everyone wants to talk to the same agent

• Cursor’s new designer tool (Harry initially credits Anthropic; Jason corrects — it’s Cursor) triggers Jason’s 2026-27 thesis: “massive convergence of categories.” It already happened in e-commerce — marketing, sales, and support “have already converged to one agent,” which is why Klaviyo ($1.3B growing 30%) brought in a Workday exec so the CEO could return to product full-time. And the design-code gap is visible everywhere: “30% of the last YC class looked vibe-coded to me — I could see the Claude artifacts all over their homepage.”
• The deeper insight, which Roy calls “just a huge insight”: “We all want to talk to the same agent — designers, product people, engineers, DevOps.” Roy’s extension: silos existed because humans were siloed; if you sell AI software to automate an outcome rather than work, you sell one agent that prospects, sells, and supports — “the single view of the customer, on the customer’s side.”
• On who wins the collapsed design-to-production platform: “Figma feels behind… tiptoeing into vibe coding just like Canva,” and Jason’s honest non-answer is “who wants it the most — everyone can copy each other in weeks now, not months or quarters.” The momentum bet favors Cursor: Figma took a decade to $1B, Cursor a year. Cursor skipped the Replit/Lovable market deliberately — those two will do ~$500M combined in a year; Cursor got there in 9 months with ~160% net revenue retention because “nobody leaves Cursor.”
• The sizing data (Menlo, via Rory): coding is ~55% of all enterprise AI end-user spend — “this is the epicenter of the enterprise AI revolution.” Roy’s bigger nut: ~$15-16B of app spend plus ~$15B of infrastructure, versus $400B spent making AI — enterprises must find not $15B but $150B of budget, “otherwise the people investing in capex are going to have a sad day.”

8. You don’t get killed — you get maimed

• Jason’s frame for incumbent risk, the episode’s most reusable idea: Cursor probably won’t displace Figma — “the bigger risk is that it maims Figma. The old customers don’t leave… they renew. They don’t buy as many seats… NRR drifts down, and the new customers — the kids from YC — defer that purchase.” He sees it across his older portfolio, and “a lot of founders aren’t being honest about how they’re being maimed.”
• The roll call: Atlassian slowing, GitLab arguably maimed, and MongoDB (likely — the name is bleeped, but the Postgres/Supabase competitor context fits) fought back yet “should be growing 50% — the explosion of apps we’re building today is unprecedented. Why isn’t it?” Rory’s mechanism: CIO surveys — as AI climbs the priority list, your item slides from three to six “and then it doesn’t get funded.” Co-attach to AI budgets or live a slow-growth secular story.
• UiPath as the archetype: not hemorrhaging — back to 16% growth, stock up 27% this year — but RPA got replaced by agents. Harry’s math for Daniel: $1.8B ARR, 98% GRR, 107% NRR (down from ~140% at IPO) means he has time — “we’re in the first inning for AI B2B… you’ve got 6, 12, 18, 24 months to roll out high-ROI agentic products and you’re back to 30% growth.” Databricks runs 150% NRR at $5B ARR.
• Jason accepts Alex Rampell’s line — can the incumbent acquire innovation before the startup acquires distribution — but says it hides the grind: lifting growth from 9% to 11% while “the stock doesn’t give a [bleep],” pushing “that Sisyphean rock up the hill for four or five years.” The market won’t grant coolness for inventing the magic thing: “You got to make yourself cool, cuz 30% growth — that’s what’s cool.”

9. Boom Supersonic: engines on trailers and the whiplash of hard tech

• Boom takes a data-center power order from Crusoe and raises $300M on it. Jason, delighted (“anyone who chooses anything other than the supersonic plane has no soul”), says the pivot isn’t crazy: GE and Rolls-Royce all sell both, because most of a jet engine is a generator — “it’s a lot easier to take an engine and plop it on the ground and have it generate electricity than put the same damn thing in a plane.” Boom is designing both plane and engines — “the full enchilada” — with a prototype built at under 100 headcount.
• Jason supplies the whiplash arc: an “arguably fake” $1B round from airlines with no revenue in December 2024, a crash to maybe $500M with a cramdown, “and now you’re back up to $1.5B” on the AI lift. The sobering fact: “They’ve sold zero of either — no planes and no jet turbines.” Don’t lump it in with SaaS reacceleration stories; it’s an ultra-high-risk, ultra-ambitious profile.
• The hard-tech tie-forward: when hard-tech bets work, “you have a wildly compelling business” — the largest IPO in history is about to be a rocket company — “but as the Boom plane part of the story makes clear, they’re damn hard problems.”

10. SpaceX at $1.5T: pricing the Elon Option Value

• The chronology humbled Harry: last Tuesday he called the $800B secondary rich; Wednesday, before the episode even aired, the $1.5T IPO leak — “you can be wrong by 800 billion in a day.” The numbers don’t close: roughly $156B of revenue (Starlink the driver, growth maybe early-mid-20s next year) puts it at 70-80x 2026 revenues — so he coined the EOV, the Elon Option Value: value the business normally, and the gap to the print is EOV. Tesla on a normal multiple might be $300B; the rest is EOV. “I’ve let go of valuation.”
• He’s earned it, Harry argues, because “he pulled the Starlink rabbit out of the hat” — a rocket company that became a communications company — making Musk the only founder since Jobs (Apple, Pixar, Apple again) to do it three times, plus founding OpenAI. The tail risk stays: “If that premium ever evaporates — if you were ever to die — the stock gap would be something horrific.”
• Jason’s read on the next act: Musk telegraphing space-based data centers ahead of the IPO is deliberate — if the Oracle/CoreWeave/Nebius revenue “all goes up in the sky… he’s the only one that can build data centers in space. And I don’t think he’s joking.” Mechanics are daunting either way: a $30B raise is only 2% dilution; a typical 8% IPO means finding $120B of raw risk capital that wants 70x run-rate revenue. Jason’s banker scenario: Google (already ~10% holder) anchors with $10B for TPUs in space, Nvidia adds $5B, Fidelity $2B — “the round is sold out.” Harry: or the valuation “gets walked back to the merely outrageous.”
• Harry’s coda insists on the lesson: Peter Thiel fired Musk from PayPal but fully vested his stock and walked him out with respect — so when SpaceX was dying, Founders Fund wired money “in an hour,” and now owns ~10% of possibly the biggest IPO ever. “In today’s age of extreme greed — be kind. It might pay off.” Jason’s caveat, kept as hedged: maybe less kindness than Musk being a big paid-in stockholder — but the PayPal talent density and the rational handling of the split are “about to pay off to the tune of a hundred billion dollars.” And the final would-you-rather splits them: Figma at $17B or Cursor at $29B — Rory takes Cursor without argument; Jason takes Figma, because “we’re going to look back in 24 months and see there was a lot less stability in these so-called leaders than we thought.”