Eventbrite Sold for $500M, Databricks $5B Raise at $134B Valuation & Why SaaS is Like Japan

Summary

• Databricks at $134B is the episode’s worked example of paying for growth: 32x on $4.1B of 2025 sales growing 55%, against Snowflake public at the same ~$4B revenue, 28% growth and 20x. Rory’s frame — “how much extra in multiple do you pay for how much extra in growth” — has no public comp set, since “there is literally only one public company growing more than 30%” (likely Palantir, at 50%, 80x sales). The kicker is reacceleration at scale: “if it continues to re-accelerate, just for the record, it’s infinitely valuable, because that’s just what the math says” — the same dynamic behind Anthropic’s step-function repricing this year.
• The majority of public SaaS is in a “TAM trap”, Jason’s coinage with average public SaaS growth at ~16% — “we’ve never grown more slowly.” Rory’s answer to how the Aaron Levies and Drew Houstons missed it: they didn’t — venture simply saturated every market and its adjacencies (“Hunger Games in SaaS”). The investing rule that falls out: “overpayment only works when the TAM is huge. In finite TAMs, you’ve got to bid more tightly.”
• Takeout season is on: Eventbrite sold for ~$500M at 1.5x revenue despite a 50% premium, PagerDuty sits near 2x ($1B on $500M ARR growing 4%), Semrush went to Adobe. The bull case is that smart PE money finds 2x revenue “stupidly cheap” and bolts on AI — but Jason notes “we haven’t seen pagerduty.com and pagerduty.ai magically mashed together into a winner yet.”
• Security is the revenge of the enterprise: after Gainsight’s two-week Salesforce lockout and the Drift breach (700 orgs’ data downloaded, with ransom demands including Cloudflare; “Drift will never come back. It is dead”), Jason argues security structurally favors incumbents — and both concede it’s also “the best excuse there is” to cut off data-slurping startups and sell your own agent instead.
• “SaaS has become like Japan” — ARR per employee only rises (HubSpot 2.8x more efficient than 2021, Salesforce 2x, Microsoft “permanently past peak employee”), so seat-based pricing shrinks organically: “if everyone only has 0.9 kids… there’s only so many seats to go around.” Rory’s twist on why multiples haven’t fallen with growth: companies flipped from twice as inefficient at the margin in 2021 to twice as efficient now.
• Model providers may focus on coding: OpenAI’s code red is read as an admission to focus the core — “win the ChatGPT wars and you are worth $2 trillion” — so verticals like AI wealth management may stay open. But clone speed is brutal: Google shipped a Lovable/Replit competitor in under 10 months (albeit “no database, no auth”) — “you don’t get five years anymore… now you don’t even get a year.”
• AI wealth management is the pitch both bulls converge on: the gap between Wealthfront’s 10 bps and Goldman/Morgan Stanley’s do-nothing 1% is enormous, the average American retires with ~$1.8M, and if AI does trusts, taxes and estate under one roof, Jason sees a “$20-40-50 billion company” — with the caveat that $8-10K pricing against $30-50K incumbents is not 10x headroom.
• Venture has become “a relevance game,” per Harry: 40% of Q1’s first-time unicorns already raised a follow-on, and LPs are seduced by momentum. Rory’s counter-rule: filter on certainty of a big outcome — “you’re only as accurate as your least accurate variable” — and hold the slow compounders, because Schwab outlived every tech IPO class of 1983.

Deep dive

1. OpenAI’s code red resets the story — Thrive is power law on steroids

• Rory’s zoom-out on the Thrive–OpenAI partnership: “that’s not the OpenAI story anymore… the OpenAI story today is almost the exact opposite” — a code red focus on the core product, pushing on ads, pushing agents-for-healthcare back, “no distractions.” The turnabout in one line: “Google did a code red three years ago on them and now they’re doing a code red back.”
• Jason’s read on the deal mechanics: only a couple of deals ever matter to a VC, so going as deep as humanly possible with your one or two winners is “power law on steroids” — Thrive is effectively turning the fund into a holding company and doubling down on the best company it ever backed.
• Rory on the asymmetry: great for Thrive — “the best marketing you can do is get as close as possible to your biggest deals,” and Sam is “pretty darn loyal” after Thrive put money into the ~$70B round and stood by him “in the great fiasco of two years ago.” The advantage to OpenAI is “much less clear”: “I know who was ecstatic when that was announced and who was like, ‘Yeah, whatever.’”

2. Databricks at $134B: the worked example of pricing extra growth

• The setup Rory calls “very convenient”: Databricks rumored raising $5B at $134B — 32x 2025 sales of $4.1B, growing 55% — while direct competitor Snowflake trades publicly at roughly the same ~$4B revenue, 28% growth, $80B, 20x. “How much extra in multiple do you pay for how much extra in growth… right here is the worked example.”
• If the extra 25-30 points of growth persist three or four years, the premium is “worth every dollar and then some.” But there’s no public data set to price it: “there is literally only one public company growing more than 30%” — likely Palantir, at 50%, wildly profitable, 80x sales. Jason’s shorthand: Databricks “would be the second best public company if it were public today.”
• The killer fact is that Databricks “continues to modestly accelerate,” and reacceleration at scale makes valuation models unreliable — “if it continues to re-accelerate, just for the record, it’s infinitely valuable, because that’s just what the math says.” Same dynamic as Anthropic’s reacceleration this year, which forced the step-function valuation rerating. Base rates: maybe one in three companies reaccelerate for a single year, only one in ten for two years — and it’s very rarely seen from a 50% base.
• The relative-value kicker — “seed’s for suckers”: risk- and time-adjusted, a coherent case says Databricks’ 3-5x from here beats a far less certain Series B with seven-to-ten-year duration, or “a seed deal at 60 post on a SAFE.” Jason: “Databricks seems like a better deal… quite cheap, honestly.”

3. Snowflake vs Databricks: a ten-year slugfest, with agents 1% into the journey

• Rory’s answer to “peaceful coexistence”: no — “they probably hate each other, in fact, and we know they hate each other because they time their sales events to overlap.” Expect a 10-year grind like SAP vs Oracle: neither folds, both want the other’s adjacent market, and “at some point some of those margins get dinged.” He rates Databricks CRO Ron Gabrisco’s claim that the “technology is 5 years ahead” as correct on AI-centric data workloads.
• Jason’s bigger point: the ground just moved — “all the vibe platforms now can directly access Snowflake data… that wasn’t even possible a couple weeks ago.” What agents on all your data mean for CRM and hosting: “I’m just not smart enough to even predict what that means in a year,” and “I think we’re 1% on this journey.”
• On whether CRMs become mere databases agents feed on: Jason points to Benioff’s 2,000 people on Agentforce as “the future right there” — met with the retort worth keeping: “it tells you his intentions. It doesn’t tell you his ability.” Salesforce has roughly two years to unlock it.
• Rory’s architecture call: single-app agents get bundled by Salesforce Microsoft-style, but multi-source enterprise agents (deposits + CRM + more) will “stuff it all in Snowflake and run an agent directly against that” — plus bespoke builds at the high end that “will make systems integrators rich for the next decade.”

4. Security is the revenge of the enterprise

• Jason’s alarm: Gainsight has been locked out of Salesforce for two weeks with no known resolution; Drift was kicked off five months ago and “Drift will never come back. It is dead.” The Drift breach saw 700 orgs’ data downloaded — including Cloudflare — with a pirate group demanding millions per instance.
• His conclusion: security benefits the incumbents. “I might want my agents from Salesforce and Snowflake and Databricks.” The escalation logic: one breach, you blame the twice-PE-flipped vendor; “two times I might start locking down my platform. The third time I might say I’m just going to own all the agents.”
• The cynical read, floated by Rory as “the best excuse there is”: use security to cut off the Gleans of the world slurping your data — “but lo and behold, I have my own agent product right here, which you can now safely buy, Mr. Customer.” Rory’s irony: both breaches were mature first/second-generation SaaS companies, yet new AI companies will get tagged with the restrictive policies. “Life is unfair, but there you go.”

5. Takeout season: smart money is buying SaaS at 2x

• The tape: Eventbrite acquired for ~$500M — 1.5x revenue with a 50% premium; PagerDuty near 2x at a ~$1B valuation on $500M ARR growing 4%; Semrush to Adobe two weeks ago. Rory on public-company vulnerability: when your stock floats at a low valuation and a 50% premium arrives, the lawyer gives the board “the speech about fiduciary duties” and “you’re very forced to take it.”
• The positive spin: “someone else — very smart money — thinks these things are worth buying,” and an aggressive PE firm could buy PagerDuty, bolt on a hot AI startup, get it back to 20% growth “and look like a hero.” Jason’s check on the fantasy: “we haven’t seen pagerduty.com and pagerduty.ai magically mashed together into a winner yet, have we?”
• Rory’s confession from the archive: his firm’s PagerDuty memo from ten years ago modeled revenues accurate within 3% — all that changed was what the market would pay. And the memo already said the add-ons were obvious: “literally every ops team on the planet uses PagerDuty… it’s pretty obvious what to add here. People, get it done.”

6. The TAM trap: the majority of public SaaS is in it

• Jason’s coinage of the episode: with average public SaaS growth at ~16% — “no one’s ever grown this slowly” — “the majority of the public SaaS companies, I think, are in a TAM trap.” If the Aaron Levies and Drew Houstons couldn’t escape it, “what hope is there for the rest of us?”
• Rory’s rebuttal — “maybe there’s no answer”: it’s not that the CEOs were idiots; venture funded so many SaaS companies that markets saturated and every adjacency was already occupied (his 2019 post, “Hunger Games and SaaS”). The quintessential case: Zoom — “everyone who needed a Zoom account has one, and everyone who hasn’t a Zoom account has a Teams account, the poor bastards.”
• Jason’s painful Zoom question: Eric is the best technical founder-leader he can name, yet “there was no great second act… why didn’t they find a way to add four billion of note-takers when there’s a trillion note-takers?” His practical conclusion: he tells founders to take their exits by default — and start the second product far earlier than feels natural.
• The rule and the escape hatch: “Overpayment only works when the TAM is huge. In finite TAMs, you got to bid more tightly.” The possible escape is AI pricing an order of magnitude up — Gamma at $100/month vs $8 for Canva, Cursor at $500 vs $3 for Jira. Rory’s caution: once three providers save the same $1,000 of labor, “they’re all willing to do it for a hundred bucks” — the $500 erodes.

7. SaaS is Japan: efficiency is the new margin, and nobody needs people

• Jason crunched the numbers: ARR per employee only rises — HubSpot 2.8x more efficient than 2021, Salesforce 2x, Microsoft “permanently past peak employee.” Hence Workday calling seat reductions an existential threat, and Jeff Lawson’s “unwavering” call that pricing moves off seats. Jason’s metaphor: “SaaS has become like Japan — it’s a great economy, but if everyone only has 0.9 kids… there’s only so many seats to go around.”
• Rory’s explanation for why multiples haven’t fallen as far as growth: in 2021 every plan was twice as inefficient at the margin; now it’s twice as efficient — investors trade slower growth for “wildly more efficient SaaS companies.” The endpoint is value-based pricing, which is much harder to measure: “you can count butts in seats pretty easily — every login is a butt.”
• On growth-stage startups, Jason is blunt: “in the fastest-growing companies that I’ve invested in, no one gives a rat’s ass about the bottom line” — the fastest AI companies carry the lowest burn multiples even with inference costs. His real red flag is bloat, not burn: the CMO who needs 50 people gets “a nice package and a good recommendation”; the healthy bar is 100% growth on 50% headcount growth, against the 2020-2023 DNA of the reverse “still ricocheting around” most executive teams.
• Rory’s three-category synthesis with a dark punchline: public companies must grind free cash flow, model labs spend on Nvidia not humans (“they just need geniuses and GPUs”), and app startups grow faster than they can hire. “The one thing they all have in common is they all don’t need people… that’s not great if you’re people.” He stays an AI optimist on employment long-term but concedes a “toughish market” for tech labor now.

8. Google clones Lovable in ten months — but model providers may focus on coding

• Jason actually tried Google’s new Lovable/Replit clone: shipped with no database and no auth — “it wasn’t impressive on itself,” and big companies “only have so many priorities.” But the meta-lesson stands: Google responded in under 10 months, and Datadog launched a PagerDuty competitor within the last 24 months against a company founded in 2008 — “you don’t get five years anymore… now you don’t even get a year.”
• Rory’s counter-comfort: the model providers may not be the competition in many apps. The OpenAI code red is “frankly tantamount to an admission” to focus the core — “win the ChatGPT wars and you are worth $2 trillion. Let’s not fuss around with little vertical markets that can be worth a couple hundred million bucks.” His aside: “that sound you might hear is the consumer hardware product slipping out.”
• Where they are most likely to go is coding — and Jason’s origin note: “we didn’t even figure that out. Claude figured that out. Cursor didn’t… Replit and Lovable both didn’t figure it out” — everyone grafted on. Where they won’t go: Rory’s new AI wealth-management deal, Range — “Google isn’t going to copy that.”

9. The gap between Wealthfront and Goldman is the next AI market

• Jason’s takedown of high-end wealth management: Goldman and Morgan Stanley are “the world’s crappiest product” — the only real service is the loan (sit on $50M of Nvidia stock, sell and pay $25M in tax, or borrow at 6%). “They’ll tell you they’ll help with your trusts. They don’t.” His mock pitch from an adviser: “Have you ever looked at my account? Do you know anything about me?”
• Rory’s Range thesis: don’t sell software to wealth managers — automate the business of wealth management and go “much further down the wealth continuum” to the doctor, dentist and entrepreneur whose affairs are “modestly complex” but can’t justify $20K lawyers. The evergreen rule: “whenever you see a product that only really rich people have, if you can find a way to get that in the hands of the rest of us, we all want it too.” Jason’s supporting anecdote: 11 months to set up three trusts, and the celebrated Silicon Valley trust lawyer’s consolation — “most of my clients never even finish them.”
• The prize and the discipline, both Jason’s: the average American retires with ~$1.8M in cash and equity (per the Wall Street Journal); take all the friction out of retirement, trusts, taxes and QSBS and “you could build a $20-40-50 billion company.” But charging $8-10K against incumbents’ $30-50K “is not 10 times the price of the existing product” — even Wealthfront, in theory a $10 trillion company, has a real TAM limit. And until the category is proven: “maybe just don’t spend all of it.”
• Harry’s skepticism — “what big business has been built in the wealth management space?” — draws Rory’s compounding defense: Wealthfront at 10 bps vs 50-70 for humans, adoption cycles of 10-15 years rather than 5, and at scale “asset management is a wonderful business,” a compounding machine like Schwab from the 1970s — “it’s going to be there for the next 30 years… in a way a lot of pure tech companies won’t.”

10. The relevance game vs the certainty rule — and Supabase vs Lovable

• Harry’s savage framing of new-age venture: “we’re playing a relevance game” — Ramp raises four rounds in a year, media matters more than ever, and LPs are seduced by fast markups. Rory’s own stat backs him: 40% of Q1’s first-time unicorns already had a follow-on round (corrected up from the 23% he cited last week). Harry: “I’d rather be playing that game than the ‘it’s coming.’”
• Rory’s counter-rule: filter on certainty of a big outcome, then everything else is adjustable — “there’s a rule in engineering that you’re only as accurate as your least accurate variable.” He invokes Thiel’s claim that “all that matters is can you build a big company here,” followed by the point that “they have no other rules.” His Schwab test — name five tech companies that IPO’d in 1983 (Harry: “Dude, are you kidding me? I was born in ‘96”) — is the point: tech companies “come quick and most of them go quick.” The LP’s brutal line he still carries: “there’s no such thing as blue-collar venture.”
• The coda both agree on, from Jason via Rory’s endorsement: “your discipline as a founder on capital is exactly proportional to how hot the market is perceived to be” — in an unproven category, stay capital efficient until you prove it, because “if you prove it, the world will beat a path to your door.”
• Closing quickfire, Supabase ($5B) vs Lovable ($6B): Rory takes Lovable — either vibe coding isn’t a category and both are screwed, or it is and the front end keeps more of the money: “you may as well be in for a penny versus in for a pound.” Jason takes Supabase — “hard problems are reassuring,” databases are painful to churn — while conceding Neon re-forked Postgres and Databricks bought it for a billion. Harry, an investor: “Lovable all the way.” Jason: “loyal to his paycheck.”