Aug 21, 2025 · 1h 34m · 20vc
Databricks at $100BN, CoreWeave’s $11B Debt Bet & Nubank’s $2.5B Profit Shocker - Ep.19 · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the 20VC podcast, host Harry Stebbings and tech investors Jason Lemkin and Rory O'Driscoll O'Driscoll dissect the modern venture capital landscape, examining the multi-billion-dollar valuations of Databricks and OpenAI, the infrastructure debt of CoreWeave, and the operational realities of AI-driven SaaS consolidation and global fintech models.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 12.1% of the talking time here. How this is scored →
Harry as informed peer 3.6 Guest teaching 2.9 Guest disagreement 2.9 Harry pushing back 2.0
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:22:35 Rory calls anyone naive about platform risk a baby
Rory aggressively dismisses the premise of being surprised by platform risk, stating that anyone who invested in Abridge without expecting Epic to build competing features is a baby whose money will soon be separated from them.
Hardest push from Harry ▶ 56:00 Harry directly challenges Rory's CoreWeave debt skepticismHarry refuses to accept Rory's skeptical framing of CoreWeave's $11.2B debt load, pressing him on what the actual issue is if the debt is backed by long-term customer demand.
Biggest teaching moment ▶ 56:13 Rory explains duration mismatch risk in debt financingFollowing Harry's challenge, Rory educates the host on fundamental banking risk, explaining how matching long-term lease liabilities against short-term inference customer demand creates severe duration mismatch exposure.
Harry holds his own ▶ 5:36 Harry cites firsthand interview with Databricks CRO Ron GabriscoHarry demonstrates strong subject matter expertise and access by citing his interview with Databricks CRO Ron Gabrisco, who claimed Databricks holds a five-year technological lead over Snowflake.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| The Holiday Edition Banter | 4 | 2 | 2 | 1 | Harry opens with strong contextual knowledge, citing his interview with Databricks CRO Ron Gabrisco and questioning if 25x ARR is actually cheap relative to growth rate. The guests offer light reframing on growth persistence, but the dynamic is overall friendly holiday banter. | |
| The Looming IPO Wave & Scarcity Premiums | 3 | 3 | 3 | 2 | Harry asks whether upcoming IPOs will command even higher public market pops. Rory politely pushes back with a counter-argument regarding public market scarcity premiums versus balance sheet absorption capacity. | |
| Power Law Outcomes and VC Fund Sizes | 4 | 3 | 4 | 2 | Harry quotes David from Vencap regarding outcome sizes scaling with fund sizes. Rory playfully calls Harry out for constantly repeating this point and being guilty of over-extrapolation, before explaining power-law distribution in mid-sized funds. | |
| Early VC Economics & Private Pricing Risk | 3 | 2 | 2 | 1 | Jason and Rory debate Andreessen's early Databricks entry valuation and timeline. Harry contributes Josh Kirshner's investment framework of buying the best house on every block. | |
| SPVs and the Broader Wealth Effect | 4 | 3 | 3 | 3 | Harry highlights Chamath's new SPAC prospectus quote about no crying in the casino and asks if it indicates peak bubble. Rory agrees SPAC structures encourage adverse selection while poking fun at the host's market euphoria. | |
| The "Casino" Mentality of Retail Investing | 2 | 3 | 5 | 1 | Jason questions why billionaires issue SPACs, calling it near grift. Rory goes on an impassioned rant rejecting the premise that capital markets are a casino, forcefully stating he would never let someone manage his money who thinks Wall Street is a casino. | |
| OpenAI's $6B Secondary & Employee Liquidity | 2 | 2 | 2 | 1 | Harry brings up OpenAI's $6B employee secondary and asks if multi-millionaire wealth impacts work ethic. Rory explains how $10M to $100M compensation offers from Meta force companies to offer cash liquidity to retain talent. | |
| Story Protocol & Saturday VC Deals | 3 | 4 | 5 | 2 | Harry brings up Story Protocol's founder stepping down after raising $130M. Rory completely dismisses the topic as uninteresting normal VC loss rate noise, refusing to treat it as a significant story. | |
| Nubank's Blowout Earnings & Global Fintech | 4 | 3 | 2 | 2 | Harry introduces Nubank's blowout earnings report of $2.5B net income and 123M customers. Rory contrasts Nubank's full-stack banking victory in LATAM against Chime in the US and Revolut in Europe. | |
| Neobank Demographics & Revolut's "Snack" Playbook | 5 | 2 | 1 | 1 | Harry demonstrates deep guest context from interviewing Revolut's Nick Storonsky, explaining their banking like snacks framework and product lab structure. Rory and Jason express strong admiration for the strategy. | |
| On Holding: The Consumer Goods Exception | 5 | 4 | 3 | 5 | Harry challenges Rory on CoreWeave's $11.2B debt load, asking what the issue is if demand is guaranteed. Rory schools Harry on classical banking duration mismatch risks between long-term leases and short-term inference clients. | |
| The AI Bubble & Macro Market Concentration | 3 | 4 | 4 | 3 | Harry tries to engage Rory in an over/under market cap game for CoreWeave. Rory explicitly rejects the premise, stating he is learning on the podcast not to give uneducated opinions just to fill airtime. | |
| Replacing Humans: The Pace of AI Labor Substitution | 5 | 2 | 2 | 2 | Harry cites Rory's previous thesis regarding the shift from tech budgets to labor substitution budgets. Jason validates this by sharing how 20VC's internal team deployed 10 AI agents to replace 5 human positions. | |
| AI Agent Consolidation and Warehouse Automation Lessons | 2 | 3 | 1 | 1 | Jason and Rory discuss vendor consolidation among AI agents. Rory shares an insight from warehouse robotics investment, noting tools must address a substantial quantum of headcount rather than small niche tasks. | |
| Vertical SaaS, Platform Risks, and the Epic vs. Abridge Dynamic | 4 | 4 | 6 | 3 | Harry asks if Abridge should be worried about platform risk after Epic announced rival transcription features. Rory delivers a sharp response, calling anyone naive about platform risk a baby who deserves to lose their money. | |
| The Kalshi Quick-Fire Round: AI Milestones, M&A, and IPO Predictions | 4 | 2 | 2 | 2 | Harry hosts a Kalshi quick-fire betting round with specific odds on AI model releases, acquisitions, and IPO races. Jason and Rory evaluate each probability analytically. |
Statements from this episode (47)
Lemkin: Snowflake recorded a $1B quarter growing at 26%
“Snowflake had a billion dollar quarter, growing 26%.”
Lemkin: Databricks crossed $4B in revenue while growing at 50%
“Databricks just said they're crossing four billion, growing 50.”
Lemkin: AI token and compute usage will 10x within 24 months
“I think we're all gonna live in AI, 24 seven, and we're all gonna use 10 times the tokens and 10 times the compute we are now in 24 months.”
Stebbings: Databricks is valued at $100B, surpassing Snowflake's market cap
“Listen, I want to start with Databricks now at a hundred billion dollars worth more than Snowflake.”
Stebbings: Databricks is undervalued at a 25x revenue multiple
“It actually feels undervalued at 25 X revenue, Given the growth rate, given the fact it's found a lead, given the macro trend where it sits, it really feels like actually very reasonably priced.”
O'Driscoll: A 25x revenue multiple requires two more years of 50% growth
“If you're paying 25 X runway for anything, you know, what you're really saying is, you know, the only thing, you know, for sure is that when things level out in growth rate, they get valued like normal companies. So implicit in that is some statement about how…”
Lemkin: Upcoming tech IPOs could trigger a double bubble
“We might get the next wave of IPOs that are even better very quickly, and the amount of froth that that could create in the ecosystem, the amount of just returns to LPs could create a bubble on top of a bubble, because times might just get much better in the n…”
O'Driscoll: Ill-informed regulation trapped high-growth tech companies in private markets
“There's very few high growth companies in the public markets because all that growth has been annexed by the private markets as a bizarre outcome of, you know, ill-informed regulation.”
O'Driscoll: Tech IPO wave will eliminate private market scarcity premiums
“If all those companies were to go public, as you say, if the canvas of this world, if the database wanted to go out, it's not clear to me that they'd get that three X pop. Instead, you might want to have is, okay, they'd probably trade at the very high, quite …”
O'Driscoll: Extended private market stays mathematically guarantee $100B+ VC outcomes
“In a power law world, the longer you let things cook in the private markets, the bigger one or two of the outcomes will be, and things have been cooking in the private market for, I mean, yes, I think SpaceX is almost 20 years old, they've been cooking a long …”
O'Driscoll: Broad venture capital returns depend on more $5B outcomes
“For the venture business to work for us all, you're going to have to see more of these five billion dollar market cap outcomes. Yeah, as you say, the hinge helps. The owners of this world start to take place, and I think it will.”
O'Driscoll: Private market investors lack stop-losses and absorb full downside risk
“The interesting comment is in the public markets, when things start going south, because there's liquidity, everybody can have a stop loss of 10% and get out, and then someone else can come in and think it's going to turn, and you know, yeah, yeah, you can hol…”
O'Driscoll: Andreessen Horowitz has invested at least $500M into Databricks
“For that guy, you probably, I'm willing to bet they have the guts of a billion dollars in that deal. Could be wrong. Half a billion at least.”
Lemkin: Buy Bay Area real estate before the next tech IPO wave
“So all I know over these cycles is you buy a house before the next massive IPO wage.”
O'Driscoll: Compensating investors for closing deals yields poor quality investments
“One thing you learn in any investment management business is if you pay people to do deals, deals get done regardless of the quality.”
O'Driscoll: The return of SPACs indicates a bubbly market
“The fact that the structure is back is indicative of a bubbly time.”
Lemkin: SPACs are at the edge of grift
“Why, why not do something more meaningful with the world rather than do something like a SPAC, which is at the edge of grift?”
O'Driscoll: If a bank explains a complex product, you are losing money
“Whenever a bank was explaining something complex, your rule of thumb should be whenever a bank was explaining something complex, you're losing money.”
O'Driscoll: Stated intentions are meaningless when faced with massive payouts
“One of my rules is this. What people say they'll do when faced with a large amount of money is meaningless and bears no relevance either way to what they'll actually do. Predicting who wants to sell their company based on what they say in the abstract is a was…”
O'Driscoll: Secondary liquidity does not reduce output for passionate employees
“I think money brings out what people really are, because it stops you having to do things you don't want to do. So the people who love their work are going to work just as hard or harder, because they'll be able to throw money at all the other little problems …”
O'Driscoll: Modern tech companies are 15-year marathons requiring private liquidity
“It made sense where it was a four or five year sprint in the mid nineties and you would go public at fifty million in revenue. I don't think it's realistic to have that attitude in a world where you can be 1215 years. So I think this evolution, you know, to al…”
Lemkin: Current tech era will see 20+ nine-figure founder cash-outs
“But it makes sense there should be double digits in this era, right? There should be 10, 20 founders who walk away with nine figures whether they're WeWorks or Hoppins, there's gotta be 20 or more of them.”
Lemkin: Hot venture capital deals are now exclusively done on Saturdays
“And it's a cost of doing deals over the, I find all hot deals now are done on Saturdays. They're all, this is the new founder vibe.”
O'Driscoll: Standard Series A and B venture loss rates are 30%
“You know, your typical venture loss rate, you know, your 30% of your deals don't work out, and that's kind of where we play A's and B's. It's probably 50% plus in seed. So, you know, and I remember in 2001 and two, the loss rate spiked to 60.”
O'Driscoll: Nubank dominated Latin America because incumbent banks were inefficient
“In LATAM, they were old, they were stodgy, they weren't serving their consumers, and they were, these guys came in over the top, and they didn't just build a deposit base like Chimebill, they've done the whole, the full monte, they do deposits, they do lending…”
O'Driscoll: Chime's valuation is capped at $11B by US bank efficiency
“Now it's at eleven billion dollars, despite the U.S. Being, you know, I don't know in the day, 10 X larger GDP country, it's, Chime is only an eleven billion dollar market cap company because, and I think that speaks to They were playing in a less profit-rich …”
O'Driscoll: Nubank's $63B market cap could double if it avoids banking mistakes
“If they can continue to do what they're doing and avoid doing dumb bank stuff, yeah, you could see a two x from here, which would be wow, which would be, you know, just a wow outcome.”
Stebbings: Revolut is incubating 26 new products in an internal venture lab
“Revolut internally has 26 new products that they are kind of baking in their venture lab, and they have insane testing, metric testing every week with Nick to determine what makes it to main stage, and they run them like little venture teams with different fun…”
Stebbings: Revolut's Nik Storonsky is the best founder I have ever interviewed
“I have all the founders that I've interviewed, and I've interviewed, you know, Jason, me and you'both interviewed thousands, 100,000 He's the best I've ever met.”
Stebbings: On Holding reached $4B in sales with 61.5% gross margins
“They have four billion of sales up 38% year on year, A software like margin, 61 and a half percent, they're worth fifteen billion dollars.”
O'Driscoll: Outlier consumer successes cannot support a repeatable venture fund strategy
“I think some things are categories of one, and they're very hard to access. And that's the kind of deal where you just go, good on them. I'm not sure you build a repeatable venture fund around that.”
O'Driscoll: CoreWeave is fundamentally a specialized real estate company
“This model, I mean, and there's nothing wrong with it. It's a real estate company. They're in the business of a very, very sophisticated, clever, risk managing real estate company.”
Lemkin: CoreWeave will show signs of an AI slowdown before hyperscalers
“Core weave, if core weave struggles, if it can't service its debt, if it can't raise more debt through this weird through, through the hyperscalers guaranteeing it, that's where we'll see, even if there's just a bump in this cycle, that's not that, that doesn'…”
O'Driscoll: AI debt models require three to five years of CapEx growth
“All this depends on three to five more years of continued AI CapEx expansion. If that's the case, everything works, and if that's not the case, all bets are off.”
O'Driscoll: 33% chance AI demand slows and multiples compress within 24 months
“I can contemplate at least a one in three chance that sometime in the next 12, maybe 24 months, that demand slows down just enough that people penciling stuff out to the sky suddenly realize maybe it will taper down. And at that point, a lot of these multiples…”
O'Driscoll: Sam Altman's $1T AI spending proposal will not pencil out financially
“It's hard for me to imagine it's a trillion, and it's also even harder for me to imagine if it is a trillion that it pencils out.”
O'Driscoll: Microsoft's capital expenditure has reached 25% of its revenue
“Microsoft CapEx as a percentage of revenue has crept up to 25%.”
Lemkin: SaaStr replaced five human employees with 10 AI agents
“We have 10 AI agents at our little team that have replaced five humans and it's going to go up.”
Lemkin: SaaStr spends $500K in list price on 10 AI tools
“We added up the nominal list price of our 10 tools, right? 500,000 dollars.”
Lemkin: Buyers will demand AI software from single vendors within 12 months
“We're already seeing this consolidation in e-commerce if we want to chat about where they're becoming one tool already. I think it's a real risk for all these cool ones. Like we just don't need five of them going to Rory's point. We may, Rippling's a good exam…”
O'Driscoll: Enterprise automation requires high ROI across large headcount volumes
“And my learnings have been, it's not enough to be ROI efficient. There's also so many, there's only so many things you can get your head around as the CFO, Or the chief of operations in a year. So you have to have a quantum. You have to be able to not just hav…”
O'Driscoll: Epic Systems holds 40% market share in large US hospitals
“Has like 40% plus market share in the electric, in the healthcare records marketplace in large US hospitals. It's dominant. I think it's a five or six billion.”
Lemkin: VCs are ignoring major platform dependency risks in B2B AI startups
“These are ones we used to shy away from in venture. These are one of the many risks we're ignoring today because the growth is epic. We're ignore, we're ignoring so many risks and platform attaches all over AI B to B. There's so much platform risk and we've gi…”
O'Driscoll: An AI adoption slowdown would trigger a venture capital Armageddon
“We're way out there on the risk curve, every one of us right now, right? And the only thing between us and Armageddon is AI adoption. If AI adoption keeps happening, this whole wagon train keeps on rolling, and we're all going to be fine, and we're going to ge…”