Oct 2, 2025 · 1h 24m · 20vc
Are Burn Multiples BS in an AI World? & Sam Altman Needs $1TRN of Energy · 20VC with Harry Stebbings
0:00 / 0:00
⌖ your search result is the highlighted band (47:37–48:03). Playback starts there
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
This episode of the 20VC podcast features host Harry Stebbings alongside venture capitalists Rory O'Driscoll and Jason Lemkin as they dissect the seismic shifts in venture capital and SaaS metrics driven by the AI revolution. The panel provides deep strategic insights on the death of legacy SaaS metrics, massive AI infrastructure costs, and the harsh realities of startup fundraising and consolidation in a highly bifurcated market.
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 11.9% of the talking time here. How this is scored →
Harry as informed peer 5.4 Guest teaching 5.5 Guest disagreement 2.5 Harry pushing back 4.3
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 17:16 Jason condemns traditional VC board advice
Jason aggressively rejects the conventional playbook given by incumbent VCs to founders, calling it terrible and borderline toxic in today's market.
Hardest push from Harry ▶ 47:39 Harry challenges $1T sovereign fund availabilityHarry bluntly rejects the premise that OpenAI can secure $1 trillion for data centers, pointing out that even sovereign wealth funds do not have that level of unallocated capital.
Biggest teaching moment ▶ 1:02:21 Jason details VC portfolio dilution mechanicsJason breaks down the precise mathematical trade-offs VCs face during unicorn mergers, showing how ownership dilutes from 20% to 8% and why single-firm overlap drives consolidation.
Harry holds his own ▶ 52:44 Harry delivers detailed critique of Meta AI strategyHarry demonstrates sharp market domain knowledge by dissecting Meta's AI strategy, contrasting its internal team friction and talent missteps with competitors like OpenAI and Microsoft.
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 Fallacy of the Burn Multiple in SaaS and AI | 5 | 6 | 2 | 5 | Harry asks a clarifying question on burn multiples and later directly pushes back on whether the metric carries any weight given its flawed inputs. Rory and Jason break down the core assumptions behind burn multiples, highlighting why fast-growing AI companies with lower gross margins distort traditional SaaS math. | |
| Fundraising Realities: The Haves and Have-Nots | 4 | 5 | 4 | 2 | Harry sets up the discussion by sharing real-world founder feedback about struggling to raise despite good metrics. Jason passionately criticizes legacy VCs who give outdated advice to founders, while Rory explains why sub-scale companies without an AI narrative lose their option value to VCs. | |
| Navigating the Kingmaker Effect and the Wall of Money | 6 | 4 | 2 | 4 | Harry articulates a strong thesis on how fear of kingmakers deters VCs from backing second or third category players. Rory admits that the sheer wall of money suckered into category leaders makes his pure evaluation model somewhat unrealistic. | |
| Valuation Madness: Peak AI Hype vs. VC Economics | 6 | 6 | 2 | 5 | Harry cites specific market data including extreme valuations for pre-revenue AI startups alongside public drops for Figma, Klarna, and StubHub. Rory draws a historical parallel to Irving Fisher's 1929 stock market proclamation to caution against assuming permanently higher valuation plateaus. | |
| The EA Take-Private and the State of Late-Stage M&A | 4 | 5 | 3 | 3 | Harry brings up the monumental EA take-private deal and lightheartedly mocks adults spending hours playing video games. Rory explains the financial leverage structure of the $55 billion LBO and praises Silver Lake's track record. | |
| Sam Altman's Trillion-Dollar Energy Quest | 6 | 6 | 2 | 6 | Harry challenges the feasibility of Sam Altman's $1 trillion energy and data center plans, stating sovereign wealth funds simply do not have that capital to dedicate to one firm. Rory offers a nuanced CEO versus CFO framework to illustrate visionary bets versus financial realities. | |
| Critiquing Meta's AI Strategy and Zuckerberg's Big Bet | 7 | 5 | 3 | 6 | Harry forcefully critiques Meta's AI execution, citing executive talent management and organizational friction compared to OpenAI and Microsoft. Rory counters by viewing Zuckerberg through a venture lens, noting his historical acquisitions justify rolling the dice again. | |
| The Trillion-Dollar Capex and ChatGPT's Commerce Experiment | 4 | 6 | 2 | 2 | Harry raises ChatGPT's new in-chat e-commerce purchasing capabilities. Rory and Jason explain that e-commerce experiments are rounding errors unless they generate billions in revenue to offset massive infrastructure costs. | |
| Unicorn Consolidation: Analyzing the Fivetran and dbt Merger | 6 | 7 | 2 | 4 | Harry introduces the Fivetran and dbt merger news with exact revenue figures and questions PE rollup strategies. Jason educates on the math of VC dilution, explaining how shared lead investors facilitate complex portfolio combinations. | |
| The Death of Static SaaS and the Impact of AI on the Seat Model | 6 | 6 | 2 | 5 | Harry questions whether private equity models buying legacy SaaS face existential risk from AI displacement. Jason and Rory analyze how software products are no longer static and why autonomous AI agents threaten seat-based pricing models. | |
| Fiduciary Duty versus Free Speech: Founders and Politics | 5 | 5 | 3 | 5 | Harry asks if founders have a fiduciary duty to restrain political commentary on social media when it risks alienating customers or employees. Jason recounts reaching out to ten CEOs about inflammatory tweets, observing that almost all chose personal expression over business risk reduction. |
Statements from this episode (28)
O'Driscoll: A $15M revenue moderate-growth company has zero VC value
“A fifteen million dollar revenue company that's perfectly good and has reasonable growth is actually of zero value to a VC because we're in the upside option game.”
Lemkin: Relying on 'triple, triple, double, double' growth is terrible advice in 2025
“I hear too many folks, and they're like, oh, you're triple, triple, double, double, or better. You're golden. Don't worry, kids. And I think that's terrible advice in 25. Terrible advice.”
Lemkin: AI-native startups under $100M ARR average -126% cash margins
“AI native companies, the ones we're all so excited about growing so quickly under a hundred million ARR. Have terrible free cash flow margins -126%. It's much worse. Non-AI companies are -56%.”
O'Driscoll: Legacy SaaS valuation models are broken for AI startups
“When they were all SaaS recurring revenue businesses, all seat based, all 90%, 80% gross margin with no CapEx, Right? All enterprise sales with low churn. It absolutely made sense. You could compare two companies, and that's why by 2019 or 20, it almost felt l…”
O'Driscoll: Ten of 15 year-to-date IPOs had almost no AI story
“Of the 15 IPOs year to date, 10 of them have almost no AI story, right?”
Lemkin: 94% of public software companies call themselves AI companies
“94% of public software companies call themselves AI companies.”
Lemkin: Lovable and Replit both reach nine-figure revenues
“And that's why lovable and Replit are in a death match. And it's very powerful. They're both at nine figures in revenue. They won't kill each other.”
O'Driscoll: B2B market leaders capture 67% of total market value
“The number one makes, in a business market, 67% of the market, the number two makes 20, 30, the number three makes 10, and anything after that doesn't even matter, right?”
Lemkin: Number two or three startups should restrict burn to $100k monthly
“Look, if nothing else, if you're not number one, don't spend like you're number one. Even if you're growing pretty quickly, like Harry said, if your number, if you're the clear number two or number three and 80% of VCs are going to drive by, if you burn a hund…”
O'Driscoll: SoftBank proved money alone cannot make market winners
“You know, soft bank proved to everyone's complete satisfaction, right? That money alone cannot make winners, right? And which is very kind of them to run that economic experiment and, you know, prove the negative.”
O'Driscoll: OpenAI must hit $200B-$300B revenue or AI market faces severe crash
“One of two things is going to happen in the next five or seven years, right? Either A, you know, you are going to see pretty profound productivity changes. You're going to see companies like OpenAI literally being two to three hundred billion dollars in revenu…”
Lemkin: SaaStr team replaced 11 human employees with AI agents
“I live it that we've replaced 11 people on our team with AI agents.”
Lemkin: Top public B2B SaaS companies trade at 20x ARR on 30% growth
“The bar is like the top group of public B to B companies, it trades at 20 times ARR only averaging 30% growth rate.”
O'Driscoll: Pre-2019 public 30%-growth SaaS companies traded at 6x-7x revenue
“And the public company, as Jason said, up until about 2019 was 30% growth, roughly six or seven times revenues, right?”
O'Driscoll: AI infrastructure ambitions will become unfundable in 2-3 years
“I think if you look to a finance and economics lens, at some point in the next two, three years, the scale of the ambitions becomes too hard to fund, and you can find evidence of that, right?”
Lemkin: Future GPU data centers will match city-sized power output with few humans
“So it will only be a couple years where the cities of the future don't even have humans in them. The S the next New York will all be GPUs and like 20 people managing an entire city, a New York city of data of GPUs.”
O'Driscoll: Projected AI adoption rates for the next 5 years are overly optimistic
“I personally think that the rate of adoption forecast over the next four or five, that's implicit in all these data center assumptions, will in retrospect prove to be too optimistic.”
O'Driscoll: Meta's AI bet will mirror VR flop, not Instagram success
“If there was a Calci bet, and maybe I should check it that, you know, some version of this AI strategy will not produce meaningful revenue despite a twenty billion dollar burn and will look more like meta VR and less like Instagram and WhatsApp, two of the mos…”
O'Driscoll: ChatGPT's real threat to Meta is stealing user attention
“The big picture comment is if you spend two hours a day on chat GPT, that's two hours a day that you are not spending on Facebook and we live and die on our attention.”
O'Driscoll: AI platforms like ChatGPT will inevitably launch advertising
“I'm sure the other shoe to drop at some point is advertising. So it all makes sense because it's the only way it's one of only two ways to monetize the free users. So it's going to happen.”
O'Driscoll: $1 trillion in AI capex is absurd and hard to cover
“Which, in my view, it exposes the absurdity of a trillion dollars of capex. You know, it's going to be really hard to cover that nut.”
O'Driscoll: VCs must merge unicorns to clear 30-year IPO backlog
“But at a zoom out level, this is the kind of thing that simply has to happen over and over again In everyone's venture portfolio because we have six, 700 unicorns. We've processed five, 15 out the IPO gate year to date, so 20 for the year. That implies we got …”
O'Driscoll: 8% of an IPO-bound company beats 20% of a non-public startup
“Yeah, 20% of something that's not going public is not nearly as interesting as 20, as eight percent of something that is going public, right? And if you believe that it's not a continuum of value, you know, a sliding scale of value, but rather it's a, it's lik…”
O'Driscoll: Fivetran acquiring dbt is a smart combo that reaches IPO scale
“And that's why I think this deal felt to me from a distance, like I'm not the infrastructure guy at scale, but it felt to me from a distance like that's a damn smart, obvious combo. You know, that'll be a good, that, that will get critical mass. I mean, you wo…”
O'Driscoll: PE firms pivoting to minority late-stage tech deals is stupid
“Going from what they do to making non-controlled late stage investments Just because, you know, pick a name, Thrive do that well would be, in my view, stupid because Thrive are really good at that, and they're not.”
O'Driscoll: AI startups risk constantly losing product-market fit as models evolve
“Product market fit When you locked into it in SaaS land, you just didn't unlock for 10 years. Whereas here, you can lock in and out of product market fate as the models change, as things, approaches become, and you look back on the product a year ago and you g…”
O'Driscoll: Corporate political neutrality is proven strategy after political activism failed
“There's a distinction subtle between bringing politics to work and not having that, which I think has been validated as the correct strategy. And I think that's almost, I won't say fully a given now, but I think that would be a pretty consensus because everyon…”