May 7, 2026 · 1h 36m · news
Anthropic's Raise & What It Means for Potential IPO? Mag7: Google & Amazon Up, Meta & Microsoft Down · 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 20VC, Harry Stebbings, Rory O'Driscoll, and Jason Lemkin analyze the record-breaking quarterly earnings of the Magnificent Seven, the viability of the SaaS market recovery, and the disruptive impact of AI on corporate structures, valuations, and global talent demands.
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 8.8% of the talking time here. How this is scored →
Harry as informed peer 4.1 Guest teaching 3.9 Guest disagreement 3.1 Harry pushing back 2.9
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:26:28 Jason's anti-manager 'Build or Go' rant
Jason aggressively demands that companies fire non-building middle managers, telling anyone talking about their team on LinkedIn to take four months severance and leave.
Hardest push from Harry ▶ 1:17:20 Harry forces a direct choice between Sierra and Anthropic upsideHarry refuses to accept vague venture logic around late-stage valuations, demanding Rory explicitly choose which company offers superior upside for a dollar invested.
Biggest teaching moment ▶ 1:20:40 Rory details the technical legal hurdles in Musk v. AltmanRory moves beyond headline gossip to educate the room on complex legal doctrines like statute of limitations and donor-advised fund standing that could dismiss Elon Musk's suit.
Harry holds his own ▶ 28:14 Harry cites Palantir metrics and hardware peer comparisonsHarry leads the segment with precise financial metrics, citing Palantir's 134% RPO growth and comparing its performance to hardware giants NVIDIA and SK Hynix.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Mag 7 Earnings & Google's Surprising Acceleration | 3 | 3 | 2 | 1 | Harry sets up the Mag 7 earnings overview with key figures before letting Jason and Rory analyze Google's growth. Rory highlights how hyperscalers are essentially financing capex and distribution for private LLM labs. The dynamic is collaborative with light host framing. | |
| The AI Value Chain: Hyperscalers vs. LLM Providers | 4 | 4 | 4 | 4 | Harry presses the guests on Microsoft's $190B CapEx versus $37B AI ARR. Rory calls bullshit on Satya Nadella's framing of low risk, pointing out that without AI, Microsoft's core revenue growth is flat. | |
| Meta's Capex Struggles & Mag 7 Stock Picks | 5 | 3 | 3 | 4 | Harry prompts the guests to explain why Meta was punished for CapEx expansion while Google was praised, then forces a quick-fire buy/sell stock pick. Rory and Jason contrast Meta's indirect ad-lift ROI with Google's clear cloud revenue stream. | |
| Palantir's Blowout Earnings & Large-Scale Enterprise Sales | 5 | 4 | 2 | 3 | Harry opens with detailed Palantir metrics including RPO growth and Rule of 40 figures before questioning its sky-high valuation. Rory explains why Fortune 500 CEOs award $10M+ deals to Palantir over fragmented point solutions. | |
| The AI Skills Gap & Low-Value Job Disruption | 1 | 5 | 2 | 1 | The guests lead a dialogue on the labor shift toward $250k AI-enabled SDRs and the collapse of low-skilled junior roles. Harry remains passive as Jason and Rory describe the Darwinian test facing corporate talent. | |
| Apple's Buyback Success & Memory Price Inflation | 4 | 3 | 2 | 2 | Harry interjects with a sharp observation regarding memory chip supply constraints inflating CapEx budgets. Rory builds on the point, explaining how hardware price inflation impacts consumer hardware pricing like Mac Minis and iPhones. | |
| Is the SaaSpocalypse Over? Atlassian, Twilio, & HubSpot | 5 | 4 | 3 | 3 | Harry opens with earnings metrics for Atlassian, Twilio, and Five9, questioning if the SaaSpocalypse is over. Jason and Rory debate whether current momentum reflects genuine re-acceleration or bounded multiples on mature bases. | |
| Token Intensity Economics: Coding vs. AI Workflows | 4 | 5 | 3 | 2 | Harry initiates a discussion on software developer capacity and plays a Price-Is-Right style game guessing token costs. Jason surprises the room by revealing that two full-time autonomous AI agents cost Saster just $254 a month in tokens. | |
| Anthropic's Mega Private Round & Potential IPO Delay | 4 | 4 | 3 | 3 | Harry brings LP market feedback regarding Anthropic's mega private round and probes whether this delays 2026/2027 IPO timelines. Rory candidly admits his previous prediction that Anthropic wouldn't raise privately was wrong. | |
| AI's Impact on Customer Experience TAM and the SaaSpocalypse | 5 | 4 | 3 | 3 | Harry introduces Sierra's massive $950M round at a $15.8B valuation, framing its 105x revenue multiple against a $400B market opportunity. Jason voices skepticism on whether CX TAM is truly expanding 10x due to AI. | |
| Sierra vs. Anthropic: The Debate on Valuation and Downside Protection | 5 | 3 | 4 | 5 | Harry challenges the late-stage venture multiple logic of GV and Tiger before forcing Rory to choose between Sierra and Anthropic for better upside. Rory instantly rejects Sierra, declaring Anthropic far more likely to yield multi-trillion upside. | |
| Analyzing the Musk vs. Altman Trial: Legal Nuances and Tech Gossip | 3 | 6 | 2 | 2 | Harry opens the floor on week one of the Musk v. Altman trial. Rory educates the host and listeners on technical legal mechanics, explaining why statute of limitations and donor-advised fund standing could dismiss the suit regardless of media drama. | |
| Coinbase's 'Build or Go' Mandate and the Future of AI-Enabled Lean Teams | 5 | 3 | 7 | 4 | Harry steers into Coinbase's mandate to eliminate non-building middle managers. Jason delivers a passionate, combative rant telling non-technical managers to take severance and leave, while Harry prompts with hot-take questions. |
Statements from this episode (40)
O'Driscoll: Microsoft's revenue growth is flat without its AI initiatives
“Without the AI initiative, Microsoft is flat revenue.”
Lemkin: Fire employees who brag about team size on LinkedIn
“Anyone on LinkedIn that talks about their team, fire them. My team this. They're all so precious about their team. Lead from the effin' front with AI.”
O'Driscoll: Big Tech AI CapEx growth consumes most free cash flow
“These same companies are doubling down on CapEx, letting it grow 50, 60%, such that CapEx is now eating most of their free cash flow.”
O'Driscoll: Anthropic likely grew its LLM token volume 15x in Q1
“But the truth is, Antropic probably grew tokens, 15 X in Q one, right?”
Lemkin: Public tech companies should spend all cash when markets permit
“When wall street lets you spend, it doesn't punish your stock. You should spend every dollar on the balance sheet. They'll let you because you'll build it back up, but it's trapped when growth slows”
O'Driscoll: OpenAI and Anthropic capture almost all commercial LLM revenue
“And when you look at what's actually going on, the big two guys are getting all the money. Gemini, for some reason, doesn't appear to be able to do a great job of developing code or love and just getting that kind of traction because coding is where it's all h…”
O'Driscoll: Developers consistently choose OpenAI and Anthropic over open-source models
“There are millions of developers who have that same choice every day, and they are choosing over and over again because of the model or because of the harness to go with the big two, right?”
O'Driscoll: Google outperforms other reporting tech giants in AI relevance
“I think Google has outperformed everyone in terms of being AI relevant. They are by far the best of the five people who reported this week.”
O'Driscoll: Meta's $150B AI CapEx to boost ad revenue is a mistake
“They've been using this justification For the LLM spend as being, it'll optimize our existing business, and I've always felt that's bullshit, because if it takes a hundred fifty billion in capex to give a 10% lift on a two hundred billion dollar business, it's…”
O'Driscoll: Investors should buy Amazon and sell Microsoft stock
“You buy Amazon and you sell Microsoft, but these are weekly held.”
Lemkin: Tech is experiencing the greatest application boom in history
“There is an application boom. This is the greatest boom in application building in the history of our lives. The amount has exploded.”
Lemkin: Markets will soon laugh at the predicted 'SaaS apocalypse'
“But I think in a couple of years, we'll look back and we'll laugh at the SAS apocalypse and say, my God, the explosion of B to B applications were like nothing we've seen before.”
O'Driscoll: App software must target a senior executive's top two priorities
“Whenever you're selling application software, you want to be one of the top two initiatives for the most senior person you're selling to. That's how you make money, right?”
O'Driscoll: Palantir is the only firm credibly delivering enterprise AI transformation
“Palantir right now is the only business that can credibly say, hey, Mr. Corporate America, Fortune 500, you want to move your needle with an AI initiative that has measurable, demonstrable results, That is about enterprise-wide transformation. We can deliver y…”
O'Driscoll: Palantir needs two to three years of growth to justify valuation
“Palantir is the best situated. And yes, it's wildly expensive. You know, you really have to go for two or three years to get into this valuation. But you know, if the boom has legs, they're the most likely to grow into it.”
Lemkin: The enterprise AI skills gap is the worst in modern history
“It's so it's the worst gap between in-house and external expertise in our lifetimes.”
O'Driscoll: Big Tech CapEx increases are largely driven by rising memory prices
“A significant slug of that CapEx raise is for the same physical amount of CapEx Just at a higher price, because the memory portion of anything you're doing has exploded in cost, right?”
O'Driscoll: Rising memory chip costs will drive up iPhone prices next year
“I think that you're going to see some impact of that right down to the price of your iPhone next year.”
O'Driscoll: Atlassian will never be a 5x or 10x AI growth company
“Taking it last, for example, I don't think it's like, it's not going to 10 X from here, right? It's not, and it's never going to be a AI first, you know, five X, 10 X growth company.”
Lemkin: Only infrastructure SaaS companies are winning two-pronged AI customer growth
“To be a two-pronged AI beneficiary, you have to be able to monetize your AI, and you have to attract new customers. There's two prongs, and the ones that have done it so far are close to infra, right? Cloudflare, Twilio, Mongo, Datadog and even DigitalOcean”
Lemkin: HubSpot's AI agent strategy will prove if classic SaaS can re-accelerate
“I believe over the next 12 months, this strategy, if it's real, should lead to genuine re-acceleration at HubSpot. Otherwise, it's, there's no hope for this class, these classic categories”
O'Driscoll: Revenue re-acceleration will be rare for traditional SaaS companies adapting AI
“Re-acceleration will be exception, not the rule, right?”
O'Driscoll: Non-coding knowledge jobs have less than 10% task automation potential
“And I think in anything other than pure AGI, which is too arm wavy for me, most of the other jobs have a task automation potential, and my view that's sub-ten percent.”
O'Driscoll: AI increasing developer efficiency will lead companies to hire more engineers
“If automation, if 20% spend on tokens, three X's the effectiveness of your developer, then the ROI in developers goes up, so the number of developers will go up.”
Lemkin: SaaStr spends $254 monthly on two full-time autonomous AI agents
“Two of them full-time, two full-time human equivalents, 254 dollars a month.”
O'Driscoll: AI compute token prices drop by 10x every 18 months
“Every turn of the crank on GPUs gives you a three X. Every turn of the crank on LLM optimization gives you another three X. You're looking at 10 X price falling every 18 months.”
O'Driscoll: Anthropic should accept a $50B private round instead of an IPO
“We live in a world where they can just send out a frickin' email, and people respond in 48 hours, and it's no drama, and they take the terms they get. So yes, they should grab the fifty billion.”
O'Driscoll: AI requires $3 to $4 in CapEx per $1 of revenue
“Every dollar of revenue that Antropic does means someone, either Antropic or its partners, has to invest three or four x in CapEx because it's just expensive in terms of compute to service AI.”
Lemkin: Anthropic's massive private raise decreases the likelihood of a 2026 IPO
“I think the combination of the two decreases the odds of the IPO this year.”
Stebbings: OpenAI chairman Bret Taylor is raising a $950 million venture fund
“Speaking of OpenAI, OpenAI's chairman, Brett Taylor, is out in market raising nine hundred and fifty million dollars”
Lemkin: AI has expanded the customer experience TAM by 50%, not 10x
“I'm not just a hundred percent convinced the CX market, as big as it is, four hundred billion, whatever you want to call it. I'm not convinced it's grown 10 X because of AI. I believe it's grown a bit. I believe it might've grown 50%.”
O'Driscoll: Customer support software TAM is $30B compared to $400B labor spend
“Cause just for a matter, my guess is the customer support software market Which exists today is probably 20 or thirty billion, and the customer support labor spend is four hundred billion.”
O'Driscoll: Sierra's LLM token spend is likely under 10% of its revenue
“And my guess is their LLM spend is sub-ten percent of revenues. In other words, LLMs are not the dominant portion of the value they deliver. It's the LLM plus the software, plus the whole thing, plus all the domain specific knowledge.”
O'Driscoll: Buying marquee assets at absurd prices has been the best strategy
“Buying marquee assets at absurd prices Has been by far and away the best strategy for the last three years, right?”
Lemkin: VCs consistently overstate downside protection and the number of potential acquirers
“Downside protection is real. We just overstate it. Right. And the number of potential acquirers is real. We actually overstate that too.”
O'Driscoll: Sam Altman holding zero equity in OpenAI is a mistake
“The most surprising thing is Sam Altman's worth zero on this, right, which I still think is weird and a mistake.”
O'Driscoll: Brian Armstrong's ban on political discussion at Coinbase was entirely correct
“In retrospect, an excellent call. I mean, an excellent, not only because it wasn't just a political call, it wasn't taking the other side, it was taking no side, which has turned out to be entirely the correct position.”
O'Driscoll: Executives must spend 10% of their time directly using new technology
“If you're not using the technology yourself at least 10% of your time, then you're in the business of listening to other people tell you things that you don't know if it's true or not.”
Lemkin: Modern CMOs should run marketing campaigns directly using AI agents
“A CMO today should be able to run their own campaigns, and this is why. I'm not saying you have to spin up Marketo or HubSpot. I'm saying you should be able to tell your agent, and our agent, our AIVP marketing started running its own campaigns the last two we…”