Mar 5, 2026 · 1h 25m · news
Anthropic vs The Pentagon: Who Wins? | OpenAI's $110BN Mega Round | Cursor Hits $2BN in ARR · 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, Jason Lemkin Lemkin, and Rory O'Driscoll analyze the geopolitical tensions of military AI contracts, OpenAI's historic private funding rounds, the 'SaaSpocalypse' driving massive tech layoffs, and how the rapid rise of AI coding tools is changing the venture capital landscape.
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 7.8% of the talking time here. How this is scored →
Harry as informed peer 3.8 Guest teaching 4.7 Guest disagreement 4.1 Harry pushing back 3.0
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 19:57 Rory rejects political decision-making objection
Rory forcefully shuts down Jason's claim that politicians shouldn't decide AI usage, dropping profanity and insisting on constitutional supremacy.
Hardest push from Harry ▶ 20:14 Harry challenges Rory on government competenceHarry directly counters Rory's constitutional argument by pointing out the practical reality that governments worldwide lack the technical depth to make sound AI policy.
Biggest teaching moment ▶ 40:23 Jason exposes public SaaS failure to adopt AIJason delivers an intense breakdown of public software executives failing to build competitive AI agents, arguing they deserve termination for missing obvious opportunities.
Harry holds his own ▶ 1:02:46 Harry counters VC consensus with Cursor ARR growth dataHarry brings real-world developer sentiment and Cursor's revenue doubling figures to disprove the guest narrative that developers had abandoned Cursor for Claude Code.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Anthropic's Clash with the Pentagon: Ethics vs. National Defense | 3 | 5 | 5 | 2 | Harry sets up the Anthropic vs. Pentagon debate and prompts the guests on Dario's ethics versus business realities. Rory forcefully criticizes Anthropic's position as naive and constitutionally wrong, while Jason explains internal safety team pressure. | |
| OpenAI's Pivot: Sam Altman's Deal and the Debate over Democratic AI Control | 4 | 4 | 7 | 4 | Rory passionately rejects Jason's premise that politicians shouldn't make AI policy decisions, dropping profanity to emphasize constitutional authority. Harry interjecting to defend the practical reality of government technical illiteracy provides host pushback. | |
| OpenAI's $110 Billion Mega Round: Capital Limits, the Altman Premium, and the Road to IPO | 4 | 4 | 2 | 4 | Harry presses both guests to go on record with exact public valuation and timing predictions for OpenAI's IPO. Rory and Jason break down the mechanics of Amazon's conditional funding and the Sam Altman valuation premium. | |
| The SaaSpocalypse Exposed: Declining Multiples, the AI Spend Gap, and Agentic Pivot | 3 | 7 | 7 | 2 | Jason goes on an intense, contrarian tirade against incumbent B2B software CEOs, declaring that legacy software leadership deserves to be fired for failing to integrate agentic AI. Rory tempers the rant with valuation multiple math while Harry lets the guests lead. | |
| Block's 40% Layoffs: Headcount Reductions, OPEX Efficiency, and the End of Bloated Tech | 4 | 5 | 5 | 3 | The panel debates Block's massive 40% headcount reduction. Rory pushes back against Jason's broad claim that every company holds 40% excess bloat, calling Jason's view extreme while analyzing corporate efficiency trends. | |
| Cursor's Explosive Growth and Battle with Claude Code | 5 | 4 | 3 | 4 | Harry uses broad developer feedback from social media to challenge the prevailing VC narrative that Cursor is dying. Jason acknowledges that VC portfolio feedback was out of touch with real-world enterprise adoption numbers. | |
| Enterprise Security Advantages and the Fortnite Effect | 3 | 4 | 2 | 1 | Jason outlines why Barclays and major banks adopt Cursor due to role-based access controls and CISO security guarantees. Rory draws parallels to Azure's enterprise market strategy in a highly collaborative exchange. | |
| The Rapid Acceleration of Autonomous AI Code Development | 3 | 5 | 2 | 2 | Jason details his hands-on experience building multi-agent AI marketing workflows, explaining how rapidly agent swarms are improving. Rory reinforces that software companies must reinvent themselves every six to nine months. | |
| Why Product Demos Have Lost Value for Software Investors | 5 | 4 | 4 | 5 | When Jason claims product demos no longer offer investment signal, Harry pushes back using concrete deal knowledge about Sona's pre-seed round. Jason defends his view that software velocity makes traditional seed evaluation obsolete. |
Statements from this episode (28)
O'Driscoll: Anthropic and SpaceX's next funding rounds will be IPOs
“You've got to believe that the next round for this, Antropic, and SpaceX are all public offerings.”
Lemkin: Every tech CEO believes they can cut 40% of staff
“Every single CEO I talked to doesn't think they need 40% of their team.”
O'Driscoll: Tech competition intensifies only at 60-70% TAM saturation
“The knife fight doesn't start until the TAM is like 60, 70% saturated.”
Calacanis: Anthropic delayed Claude launch before ChatGPT over safety concerns
“In fact, they refused to release Claude before ChatGPT because he didn't think it was safe. They only released Claude under, under duress.”
O'Driscoll: Anthropic retained all seven founders while OpenAI lost most
“Which is why they still have the seven founders when the other guys across the street have lost almost all but one or two of the founders.”
O'Driscoll: Anthropic was naive to dictate safety limits to the Pentagon
“Given that, I think it was naive to try and sell the government at all at the Department of Defense at all.”
Calacanis: OpenAI staff forfeit eight-figure equity packages to leave early
“I mean, people are leaving, you know, eight figures of vesting stock at OpenAI because they just, they want to work somewhere else after 13 months or six”
Calacanis: OpenAI revokes sales team keycard access to research buildings
“Open a don't even let the salespeople talk to the engineers, the researchers, they're not even allowed their cards don't work at the building of engineers because they don't want to be bothered by the go to market team.”
O'Driscoll: State power is a more tangible threat than theoretical AI risks
“AI might be scary in the theoretical after, oh, it could be used for this, it could be used for that. The state is scary in the real sense of we have laws, we have men with guns to enforce those laws, and we can take your company if we want to.”
Calacanis: xAI lost military deal when Palantir chose Anthropic
“He wasn't even approved at the Department of Defense for this application, right? Only XAI was, and they had lost. They had lost because Palantir picked Anthropic.”
O'Driscoll: Musk's death would cut Tesla 80%, Altman's cuts OpenAI 25%
“Tesla trades at a trillion today. I think if Elon died tomorrow, they traded two hundred billion. OpenAI trades at eight hundred billion today. I think if Sam Altman died tomorrow morning, we traded six hundred billion.”
Stebbings: OpenAI will IPO in October 2026 at $1.5 trillion valuation
“I'm going to say October and I'm going to say it's going to go out at 1.5 trillion.”
O'Driscoll: SpaceX generates $18 billion in revenue growing at 20%
“The existing business is eighteen billion growing 20%, so it's not obvious that it's worth a hundred times trailing revenues for 20%.”
Lemkin: Public software companies will perform worse throughout 2026
“I think all, almost all the public software companies are just going to do worse and worse for the rest of this year.”
Lemkin: Vibe coding is only a minor threat to software companies
“I think vibe coding is one of those minor threats to software companies in our lifetimes, and I vibe code constantly. It's a minor threat.”
O'Driscoll: MongoDB stock dropped 25% due to 40x EBITDA valuation
“Yamango had a very strong quarter, modest delta in terms of future guidance versus expectations, dropped 25%. Why? Answer, it was trading at 40 times next year's EBITDA.”
O'Driscoll: Software valuations averaged 6x revenue historically versus 20x during COVID
“There was 15 or 16 years of the bucket growing at an average of 30% valued at an average of six times revenues. Then there was two years of COVID bucket growing 40% average of 20 times revenues.”
Lemkin: Most public software CEOs will abandon growth targets in 2026
“Which I think most public software companies will get to during the course this year. They will give up. The CEOs will give up. This is my point. They will give up on returning to growth.”
Stebbings: Three mid-sized tech CEOs plan minimum 20% headcount cuts
“I spoke to three CEOs between 500 and a thousand people. And all three said to me that they were cutting by a minimum of 20%.”
O'Driscoll: Block's 40% cut is largest public tech layoff in 20 years
“This is the biggest percentage change for a publicly-held tech company in the last 20 years, which is just an astonishing certificate. The statistics.”
O'Driscoll: Apple generates $2.6M revenue per employee, Google generates $2.1M
“Apple's about 2.6 million revenue per employee. You know, Google is 2.1.”
Stebbings: Cursor doubled revenue from $1B to $2B in three months
“Cursor announces their movement from a billion to two billion in revenue in, in three months. Rumors of secondary round being done at fifty billion.”
Calacanis: Barclays approved Cursor as its first agentic AI coding tool
“Barclays had just rolled them out. Okay. That was the first agentic coding product that was approved at Barclays.”
Calacanis: Every autonomous AI agent will leak data if allowed
“Like every agent will leak data if it's allowed to, like every single agent will. That's how goal seeking works.”
Calacanis: 60-70% of Cursor model calls use Anthropic's API
“Apparently, 60 to 70% of Cursor model calls are still using Anthropix API, okay?”
O'Driscoll: AI startups fall behind without major product evolution every 6-9 months
“If your story is not changing significantly every six to nine months, you are probably falling behind. In fact, not probably, you are falling behind.”
Calacanis: Enterprise and developer buyers of frontier AI models are not price-sensitive
“People are not that, no matter what anybody says, we're not that price sensitive today.”
O'Driscoll: Software code alone will no longer provide a competitive moat
“Software alone will not be a competitive advantage. So there will still be extremely large demand for software, and there as a result of that will be very big software companies, but they will have to do something more than, hey, I wrote a bit of code for this…”