Interview

Stripe's $8B OpenRouter Bet | Anthropic's First Profit & The Math Behind Reaching $600B in Revenue?

Harry Stebbings is the podcaster of 20VC.

Jason Lemkin is one of the leading SaaS investors of the last decade with a portfolio including the likes of Algolia, Talkdesk, Owner, RevenueCat, Saleloft and more.

Rory O’Driscoll is a General Partner @ Scale where he has led investments in category leaders such as Bill.com (BILL), Box (BOX), DocuSign (DOCU), and WalkMe (WKME), among others.

Source attribution: Original source

Interview details

Publication date: Aug 22, 2026
Duration: 1h 17m

Speakers

Harry StebbingsPodcaster, 20VC Jason LemkinInvestor Rory O’DriscollGeneral Partner, Scale

Viewpoints in this interview

Cursor's survival shows the importance of extreme adaptability

Lemkin argues that Cursor's trajectory demonstrates how quickly AI markets can change and how companies must repeatedly adapt rather than give up when new competitors emerge.

Jason Lemkin Cursor and SpaceX Acquisition Source at 1:48

Market growth can outweigh serious margin weaknesses

O’Driscoll argues that Cursor's gross-margin problems remained real, but its position in a rapidly expanding coding market and its strategic fit with the buyer outweighed those weaknesses.

Rory O’Driscoll Cursor and SpaceX Acquisition Source at 5:49

Major AI acquisitions require unusually fast executive conviction

Lemkin argues that acquisitions on Cursor's scale can only be executed by a small number of leaders able to commit enormous amounts of capital within days.

Jason Lemkin AI M&A Strategy Source at 8:04

Microsoft has the strongest strategic need for leading coding tools

O’Driscoll argues that losing leadership with developers is more strategically consequential for Microsoft than for Amazon or Meta, making ownership of a leading coding product particularly important.

Rory O’Driscoll AI M&A Strategy Source at 13:33

Fast markets reward buying infrastructure instead of building it

O’Driscoll argues that when an emerging market is expanding extremely quickly, strategic buyers may rationally pay prices that appear excessive on standalone financial analysis because acquisition accelerates entry and monetization.

Rory O’Driscoll Stripe and OpenRouter Source at 17:33

OpenRouter may remain valuable but fundamentally niche

Lemkin argues that OpenRouter is highly useful for developer tools, chatbots and fallback routing, but high-accuracy enterprise workflows often standardize on only a few models, limiting universal routing demand.

Jason Lemkin Stripe and OpenRouter Source at 21:18

Limited enterprise model diversity could constrain routing value

O’Driscoll agrees that OpenRouter faces a structural risk if enterprises qualify only two or three preferred models instead of maintaining a broad portfolio requiring sophisticated routing.

Rory O’Driscoll Stripe and OpenRouter Source at 23:00

OpenRouter could become infrastructure rather than a standalone product

Lemkin predicts that OpenRouter's current product may disappear as a distinct offering and become absorbed into a broader Stripe token-management platform.

Jason Lemkin Stripe and OpenRouter Source at 26:03

OpenRouter represents a potentially significant market expansion for Stripe

O’Driscoll views the acquisition as a strategic attempt by Stripe to enter a new market, with meaningful upside if AI routing develops into a major second business.

Rory O’Driscoll Stripe and OpenRouter Source at 26:46

Rapid gross-profit growth makes Anthropic profitability unsurprising

Lemkin argues that Anthropic's extraordinary revenue growth combined with meaningful gross margins makes operating profitability difficult to avoid because expenses cannot scale as quickly as gross profit.

Jason Lemkin Anthropic Economics Source at 29:17

Anthropic's valuation depends primarily on sustaining extraordinary revenue growth

O’Driscoll argues that compute commitments and other liabilities remain manageable if revenue continues growing rapidly, making future revenue assumptions the central variable in Anthropic's economics.

Rory O’Driscoll Anthropic Economics Source at 31:43

Six hundred billion dollars of AI revenue requires extreme spending

O’Driscoll argues that Anthropic's much larger long-term revenue targets become difficult to justify because they would require AI spending to consume an exceptionally large share of global knowledge-worker compensation.

Rory O’Driscoll Anthropic Economics Source at 34:47

Hundred-thousand-dollar AI budgets per engineer could become normal

Lemkin predicts that companies may normalize roughly $100,000 of annual AI inference spending per engineer when that spending materially accelerates software delivery.

Jason Lemkin AI Engineering Spend Source at 37:44

Anthropic benefits strategically from going public before OpenAI

O’Driscoll argues that Anthropic would be better positioned as the first company to reach public markets, especially while it can present strong growth, enterprise leadership and at least intermittent profitability.

Rory O’Driscoll AI IPO Strategy Source at 42:56

OpenAI may accept going public after Anthropic

Lemkin suggests OpenAI may have accepted being the second major AI company to list publicly rather than rushing while addressing internal and operational issues.

Jason Lemkin AI IPO Strategy Source at 43:32

Workday's system of record provides retention rather than growth

Lemkin argues that Workday's system-of-record position makes customer churn unusually difficult and its future relatively predictable, but does not itself create reasons for customers to increase spending.

Jason Lemkin Workday and Private Equity Source at 51:23

Workday's closed architecture strengthens its private-equity defensibility

Lemkin argues that Workday's relatively closed ecosystem can help it retain more economic value and shield it from third-party AI agents that could otherwise abstract away an open system of record.

Jason Lemkin Workday and Private Equity Source at 54:20

Private equity must preserve customer value after acquisition

O’Driscoll argues that Workday's ecosystem can remain highly profitable, but excessive price extraction or insufficient product investment could encourage customers to seek newer alternatives.

Rory O’Driscoll Workday and Private Equity Source at 55:53

AI application platforms are becoming harder to displace

Lemkin argues that products such as Cursor, Lovable and Replit are rapidly accumulating functionality and complexity, turning from simple applications into richer platforms that become progressively harder for competitors to reproduce.

Jason Lemkin AI Application Valuations Source at 63:51

Cursor provides a valuation benchmark for Lovable

Lemkin argues that Cursor's acquisition has established a market comparison that makes Lovable's valuation appear less extreme than it would have previously.

Jason Lemkin AI Application Valuations Source at 64:38