20VC: Dario and Anthropic Declare War on Open-Source | Coinbase Slash AI Spend by 50% | Kalshi's $40BN Valuation and Impending IPO | Bending Spoons: Smartest IPO of 2026 and the Year for SaaS Roll-Ups
Harry Stebbings, Jason Lemkin, and Rory O'Driscoll discuss major tech developments including Coinbase's 50% AI spend reduction, Anthropic's concerns about Chinese model distillation, Microsoft's declining growth, Bending Spoons' $20B IPO valuation, and emerging opportunities in B2B SaaS consolidation. The conversation explores tensions between frontier model economics, open-source competition, regulatory dynamics, and venture funding standards in the AI era.
Summary
The episode opens with Jason Lemkin expressing frustration with performative AI announcements from struggling CEOs without corresponding revenue growth. Harry and Rory defend Coinbase's announcement about cutting AI spend 50% in two months while maintaining token usage, viewing it as valuable data about cost optimization that will pressure other CFOs. The discussion reveals disagreement about whether this signals frontier model vulnerability.
The panel then addresses Anthropic's letter to the Senate Banking Committee alleging that Chinese open-source companies are breaching terms of service by sending millions of prompts to Anthropic's models and using responses as training data. Rory explains the dual legal issues (breach of contract and potential IP/trade secret violations) and speculates Anthropic may be laying groundwork for government bans on Chinese models used by U.S. companies, conflating distillation ethics with national security concerns. Jason argues this would constitute regulatory capture—that banning Chinese open-source models while ignoring Anthropic and OpenAI's own training data practices is hypocritical and would ultimately harm broader AI adoption through artificially high prices.
On Microsoft, the group discusses why the stock has declined 16.5% in a month—the worst month since 2000. The consensus is that despite substantial OpenAI investment, Microsoft lacks a standalone frontier model and compelling AI products for its core knowledge worker and developer segments, which are being disrupted by Cursor and Claude Code respectively. Azure growth deceleration signals broader concerns about whether AI productivity gains are materializing into revenue lift. Jason emphasizes that companies must show either revenue acceleration or cost savings from AI spending, and Microsoft's guidance of declining growth rates represents a failure in this environment.
Bending Spoons' $20 billion IPO valuation for a company with $1.5 billion trailing revenue (8-9x forward revenue multiple) prompts discussion of a different IPO model: the anti-AI rollup of mature SaaS companies. The group analyzes why this pricing makes sense given a potential 1,000+ acquisition targets and identifies missed opportunities in B2B, particularly in products like Marketo, PagerDuty, and Asana where broken cultures and stagnant products could be revitalized by motivated operators. Jason emphasizes that cost-cutting alone won't work in AI-era B2B—new revenue generation through AI-enhanced features is required.
Regarding Chamath Palihapitiya's $135 million fundraise for AI software factory 8090, Jason voices skepticism about wealthy founders splitting attention between multiple ventures, citing personal experience that such arrangements consistently underperform. However, Rory and Harry credit Chamath for entering the arena and acknowledge the opportunity in software development automation.
The panel discusses Harry's controversial tweet that 1.5M-5M ARR growth isn't sufficient for Series A in the current market. Jason reframes this as a reality check about opportunity cost of capital and the need for founders to plan accordingly, though he advocates for honesty from VCs about true market dynamics rather than sending founders on fruitless fundraising rounds. The discussion explores tension between unflinching market facts and compassionate founder support.
Finally, on Anthropic's Claude Tag in Slack, Jason expresses skepticism about whether this product is core to Anthropic's strategy, noting that for a company approaching $100B in revenue, products that don't generate $10B+ annually may not merit serious resource commitment. He suggests this could either be existentially disruptive if fully resourced or merely an experimental feature. The group closes on the observation that as frontier models scale toward trillion-dollar revenue targets, traditional SaaS disruption concerns may become immaterial to their strategic calculus.
About this episode
<p dir="ltr"><span style="text-decoration: underline;"><strong>AGENDA:</strong></span></p> <p dir="ltr">00:00 Coinbase Slashes AI Spend 50%—Is the AI Token Bubble Bursting?</p> <p dir="ltr">12:55 Anthropic Warns Open Source Could Destroy the AI Business Model</p> <p dir="ltr">18:10 Dario Escalates the AI War with China & Open-Source</p> <p dir="ltr">22:00 Should the US Ban Chinese AI Models?</p> <p dir="ltr">33:15 Microsoft's AI Strategy Is Breaking Down</p> <p dir="ltr">38:00 Kalshi's $40B Valuation Signals a New Consumer Gold Rush</p> <p dir="ltr">41:15 Has SpaceX Frozen the AI IPO Market?</p> <p dir="ltr">43:20 Why Bending Spoons May Be the Smartest IPO of the Year</p> <p dir="ltr">46:00 The $100B Opportunity to Buy Broken SaaS Companies</p> <p dir="ltr">53:30 Which Software Companies Would Jason Buy Tomorrow?</p> <p dir="ltr">1:04:30 The Great AI Talent War Is About to Get Worse</p> <p dir="ltr">1:11:20 Every Company Is Becoming an AI Company—or Dying</p> <p dir="ltr"> </p> <p> </p>
Key Insights
- Jason argues that software companies in the age of AI are either accelerating or irrelevant, and companies without AI-driven revenue lift should face scrutiny regardless of cost-cutting achievements.
- Rory posits that Anthropic's Senate letter about Chinese distillation is laying groundwork for government intervention rather than purely addressing fairness, with potential quid pro quo linking compliance with regulatory benefits.
- Jason contends that banning Chinese open-source models while Anthropic and OpenAI trained on unlicensed IP constitutes regulatory capture that would ultimately harm broader economy through artificially elevated AI prices.
- The group observes that Microsoft's stock decline reflects market recognition that without a standalone frontier model, the company's AI story is primarily OpenAI revenue pass-through rather than organic product innovation.
- Rory explains that Microsoft's Azure deceleration from 40% to 37% growth fails market expectations because investors require acceleration, not merely maintained growth, in the current AI-focused landscape.
- Jason identifies that Bending Spoons succeeds because it targets sticky B2B software companies with broken cultures, using low-cost Italian engineering and motivated operators to improve NRR rather than requiring product innovation.
- The panel suggests a Bending Spoons B2B model in the AI era would require actual feature development and revenue generation from AI, not just cost optimization like previous PE rollups, making execution significantly harder.
- Jason claims that venture capitalists frequently lie to founders about Series A viability by encouraging rounds they privately believe won't succeed rather than honestly discussing market realities and opportunity costs.
- Jason argues that Chamath's ability to execute on 8090 is doubtful given historical pattern that wealthy founders splitting attention between ventures consistently fail, despite some counterexamples.
- Harry contends that a founder growing from 1.5M to 5M ARR is still building a potentially generational company, and opportunity cost of capital doesn't negate that founders should pursue uncommon paths.
- Jason notes that as Anthropic approaches $100B revenue, specific products may not merit strategic focus unless they generate $10B+ annual revenue, explaining potential indifference to some enterprise integration initiatives.
- Rory observes that Claude Tag could be either existentially disruptive to SaaS architecture or merely experimental depending on Anthropic's resource commitment, with current evidence insufficient to determine strategic importance.
Topics
Transcript
I am getting burnout on struggling CEOs on Twitter sharing performative AI data when they're not AI companies. It's like, show me the money. If you can be the largest tech company on the planet and still not make money, you might have oversized your ambitions a little and it might pay to come back a bit. Software companies in the age of AI are either accelerating or irrelevant. AI is going to be like the oil situation in the Persian Gulf today. Poor old Warren Buffett is like, it's time for me to die because you people have lost the plot. As we record this, greed will still trump fear, right? This is 20VC with me, Harry Stebbings, and…
Full transcript available for MurmurCast members
Sign Up to AccessMore from The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
20VC: Is Seed Investing Dead Without a $1BN Fund? | Does Ownership and Price Matter When Companies Can Be $1TRN Exits | Are AI Revenue Numbers Real and What to Watch Out For with Venky Ganesan, Menlo Ventures
Venky Ganesan from Menlo Ventures discusses how venture capital has fundamentally changed in the AI era, arguing that seed investing still exists but founders must understand that venture is now focused on outlier companies with explosive growth rather than steady revenue growth. He emphasizes the importance of IRR over ownership percentages, the dangers of gaming metrics, and the need for VCs to play the current market game while maintaining authentic values.
20VC: The Future of Datacentres: What You Need to Know | Why Everyone Gets GPU Depreciation and AI's Energy Costs Wrong | Who Really Makes Money From AI & Why Most Moats Don't Exist with Chase Lochmiller
Chase Lochmiller, CEO of Crusoe Energy, discusses how data centers and AI infrastructure are becoming critical bottlenecks, arguing that energy-distributed computing rather than centralized hubs will power AI's future, and that vertical integration across data centers, GPUs, and inference services creates sustainable competitive advantages in the AI infrastructure space.
20VC: $1BN ARR in 18 Months; The Untold Story of Higgsfield | Spending $4M Per Month on Models | Why Moats in AI are BS | Scaling a Content Team to 150 People with Alex Mashrabov
Alex Mashrabov, founder of Higgsfield, discusses how his AI video generation startup achieved $1 billion ARR in 18 months—faster than Coursera—by focusing on product-market fit, strategic model routing between open-source and proprietary AI models, and building a 400-person team primarily based in Kazakhstan. He shares insights on AI moats being largely meaningless, the $4 million monthly spend on model inference, and his vision for Higgsfield to become bigger than Shopify by owning the distribution layer for AI-generated content.
20VC: Jensen Huang Declares AGI Has Arrived | GPT Astra and Fable 5.1 Accelerate the Model Race | Tesla Launches Cybercabs | Index Pulls Out of Town & Anthropic Pulls From Descartes Acquisition
In this 20VC episode, Rory O'Driscoll and Jason Lemkin discuss major tech news including Jensen Huang's AGI claims, AI agent risks and rule-breaking behavior, Elon's Cybercab launch, venture capital conflicts around AI assistants, and high-growth companies like Wonderful raising at massive valuations with significant secondaries.
20VC: The $100 Billion AI Assistant Race: Town vs Instinct vs GrokBot | We Spend $75K Per Engineer on AI Tools | Why the AI Assistant Market Is Not a Bubble & AI Assistants Will Replace Every App on Your Phone with JD, Founder of Town
JD, founder of Town, discusses how his AI assistant platform operates within email and calendar to automate work tasks, argues that network effects at the agent level will be the true moat in the AI assistant market, and explains why the speed of market development means companies can only learn at human speed while building at machine speed.