This Week in Startups
Is AI the new cigarettes? Our VC panel digs into Anthropic’s S-1 | E2356
This debrief was written and voiced by AI from the episode's audio.
🎧 AI Debrief
🎙️ Full Episode
Anthropic's S-1 filed: $16.3B revenue, $80.7B reported loss explained
Jason introduces Anthropic's S-1 filing with key financials. The company shows 14x revenue growth but a massive non-cash accounting charge on convertible notes. This sets up the core debate: is the profitability real?
Key Moments
Jason introduces Anthropic's S-1 filing with key financials. The company shows 14x revenue growth but a massive non-cash accounting charge on convertible notes. This sets up the core debate: is the profitability real?
DA Wallach frames the opportunity but raises a critical question: if AI is as valuable as cigarettes (unlimited demand, huge margins), why are enterprises spending only 0.1% of budgets on it? The answer hints at structural constraints, not price.
Ben Black explains Anthropic's shocking 900% NRR—same customer cohort spending 9x more year-over-year. He calls it potentially the most useful product ever created and raises questions about whether this growth is sustainable.
Dave McClure and Jason dig into why enterprises aren't spending more on AI despite its value. The constraint isn't cost—it's organizational capacity to actually deploy it. Engineers and business leaders are the limiting factor.
DA Wallach makes a provocative claim: Anthropic's founders are ideological true believers in AI's potential to save humanity, while OpenAI's Sam Altman is a venture capitalist playing to win. This philosophical difference shapes business decisions and competitive dynamics.
Ben recounts Dario giving a public scolding to Sam Altman for irresponsible compute deals. But when market pressures mounted, Dario did exactly the same thing: locked up massive CapEx with Google and Amazon. Ideology bends to business reality.
Lon breaks down the technical accounting for the $72.4B non-cash convertible note charge. It's real for financial reporting but not real cash. The deeper question: how will public markets learn to interpret AI companies' financials?
Ben explains why he stays away from Neo-cloud infrastructure investments despite high potential returns. Cost of capital and timing are out of your control. He takes the long view and prefers businesses with simpler risk profiles.
Jason introduces the concept of 'alicorns'—AI-native founders who reach $1B valuation by building capital-efficiently and skipping Series B rounds entirely. They're becoming the next generation of world-changing companies without traditional VC permission.
Search Episode
Summary
This Week in Startups analyzed Anthropic's S-1 filing with a VC panel, examining whether the AI economy's numbers add up. Anthropic showed operating profitability, a $72.4B non-cash accounting charge on convertible notes, and a shocking 900% net revenue retention rate. The company projects $65-80B annualized revenue with only 0.1% of corporate America's operating budgets currently spent on AI, suggesting massive headroom for growth. The debate centered on whether "intelligence is the new cigarettes"—an addictive product with unlimited demand and gross margins. Key tensions emerged: whether Anthropic's growth can sustain massive CapEx commitments, whether the company's founders are idealists or capitalists, and how enterprise adoption will scale. The panel discussed emerging managers, AI-driven entrepreneur efficiency, and the shift from traditional VC to SPVs and secondary markets.
Key Takeaways
Anthropic achieved operating profitability on an actual cash basis despite an $80.7B reported net loss, driven by non-cash convertible note revaluation. This reframes the company's financial health and addresses investor concerns about AI profitability.
high900% net revenue retention is extraordinary—same customers spending 9x more year-over-year. This indicates AI is becoming embedded and essential, not experimental, suggesting strong long-term demand.
highCorporate spend on AI is only 0.1% of operating budgets today, implying 10-100x upside if adoption reaches 1-2%. However, the constraint is organizational capacity ("meat sacks") to deploy AI, not just price.
highAnthropic's $100B+ CapEx commitments to Google and Amazon lock in long-term capacity but create execution risk. The payback period and amortization schedule are unclear, unlike SpaceX's sub-12-month payback.
highAnthropic's founders operate from idealistic, almost religious conviction about AI's potential to cure disease and save humanity. This contrasts with competitors like Sam Altman (venture capitalist mindset) and Elon Musk (ambitious but pragmatic).
mediumEmerging managers are shifting capital to SPVs rather than raising traditional funds, commoditizing late-stage deal access. Meanwhile, AI-enabled young founders ("alicorns") are reaching unicorn status with minimal capital, shifting power back to seed investors.
mediumThree subscription services by revenue: Apple Services ($150B), Anthropic ($65-80B run rate), and OpenAI ($50B). This contextualizes Anthropic's scale relative to the most successful subscription businesses in history.
lowNotable Quotes
“I think this is the greatest business created by humans on planet Earth ever.”
— Jason Calacanis
Reflecting on Anthropic's path from zero revenue in 2023 to ~$100B+ annualized run rate by 2026, with profitability and massive cash position.
“The fundamental question for Anthropic and all the players in this market is basically is intelligence the new cigarettes.”
— DA Wallach
Framing the comparison to Philip Morris/Altria as the greatest business in stock market history, which had enormous margins and virtually unlimited demand.
“I liken this to teenagers in high school talking about sex. They wanna talk about everybody's doing it, but not everybody's doing it. And eventually, they do get there.”
— Dave McClure
Addressing the adoption gap—explaining why enterprise AI spend is still only 0.1% of budgets despite hype, and predicting acceleration as organizations move up the stack.
“These are dyed in the wool, legitimate religious extremists, and they might be right. And if they're right, we should thank them because they're saving civilization.”
— DA Wallach
On Anthropic's founders and their commitment to building safe, beneficial AI even at personal cost (smoking cigarettes to signal confidence in AI solving cancer).
“These 16 year olds right now in their bedrooms who are building all this shit with agents, and they're token maxing, and they're starting companies, and they have no money and no one's permission. These are going to be the greatest fucking entrepreneurs in human history.”
— Jason Calacanis
Arguing that AI has become so accessible and capital-efficient that the next generation of founders will outcompete traditional VC-backed companies and don't need permission or capital to succeed.
Predictions
Anthropic will reach $150-200B annualized revenue by end of 2027, not the previously guided $200-300B, due to slower growth trajectory and market share pressure from OpenAI.
— Dave McClure · Confidence: medium · End of 2027
Enterprise AI adoption will accelerate from 0.1% to 1-2% of corporate operating budgets within 3-5 years as organizations solve the organizational deployment challenge.
— Panel consensus · Confidence: medium · 3-5 years
Apple will launch integrated AI directly on iPhones at parity with Claude/ChatGPT, leveraging its distribution advantage and avoiding the $20/month subscription model.
— DA Wallach · Confidence: high · Within 1-2 years
Anthropic's founders will remain focused on safe AI development over profit maximization, potentially losing market share to ruthlessly commercial competitors like OpenAI and Meta.
— DA Wallach · Confidence: medium · Long-term (5-10 years)
Emerging managers will be squeezed out by capital consolidation; the future of seed investing belongs to solo operators and small syndicates backing AI-enabled founders who need minimal capital.
— Jason Calacanis · Confidence: high · 2-3 years
Bitcoin will never break its all-time high again and will stabilize as a store of value (debasement trade) rather than appreciating asset, with attention shifting to AI and superintelligence.
— Jason Calacanis · Confidence: low · 5+ years
Action Items
If you invest in late-stage companies or secondaries, focus on profitability metrics and cash burn rather than growth-at-all-costs. Model the CapEx amortization schedule (payback period) for hardware-heavy AI businesses.
The panel emphasized that understanding compute costs and payback periods is critical to valuing NeoCloud and AI infrastructure plays. A 2-3 year payback is reasonable; longer timelines introduce execution risk.
Evaluate AI adoption readiness within your portfolio companies. Focus on organizational bottlenecks ("meat sacks") preventing deployment, not just price or model capability.
The panel identified that corporate spend is constrained by organizational capacity and hesitation, not cost. Companies solving the deployment and integration challenge will see outsized AI ROI.
For founders: prioritize revenue and capital efficiency over large fundraises. Model how to reach $10M+ revenue with minimal dilution ("alicorn" strategy) to compress funding rounds and retain control.
Calacanis argued that AI-enabled founders can reach unicorn status without raising Series B and that VCs are now asking unicorns to take secondary sales. Lean capital paths are a competitive advantage.
Monitor Anthropic and OpenAI's Q3 and Q4 2026 financials closely to assess whether the 900% NRR is sustainable and whether revenue growth continues to decelerate.
The panel noted that initial S-1 reception was positive, but revenue trends and cost-of-capital sustainability remain key questions for the IPO valuation.
Subscribe to on-chain data (open-source model adoption, token usage patterns) to track whether enterprises are hedging AI spend across multiple providers or consolidating on Anthropic/OpenAI.
The panel discussed concerns about whether leading AI companies can defend against open-weight and cheaper alternatives. Direct usage metrics will signal market share pressure early.
Review your family office or portfolio strategy for Bitcoin and crypto exposure. Consider rebalancing from speculative assets (Bitcoin) to productive assets (AI infrastructure, productive software).
Calacanis published a provocative post arguing Bitcoin will never break all-time highs and that capital should shift to superintelligence. While contrarian, it signals a broader sentiment shift among sophisticated investors.
Links Mentioned
PayPal Business—checkout and payment infrastructure
https://paypal.com/business
The Conservation Fund—land and wildlife protection organization
https://www.conservationfund.org/
Hostinger—web hosting and AI website builder
https://hostinger.com/twist
Plaud—AI voice recording and note-taking device
https://plaud.ai/twist