Why the most expensive Seed deals are the cheapest | E2299

Key Concepts

• AI-Native Growth: Startups achieving 10x annual revenue growth, driven by massive enterprise AI budgets and "net new" spending.
• Liquidity Event: The 2026 market environment featuring major IPOs (SpaceX, OpenAI, Anthropic) signaling a robust exit market.
• Token/Compute Financialization: The trend of AI labs and funds providing compute (GPU hours) or tokens in exchange for equity.
• Vertical AI: Specialized AI applications (e.g., factory floor automation) that require proprietary data and "on-the-edge" deployment.
• Blended Valuation: A controversial practice where investors structure rounds in tranches to lower their average entry price.
• Agentic Coding: The use of AI models to orchestrate complex tasks, manage memory, and federate work across multiple agents.

---

1. The 2026 Liquidity Landscape

The panel discussed a historic moment for the venture capital ecosystem, with three major companies—SpaceX, OpenAI, and Anthropic—preparing for public offerings.

• Market Demand: SpaceX’s IPO is reportedly 2.5x to 3x oversubscribed. The total capital raised by these three entities alone could exceed the total IPO dollars raised in the previous decade.
• Bending Spoons: Highlighted as an "AI holding company" model, where legacy businesses are acquired and reinvigorated with AI-native coding practices.
• Outlook: The investors agree that 2026 will be an excellent year for liquidity, with the exit market finally "de-icing" and becoming more amenable to the venture cycle.

2. Shifting Power Dynamics & Fundraising

• Growth Expectations: The bar for IPOs has risen; companies now typically require $300M–$500M in annual revenue. Early-stage startups are now expected to grow 10x annually to secure top-tier Series A terms.
• Capital Efficiency: Founders are increasingly savvy, often choosing to raise less capital to maintain control. However, the "use of funds" has shifted from hiring large teams to massive "token spend" (the cost of running AI models).
• The Role of VCs: While some argue that founders need less capital, the panel emphasized that elite VCs remain vital for navigating IPOs, board composition, and complex M&A.

3. Technical Methodologies: AI Deployment

• Hybrid Model Architecture: To manage costs, companies are using a "hybrid approach"—using state-of-the-art models (like Claude’s "Fable 5") for high-level reasoning and orchestration, while offloading repetitive tasks to smaller, local, or open-source models.
• Skill Distillation: Tomas Tongus explained that models can create "skills" (markdown files) that allow local models to perform tasks with 90%+ efficiency, significantly reducing token costs.
• Edge Computing: Paige Dhorty highlighted "Miva," which uses embodied AI on factory floors. By deploying agents on off-the-shelf hardware, they automate quality control, allowing human workers to focus on higher-level efficiency tasks.

4. Key Arguments & Perspectives

• On Nationalization: The panel expressed strong opposition to the nationalization of AI labs. Tomas Tongus argued that government-appointed monopolies lead to regulatory capture and that the history of Silicon Valley is built on "dual-use" technologies (private sector innovation with government relationships).
• On Seed Pricing: Carta data shows seed valuations reaching a 95th percentile of $174M. While some view this as unsustainable, Michael Downey argued that the top 1% of companies might actually be underpriced given the massive scale of potential AI-driven outcomes.
• On Transparency: The "blended price" practice (where investors invest in two tranches to lower their average cost) was identified as a growing, albeit controversial, trend. While not a "red herring," it is becoming more common in high-demand rounds.

5. Notable Quotes

• Tomas Tongus: "The willingness to pay and the urgency associated with the procurement of these systems or data is extreme."
• Michael Downey: "The destination point in terms of valuation, the outcomes are always way larger than what we anticipated."
• Paige Dhorty: "We’re seeing growth rates at like 10x with like 100x plus in a year from a revenue perspective."

6. Synthesis/Conclusion

The venture capital market in 2026 is defined by a shift toward "AI-native" operations where growth is fueled by massive enterprise adoption and net-new budgets. While seed valuations are hitting record highs, the panel suggests this is a rational response to the unprecedented scale of AI-driven outcomes. The future of the application layer lies in "orchestration"—bundling complex AI capabilities into simple, user-friendly software that abstracts away the underlying model complexity. Ultimately, the market is moving toward a more concentrated, high-stakes environment where GP discernment and the ability to navigate the "chasm" between lab technology and real-world ROI will determine the next generation of winners.