Why the most expensive Seed deals are the cheapest | E2299

Episode Summary

Executive Summary: This Venture Capital Roundtable focused on how AI is reshaping startup liquidity, fundraising, and value creation. The panel argued that the upcoming IPO wave from SpaceX, OpenAI, and Anthropic signals broad market de-icing, while AI-native startups are growing far faster and raising capital on very different terms than in the SaaS era. The discussion also covered token and compute financing, open-source models, vertical AI, defense tech, and rising seed prices.

Main Topics: IPO wave and venture liquidity (Priority: 5/5): The panel sees the pending SpaceX, OpenAI, and Anthropic public offerings as a major liquidity event that could reopen the exit market and accelerate capital recycling across venture. AI-driven startup growth and fundraising power shift (Priority: 5/5): Speakers argued that AI-native companies can reach revenue milestones dramatically faster, giving founders more leverage and reducing dependence on traditional venture capital. Token spend, compute financing, and new capital structures (Priority: 5/5): A key theme was the emergence of financing tied directly to tokens, GPU hours, and compute, reflecting how AI inputs are becoming the new currency for startup scaling. Model competition, open source, and application-layer value (Priority: 4/5): The panel debated whether value accrues to frontier labs or applications, with consensus that open-source models and orchestration layers are increasingly important for cost control and differentiation. Seed pricing and selection pressure (Priority: 4/5): Rising seed valuations were viewed as a challenge for early-stage investors, but not necessarily a market break if top companies can still produce outsized outcomes. Defense tech and industrial AI examples (Priority: 3/5): Portfolio examples like Saronic and Knox Metals illustrated how AI-adjacent innovation is expanding into defense, manufacturing, and reindustrialization. VC role, board help, and founder relationships (Priority: 3/5): The group emphasized that good venture investors add more than capital: they help with IPOs, acquisitions, boards, and strategic advice, though founder control is growing.

Key Arguments: The upcoming SpaceX, OpenAI, and Anthropic IPOs could bring enormous liquidity and validate demand for AI and space exposure. AI-native companies are scaling much faster than SaaS-era companies; a 3x year is no longer exceptional, and 10x+ growth can be expected in some categories. Many companies now spend heavily on tokens and inference, so later-stage capital is increasingly being raised to fund usage costs rather than hiring sprees. The strongest application-layer companies will capture value by managing workflows, context, and model routing, not by owning frontier models alone. Open-source models matter because they let startups compete on cost and performance, especially for low-margin businesses. Seed-stage investing is becoming more polarized: median valuations may be rich, but the best companies may still be underpriced relative to future outcomes. Later-stage VCs still matter for expertise, board guidance, IPO support, and introductions, even as founders gain more leverage and alternatives. Defense and industrial startups can convert urgent, real-world demand into revenue quickly, showing that AI’s impact extends well beyond chatbots and software. Compute and token-based financing could become a meaningful part of venture dealmaking as raw AI inputs become monetized directly. Founder control over boards and financing has increased, but the need for specialized advisors and crossovers remains important.

Data Points: SpaceX IPO size: $1.7 trillion - Referenced as the expected market value of SpaceX going public in the discussion Anthropic valuation/IPO figure: $856 billion - Used in a combined liquidity calculation for major AI company exits OpenAI valuation/IPO figure: $950 billion - Used in a combined liquidity calculation for major AI company exits Combined liquidity from three offerings: About $3.5 trillion - SpaceX, OpenAI, and Anthropic combined potential liquidity SpaceX IPO demand: 2.5x to 3x oversubscribed - Cited from Reuters as evidence of strong demand Average revenue for recent IPOs: $300 million to $500 million annually - Discussed as the bar for recent billion-dollar-plus exits Portfolio growth example: 10x to 100x+ revenue growth in a year - Paige described growth rates seen in AI-native portfolio companies Example seed round: $4 million seed - A portfolio company example that reached a $120 million revenue run rate Example revenue run rate: $120 million ARR - Same portfolio company example cited by Paige Example free cash flow: $750K per month - The example company was described as generating this amount of free cash flow Typical large Series A alternative: $75 million for 6% equity - Michael said he would take that deal if offered on a hot company Anthropic model performance gain: 13.4 to 29.3 - Tomas cited an agentic coding benchmark jump for the new model release Model release cadence: Every 41 days - Used to describe how frequently major model updates are arriving Seed valuation 95th percentile: $174 million - Carta data point for U.S. seed round valuations in the 95th percentile Seed valuation 75th percentile: $94 million - Carta data point for U.S. seed round valuations in the 75th percentile Seed valuation comparison: $66 million and $50 million - Referenced as 2022 comparables for high-end seed pricing OpenAI chatbot revenue benchmark: $120 ARPU per year - Tomas compared this to Google to argue for potential ad monetization

Pivotal Quotes: "I think we're going to see broad liquidity." — Tomas Tunguz: On whether the SpaceX, OpenAI, and Anthropic IPOs are isolated or signal a broader reopening of exits "You control your own destiny quite a bit." — Michael Downing: Describing how AI-native founders with small prior raises now have more leverage in fundraising "I think the expectation is, oh, there's a bug next tomorrow morning, I think it'll be fixed." — Tomas Tunguz: On how AI is changing software buyer expectations toward faster product iteration

Implications: The panel expects a major liquidity cycle and a continued shift of power toward founders and AI-native startups. Investors will need sharper selection, cheaper model routing, and new financing structures to compete in a market defined by speed, scale, and compute costs.