Why the VC Hype Cycle Always Gets It Wrong | VC Roundtable | E2307

Episode Summary

Executive Summary: This VC roundtable focused on the post-2023 venture reset: improving liquidity from IPOs and PE exits, AI’s impact on fundraising and company formation, the rise of open-weight models and model-agnostic architectures, and how founders/investors should navigate oversized seed/A rounds and hot AI markets. The panel argued that scarcity is shifting from generation to correctness, but warned of short-lived AI opportunities, inflated valuations, and growing social backlash.

Main Topics: Venture liquidity and the return of exits (Priority: 5/5): The panel debated how improved exits, including IPOs and PE acquisitions, may restore LP appetite and reduce the denominator effect, while noting early-stage funds still return relatively small absolute dollars compared with mega-funds. AI compressing cycles and changing venture strategy (Priority: 5/5): Speakers argued that AI is making product, fundraising, and competitive cycles much faster, forcing founders and VCs to reassess what qualifies as a real company, a defensible moat, and a viable growth trajectory. Open-weight models, model routing, and AI cost discipline (Priority: 4/5): They discussed the increasing use of open-weight Chinese models and cheaper alternatives to frontier models, emphasizing model agnosticism, routing, and shifting expensive model use to where it creates durable advantage. Oversized seed and Series A rounds (Priority: 5/5): The conversation criticized extremely large early rounds as often misaligned with stage, creating pressure for huge follow-on valuations and potentially trapping companies in unsustainable expectations. AI-native pivots vs. ‘burn the boats’ transformations (Priority: 4/5): The panel examined whether legacy software companies can successfully pivot to AI-native products, citing examples like Mutiny and Intercom/Finn, and highlighted the difficulty of retrofitting old products versus rebuilding from scratch. Acceptance AI, trust, and verification layers (Priority: 5/5): A central thesis was that as AI-generated output becomes abundant, value may shift to trusted third parties that validate correctness, compliance, and provenance—especially in enterprise settings. Public-policy, labor, and social backlash to AI (Priority: 4/5): The group closed on concern that AI is worsening inequality, fueling public distrust, and inviting heavy-handed regulation, while also noting the U.S. must stay competitive with China.

Key Arguments: Liquidity matters for LP fundraising, but early-stage funds return small checks; mega-liquidity events like SpaceX matter more for institutional allocators than seed funds do. Seed and Series A rounds have become mislabeled in AI; the dollar size now matters more than the stage label and can create dangerous valuation and follow-on pressure. The best venture strategy depends on fund size: your fund size implicitly constrains the exit size you need to deliver a good return. AI is shifting the bar for venture-scale growth; companies raising from top-tier investors must often show much faster traction than traditional SaaS companies once did. Many AI startups are solving problems that may only exist for months, which is making investment less durable and more crowded. Model-agnostic architectures and cheaper open-weight models reduce dependence on frontier labs and can lower token costs over time. The most defensible AI businesses will combine customer data, routing, and workflows with an increasing-returns or trust/verification moat. Legacy companies can survive an AI transition only if they actively re-architect products and incentives; passive ‘wait and see’ approaches will likely fail. Non-IPO exits and PE roll-ups are likely to become more common for older software businesses with slower growth or lower strategic relevance. The panel sees public trust, labor displacement, and inequality as major risks that the industry has not yet adequately addressed.

Data Points: Cowboy VC Fund 4: $230 million - Aileen Lee noted Cowboy was still investing Fund Four after raising it in 2023. Cowboy opportunity fund: $140 million - Raised alongside Fund 4 in 2023. Floodgate Fund 8 filing: $130 million - Mike Maples said Floodgate filed to raise Fund 8 in May. Lerer Hippeau Fund 9: $200 million - Ben Lerer said the firm closed Fund 9 last year. Floodgate liquidity returned: $350 million+ - Maples said Floodgate returned a little over $350 million in the last two years. Bending Spoons IPO price: $29/share - Mentioned as pricing above its $26-$28 range. Bending Spoons valuation: $18.5 billion non-diluted - Price implies this valuation versus a prior $11 billion mark. Bending Spoons revenue growth: 2x YoY - Alex cited S-1 data showing revenue doubled year over year. Bending Spoons operating/net income: Positive in Q1 2025 on GAAP basis - Cited as evidence of strong health. Typical AI seed/A size cited: $20 million seed at $100 million pre - Aileen described a new pattern of massive early rounds. Next-round target for large seed rounds: $200 million to $300 million - Aileen said companies raising big seeds must be prepared for much larger follow-ons. Startup growth bar in AI era: 1-to-4.5 or 1-to-5 - Aileen said investors now expect much faster growth from seed/A/B companies. Former enterprise rule of thumb: 1-to-3 then 1-to-9 then 1-to-25 - Referenced as the old triple-triple-double growth model. Labor share of GDP: ~57% - Alex cited a Fred chart showing a sharp decline from the 1950s. Historical labor share of GDP: High 60s - Referenced as the earlier postwar level. KeepSafe investor return: $1.5 million invested; >$10 million returned in dividends - Mike used this to illustrate profitable-growth and dividend strategies. Data center/AI infra raises: $800 million (Etch), $850 million (Lightmatter) - Examples of massive infrastructure financing tied to AI compute demand. DG Matrix raise: $20 million - Example of funding for solid-state transformers for AI infrastructure. Aileen’s cited LP exposure shift: 25% target to 45% actual venture exposure - Used to illustrate denominator effects and venture overexposure from paper gains. SmarterDX exit: $1 billion - Mike cited it as a great seed outcome that mostly matters to early-stage funds.

Pivotal Quotes: "Your fund size is your strategy." — Mike Maples: He argued that a fund’s size determines the scale of exits it must pursue and what types of companies it can support. "Companies get bought, not sold." — Ben Lerer: He used this to explain why founders should not force exits but should seriously consider strong inbound acquisition offers. "What’s scarce is correctness and proof of correctness." — Mike Maples: He argued that as AI output proliferates, verification and trusted validation become the valuable layer.

Implications: Listeners should expect faster venture cycles, more selective capital, and increasing pressure for real moats. In AI, winners may be companies that combine frontier models with trust, data, and workflow advantages—not just flashy demos or massive early rounds.