定量视角：少数伟大人物（英）2026

QUANT RESEARCH The Magnificent Few A quantitative perspective on US VC Introduction PitchBook Data, Inc. Research Though liquidity has begun to show signs of life, venture’s recovery remains uneven. Macroeconomicvolatility, fueled by geopolitical uncertainty, continues to complicate capital planning and exit timing.Despite a handful of high-profile debuts, the median VC-backed IPO has underperformed, reinforcingissuer hesitation and leaving the exit pipeline drier than expected. Susan HuQuantitative Research Analystsusan.hu@pitchbook.com At the same time, venture’s top-heavy capital concentration has reached unprecedented levels. In 2025,OpenAI, Anthropic, and xAI captured more than $50 billion in deal value. That historic figure was quicklyeclipsed in February 2026, when OpenAI alone raised a $110 billion venture round, equivalent to roughlya third of all deal value raised in 2025. Jacobie FullertonQuantitative Research Analystjacobie.fullerton@pitchbook.com The market has rarely exhibited this degree of sector-level concentration, reflecting AI’s uniquepotential and endless opportunities to be implemented across every sector. The central question nowis whether incumbents can adapt quickly enough for AI to serve as a tailwind rather than a source ofdisplacement. The market’s recent “SaaSpocalypse” suggests investors are, at least for now, bracingfor the latter. Emily ZhengSr. Analyst, Venture Capitalemily.zheng@pitchbook.com Beneath the headline mega-rounds lies a more sobering reality for VC fundraising. Nearly every vintagesince 2016 remains behind historical pacing in returning capital to LPs, with recent vintages remainingparticularly slow to distribute, as many pandemic-era markups have yet to fully reset to reflect currentfundamentals. Similar to dealmaking, fundraising has been dominated by the largest firms. The top 10 venture fundsby capital raised accounted for 32.7% of total capital raised in 2025, near decade highs forconcentration. With overall fundraising subdued, emerging managers and smaller firms face mountingpressure in 2026 as cautious LPs funnel commitments to established, multi-billion-dollar platforms withoften privileged access to competitive deals. Venture’s concentration creates a paradox. On one hand, scale is an asset in a capital-intensive AI racewhere compute, talent, and data advantages compound. On the other, elevated entry valuations andcrowded cap tables increase the risk of both big wins and big losses. In a market defined by delayedliquidity and extended holding periods, underwriting discipline and portfolio construction matter morethan thematic exposure alone. Contact Key takeaways •Rate uncertainty clouds the liquidity recovery. Dispersion in policymakers' 2026 projectionshas made the easing path less predictable. For venture, this complicates the stabilizationin discount rates needed to support exit timing and pricing, particularly through publiclistings, which historically cluster during periods of low volatility. (page 5-7) •AI is reshaping venture less as a vertical and more as a horizontal force. Across SaaSsegments, AI-tagged companies account for as much as 69% of VC-backed SaaScompanies within select segments. Yet public markets are skeptical that incumbents willcapture the value—nearly all recently listed VC-backed SaaS stocks have declinedsignificantly year-to-date, suggesting investors view AI as a displacement risk rather than atailwind. (page 23-24) •Exit activity is recovering on the surface, but the gains are narrow and fragile. Short-termexit trends have crossed above long-term baselines, yet acquisitions still account fornearly three-quarters of exit count and two-thirds of exit value. Public listings represent ashrinking share of total exits, and the median VC-backed IPO has underperformed itsbenchmark by 44.5% within 120 days of listing—reinforcing hesitancy across the pipeline.(page 10-13) •Bargaining power has bifurcated across sectors. Our early-stage dealmaking indicatorshows a gap between AI deals, which remain firmly startup-friendly, and non-AI deals,where investors retain significantly more leverage. Investor protection provisions likecumulative dividends and liquidation participation have risen post-2022 but remain notablylower in AI transactions, reflecting a two-tier negotiating environment. (page 25-26) •Secondary markets are expanding to fill the liquidity gap, but concentration persists.Annualized direct secondary market value reached $91.7 billion in Q4 2025. Yet the topfive companies accounted for 55.6% of trading value, and companies that last raised at the2021 peak are trading at median discounts exceeding 60%. (page 14,17) •Valuation discipline varies sharply between AI and non-AI, and paying up changes thereturn profile. At Series D+, median AI pre-money valuations exceed non-AI peers by morethan 3x. Higher-priced seed rounds have historically correlated with better exit rates, butportfolio simulations based o