220 unicorn startups have lost their billion-dollar status. Here’s why

More than 220 U.S. unicorns have lost their billion-dollar status, with startups that raised in 2021 now worth 68% less on average and those from 2022 down 52%. The AI boom has redirected more than $250 billion into firms like OpenAI and Anthropic, pressuring pre-ChatGPT business models, especially SaaS and direct-to-consumer brands. Recent transaction comps underscore the reset: Stash was sold for $425 million versus $660 million invested, while Step changed hands below its $500 million fundraising.

Analysis

The market is not just repricing failed venture promises; it is repricing the terminal value of labor-intensive software in an AI-native world. The biggest second-order effect is that the cost advantage is shifting from capital-efficient growth to model access and distribution, which means incumbents with embedded workflows face a slower, harsher multiple compression than headline revenue declines suggest. That favors platform owners and infrastructure enablers while creating a hidden liability overhang for any SaaS business whose usage is tied to seats rather than outcomes.

The damage is likely to cascade through the venture ecosystem before it shows up in public comps. Late-stage funds that marked portfolios off 2021 metrics will be forced into down-rounds, structured financings, and acqui-hires, which can create a feedback loop of talent and customer churn as weaker companies cut support and product velocity. The more exposed segment is not consumer DTC but software businesses with high gross margin and low switching costs; if AI compresses implementation time and service intensity, their moat erodes faster than revenue growth can be replaced.

There is a near-term bifurcation in beneficiaries: AI infrastructure and adjacent software enhancement tools should continue to gain share, while generic workflow SaaS should see multiple pressure for quarters, not days. The risk to the bearish case is that incumbents buy time by bundling AI features and shifting pricing to usage or value-based contracts, but that usually requires 12-24 months and still leaves the incumbent ceding economics to model providers. In M&A, distressed VC assets become cheap option value for strategic buyers, which supports a wave of sub-$1B takeouts but also locks in a lower clearing price for the entire private market.

The contrarian view is that the market may be overdiscriminating on “AI threat” versus “AI adoption,” especially for enterprise software with data gravity and compliance moat. Some fallen unicorns are less broken business models than broken financing structures; if they can cut burn aggressively and re-underwrite to profitability, public-market analogs may stabilize before private marks do. But the burden of proof has shifted decisively: companies now need to show AI-driven productivity gains translating into margin expansion, not just narrative relevance.

Key Decisions for Investors

• Short basket of workflow-heavy SaaS names with low net retention and high sales efficiency dependency; express via equal-weight short or put spreads over 3-6 months. Target names where AI features are easy to copy and customer seat counts are stagnant; risk/reward is strongest because multiple compression can outpace modest fundamental beats.
• Long NVDA on pullbacks versus short a SaaS basket as a pair trade for 2-4 quarters. The thesis is that model spend and inference demand accrue to infrastructure first, while application-layer monetization gets competed away; use the pair to isolate AI capex winners from AI margin victims.
• Maintain selective long ADBE on 6-12 month horizon as a beneficiary of workflow augmentation rather than replacement. Adobe has a better chance than pure-play SaaS to monetize AI through pricing power and embedded creative workflows, but size modestly because valuation can still de-rate if usage monetization stalls.
• Avoid chasing public-market takeout stories in distressed venture names; prefer options over outright equity if expressed at all. Event-driven upside exists, but the clearing-price signal is negative for the group, so upside is capped while downside remains tied to further down-rounds and refinancings.
• Watch GRABW only as an opportunistic long if the market overreacts to private-market distress; it benefits if distressed fintech/consumer assets are acquired at lower multiples. Keep it small and tactical, because the broader read-through is still negative for growth-duration assets.