Source

Most rigorous quantitative forced-selling model found for the AI mega-IPO lane. Paul Dunn (founder/lead engineer, AlgorithmicFIRE) models the three-IPO stack as a single liquidity event.

Key claims (author's model, explicit assumptions):

• Combined SpaceX ($1.75–2.0T) + OpenAI ($852B–1.0T) + Anthropic (~$950B) = ~$3.0–3.8T, or roughly 5–6% of S&P 500 market cap.
• ~$49B forced S&P rebalancing for SpaceX alone; ~$97B cumulative 12-month forced selling across all three names as they phase into index families.
• Applies the Inelastic Markets Hypothesis ~5× multiplier → ~$485B implied market-cap displacement.
• Models rebalance-day selling as a share of daily volume for funding-pocket names (e.g., AAPL ~21% raw / ~105% amplified; GOOGL ~20% / ~100%).

Lane use: this is the liquidity-drain hypothesis quantified. The ~$49B SpaceX-alone figure sits at the low/middle end of the wide public range (cf. the fragile "$950B JPMorgan" number carried by TradingKey). Treat as one modeled estimate with stated assumptions, not a primary disclosure. Pairs against the Friedman float-math "nothingburger" steelman as the two falsifiable poles.