Anthropic lists tomorrow at $83/share — $81.3B valuation, largest IPO ever, ahead of SpaceX's $75B in June.
The market already paid for it. SPY down 5.55% over the last 20 days. QQQ down 9.06%.
The question: does the drain continue once $81B locks into one name, or was the selling the drain?
We tested 54 mega US IPOs since 2007. Forward returns from listing-day close:
20 days after listing, mega-IPOs didn't drain liquidity — the market did better than usual.
• SPY 20-day forward: +1.90%, 83% win rate (p < 0.0001)
• QQQ 20-day forward: +2.44%, 80% win rate (p = 0.0001)
SPY on any random day: +0.78% over 20 days. After a mega-IPO it beats baseline by +1.12% (p = 0.004).
The $15B+ tier: Visa +5.16%, GM +3.62%, Meta +3.59%, Alibaba -6.09%. Three of four positive.
Now the live tests. 2026 gave us two IPOs bigger than anything in the study:
1. SpaceX ($75B, June 12): SPY +1.36%, QQQ -0.23% over 20 days
2. SK hynix ($26.5B, July 10): SPY +2.43%, QQQ -0.34%
SPY rose both times, short of the +1.90% average. QQQ flat both times. No drain, but no tech-led rally either.
Our study starts at listing, so it can't tell us whether the selling was caused by the IPO. Closest precedent is Meta 2012: SPY fell 4.98% in the 20 days before, then rose 3.59% in the 20 after.
The caveat: Anthropic is 3.7x Alibaba, the largest IPO in our historical sample. Two live tests isn't a sample. Direction has held. Magnitude at this scale is still an open question.
The 20-day clock starts at tomorrow's close. We'll post the result either way.
The market already paid for it. SPY down 5.55% over the last 20 days. QQQ down 9.06%.
The question: does the drain continue once $81B locks into one name, or was the selling the drain?
We tested 54 mega US IPOs since 2007. Forward returns from listing-day close:
20 days after listing, mega-IPOs didn't drain liquidity — the market did better than usual.
• SPY 20-day forward: +1.90%, 83% win rate (p < 0.0001)
• QQQ 20-day forward: +2.44%, 80% win rate (p = 0.0001)
SPY on any random day: +0.78% over 20 days. After a mega-IPO it beats baseline by +1.12% (p = 0.004).
The $15B+ tier: Visa +5.16%, GM +3.62%, Meta +3.59%, Alibaba -6.09%. Three of four positive.
Now the live tests. 2026 gave us two IPOs bigger than anything in the study:
1. SpaceX ($75B, June 12): SPY +1.36%, QQQ -0.23% over 20 days
2. SK hynix ($26.5B, July 10): SPY +2.43%, QQQ -0.34%
SPY rose both times, short of the +1.90% average. QQQ flat both times. No drain, but no tech-led rally either.
Our study starts at listing, so it can't tell us whether the selling was caused by the IPO. Closest precedent is Meta 2012: SPY fell 4.98% in the 20 days before, then rose 3.59% in the 20 after.
The caveat: Anthropic is 3.7x Alibaba, the largest IPO in our historical sample. Two live tests isn't a sample. Direction has held. Magnitude at this scale is still an open question.
The 20-day clock starts at tomorrow's close. We'll post the result either way.