The Collapse of Terminal Value Is a Reallocation, Not a Loss
Chamath Palihapitiya published a thought experiment this week: The Collapse of Terminal Value. The argument is straightforward. If AI erodes competitive moats fast enough, markets can no longer rationally assign value to what a company earns in year ten or beyond. Sixty to eighty percent of large-cap equity value sits in terminal value assumptions. Remove those assumptions and the S&P 500 reprices from $58 trillion to roughly $14 trillion - a 75% drawdown.
Chamath frames this as a crisis. But capital doesn't vanish. It reallocates.
Where the money goes
If incumbents lose their moats, someone is eating their lunch. The capital exits the incumbent and enters the disruptor. The total market cap disperses. Instead of $44 trillion concentrated in five balance sheets, it spreads across hundreds of smaller companies. More competitors, more surface area, more innovation.
The picks-and-shovels layer can't be disrupted by its own output. NVIDIA, TSMC, and the hyperscalers sell the weapons that destroy moats. They become the new terminal value. You can't vibe-code a fab.
Beyond equities: Bitcoin absorbs the "digital gold" rotation. Treasuries capture the flight from duration risk. The money finds new homes in assets that don't depend on terminal value assumptions. The pie changes zip codes.
Orderly vs. chaotic
The real risk is the velocity of the drawdown.
An orderly reallocation - compressed innovation cycles, capital cascading from mature incumbents to seed-stage disruptors over years - produces better outcomes than the status quo. Shorter company half-lives mean capital turns over faster. You don't park $1 billion in a 30-year moat. You deploy $100 million into 20 companies over five years and recycle the winners. Seed and Series A become the new public equities.
A disorderly reallocation - the repricing happening in weeks, triggered by a single earnings miss or model release - is a liquidity cascade. That's the scenario Chamath is warning about.
The difference between the two is whether the market processes the terminal value collapse gradually or all at once. If gradually, the end state is genuinely better: more competition, faster capital velocity, deeper allocation into the startup ecosystem. If all at once, the transition mechanics overwhelm the destination.
Moats protect mediocrity
A moat is a structural barrier that prevents competition. It doesn't mean the product is good. It means the product is protected - by switching costs, regulatory capture, network effects, or distribution lock-in.
The moat-era economy rewarded defensibility over quality. Companies optimized for entrenchment, not excellence. AI flips this. When the barrier to compete drops to near-zero, only genuinely superior products survive. GDP shifts from rent-extraction to value-creation.
This is painful for shareholders of the old regime. It is excellent for everyone else.
Chamath is right about the mechanism. Terminal value assumptions are fragile in a world where moats are temporary. But he frames the outcome as a loss when it's a transfer - from concentrated, moat-protected incumbents to the infrastructure layer, the disruptors, and the assets that don't need a moat to hold value. The market doesn't get smaller. It gets more efficient.
[^1]: Palihapitiya, Chamath. "The Collapse of Terminal Value - What Happens If AI Makes Every Moat Temporary?" X. May 2026. [^2]: Dror, Itay. "The Collapse of Terminal Value - My 2 Cents." Substack. March 2026.