Prediction market prices 78% chance Anthropic will be worth more than Berkshire Hathaway by year end

Polymarket is assigning a 78% probability that Anthropic will reach a $1.5 trillion valuation by the end of 2026, implying a potential market value roughly $500 billion above Berkshire Hathaway and near Meta's scale. The piece reflects bullish speculative sentiment around AI private-market valuations rather than a concrete operating update. Market impact is likely limited, but it underscores elevated investor appetite for frontier AI assets.

Analysis

The market is not really pricing Anthropic here; it is pricing the durability of a very small set of implied assumptions: relentless frontier-model capex, no material cost compression, and a winner-take-most equilibrium in enterprise AI. If that narrative keeps inflating, the second-order winner is not necessarily the model vendor but the infrastructure stack—GPU suppliers, hyperscale cloud, and power/thermal providers—because a $1.5T private valuation implies a multi-year demand curve for compute that is harder to arbitrate away than the equity itself.

For META, the relevance is subtle: a rising private-market mark for a top-tier model lab can actually sharpen the strategic value of in-house AI R&D and distribution. The market may read this as validation that AI monetization still has room to run, which supports multiple expansion for scaled incumbents with user funnels and ad data, but it also increases the pressure to show that spend converts to measurable product lift within 2-4 quarters. BRK.B is more of a relative loser in sentiment terms: when speculative private assets command near-megacap pricing, capital tends to rotate away from low-volatility compounders, even though that does not change Berkshire’s intrinsic economics.

The key risk is that prediction markets can become self-referential in illiquid names: a high probability estimate can be driven by narrative momentum rather than true funding capacity. If AI software revenue growth decelerates or if model training costs plateau while monetization remains experimental, the re-rating can unwind fast over months, not years, especially if one or two flagship AI launches miss expectations. The contrarian view is that the consensus is overstating the scarcity value of a model lab; over a 24-month horizon, the more durable moats may sit with distribution and balance-sheet scale, not the private company being bid up in the headlines.

Key Decisions for Investors

• Long META vs. short a basket of speculative private-AI sentiment proxies via options around the next 1-2 earnings cycles; thesis is that scaled distribution monetizes AI faster than standalone model labs can prove durable unit economics.
• Buy META call spreads 3-6 months out, funded by selling further OTM calls; attractive if the market continues to reward visible AI monetization and capex discipline with multiple expansion.
• Reduce relative exposure to BRK.B against a growth basket for the next 1-3 months only if factor flows are chasing AI momentum; use as a temporary pair, not a fundamental short.
• If Anthropic-style private valuations keep moving higher, go long infrastructure beneficiaries on pullbacks (semis/cloud/power) rather than the private AI cap table itself; better liquidity and cleaner expression of the compute thesis.
• Set a reversal trigger: if two consecutive AI-platform results show slower-than-expected revenue conversion or rising inference costs, fade the sentiment trade and rotate back into quality compounders like BRK.B and META.