Why this could go wrong
A concentrated, leveraged bet that the physical infrastructure deficit deepens faster than the market expects. If he's wrong, the leverage works in reverse.
Leverage
~3.5× gross exposure on ~$1.5B AUM. In a drawdown, small-cap BTC miner and neocloud positions with limited float face severe liquidity pressure. Q4 2025 est. −21% illustrates this.
Timeline
Entire thesis is calibrated to AGI by 2027. If GPT-5 or GPT-6 disappoints vs OOM projections, the capex supercycle thesis weakens materially. Scaling laws plateauing = thesis deflates.
Geopolitics
Taiwan invasion, Iran escalation, or Chinese chip export restrictions could break supply chains. Regulatory intervention could slow buildout. AGI timelines and Taiwan invasion windows converge around 2027.
Liquidity
Small/mid-cap concentration. 9.4% stake in Core Scientific (13D) cannot be exited quickly. Options-heavy book amplifies both gains and gap risk.