Warsh's Bluff, AI Bailout Risk & Bitcoin's Next Leg | Lepard & St. Onge

— Kevin Warsh presents as hawkish but is likely more dovish than his rhetoric suggests; rate cuts are probable before the election to support economic growth and political objectives.

— The Federal Reserve's fundamental constraint is that growing debt requires growing money supply to service interest payments—a mathematical inevitability that will force eventual monetary expansion regardless of the chair's stated position.

— AI is experiencing a commoditization crisis: competitive moats vanish in weeks, making application-layer companies vulnerable; infrastructure (picks and shovels) captures more durable value than AI application developers themselves.

— Legitimate energy cost grievances exist around AI data centers, but nuclear power expansion and competitive energy markets could resolve capacity constraints; some regions are now requiring data centers to fund their own power.

— Bitcoin's volatility is normalizing over time (corrections shrinking from 90% to ~55%); this mirrors gold's boom-bust cycles during non-dominance and does not signal fundamental failure.

— Michael Saylor and MicroStrategy: the preferred equity liquidation cascade was healthy feedback limiting leverage; Saylor's evolving playbook (including potential Bitcoin sales when stock trades at discount) reflects mature capital allocation, not deception.

— Position for rate cuts, not hikes. Market pricing assumes tightening; the institutional bias of the Fed and political pressure before elections favor cuts in July or September 2025. Reposition sound money holdings accordingly—the monetary debasement thesis remains intact despite recent weakness.

— Distinguish leverage from core Bitcoin exposure. MicroStrategy (MSTR) is a levered bet suitable only for aggressive allocators; own Bitcoin in self-custody first, then consider MSTR or ETFs if you want additional exposure and understand 70% drawdowns are possible in bear markets.

— AI infrastructure (energy, chips) is more durable than AI applications. If considering AI-adjacent investments, focus on picks-and-shovels (power, semiconductors, data center operators) rather than model developers facing rapid commoditization and bailout risk.