New Fed Chair Kevin Warsh Has Big Plans. They Could End the Trump Bull Market.

Kevin Warsh’s push to shrink the Fed’s $6.7 trillion balance sheet back toward pre-2008 levels could tighten financial conditions even if short-term policy rates are cut. The article argues that higher long-term rates and borrowing costs would pressure the capital-intensive AI build-out, which is increasingly debt financed and has helped drive recent stock gains of 18% for the S&P 500, 12% for the Dow, and 28% for the Nasdaq since March 30. The risk is a broad market rotation out of speculative assets if QT pushes yields higher.

Analysis

The market is treating AI as a duration trade, but the more important exposure is credit duration: once financing costs move up at the long end, the marginal AI project becomes less about model quality and more about hurdle rates. That shifts the winners from hyperscale capex beneficiaries to firms with internal cash generation and the ability to self-fund multi-year builds; externally financed infrastructure names, niche equipment vendors, and smaller AI adjacency plays are the most vulnerable to a reset in credit appetite.

The second-order effect is that QT can pressure equity multiples even if short rates are cut, because the discount rate that matters for 5- to 10-year AI cash flows is the term premium, not the policy rate. If balance sheet runoff is more aggressive than expected, expect the first damage to show up in spread widening, then in delayed data-center commitments, and only later in headline earnings revisions. That lag creates a window where the stocks most levered to AI optimism can de-rate before fundamentals visibly deteriorate.

Consensus is likely underestimating how self-reinforcing the capital cycle has become. A modest rise in borrowing costs can slow capex just enough to cool supplier order books, which then tightens labor and equipment demand, leading to slower revenue recognition across the ecosystem; that feedback loop would matter more than a simple multiple compression story. On the other hand, megacap cash-rich platforms may actually gain relative share if higher rates force smaller competitors to retrench, making this a dispersion trade rather than a broad market call.

Near term, the cleanest risk is not an instant equity crash but a 2-6 month valuation unwind in the highest-beta AI beneficiaries if term yields grind higher. The catalyst to monitor is any explicit signaling on balance sheet runoff pace, Treasury issuance absorption, and credit spreads in data-center financing; if financing markets tighten before earnings weaken, the move can accelerate quickly. If the Fed softens on QT, the trade loses urgency, but if runoff is real, this is one of the few policy shifts that can simultaneously pressure growth multiples and AI capex expectations.

Key Decisions for Investors

• Short NVDA vs long AMZN or GOOGL over the next 1-3 months: use this as a dispersion trade on who can self-fund AI spend; target a 5-8% relative move if long-end yields back up and capex discipline becomes the market focus.
• Buy 3-6 month put spreads on a basket of high-beta AI hardware/supply-chain names, funded by selling farther out-of-the-money puts on cash-rich megacaps; this expresses downside in financing-sensitive names while limiting theta bleed.
• Reduce exposure to levered data-center and AI infrastructure beneficiaries until credit spreads stabilize; the setup favors waiting for a post-QT widening in spreads before re-entering, rather than chasing momentum now.
• Pair long GOOGL / short INTC on a 2-4 month horizon: the former can absorb higher capital costs through internal cash flow, while the latter is more exposed to financing and execution sensitivity if the cost of capital rises.
• If 10-year yields continue to grind higher for two consecutive weeks, add tactical shorts to the most crowded AI momentum basket; stop out quickly if policy rhetoric turns dovish on balance sheet runoff.