The certainty trap ↗Oakmark argues that AI investing should avoid binary forecasts and favor bonds priced for a broad range of outcomes. It favors Meta, Oracle and select single-tenant data center bonds for their resilient credit profiles and contractual protections, while avoiding highly leveraged NeoCloud issuers whose economics depend on sustained compute demand.2Q 2026 —META$1.4TORCL$413B
Wrong in the right direction ↗Oakmark retained similar weights in Samsung and ASML after revising valuations upward to reflect AI-driven demand for high-bandwidth memory and lithography equipment. It credited Samsung’s memory scale and process leadership, ASML’s EUV-tool monopoly, and Intertek’s board for accepting a private-equity offer that recognized self-help opportunities and execution risks.2Q 2026 —ASML$673BSSNLFdelistedIKTSF$12B
The discipline to stay boring ↗Oakmark Fund retained its value discipline rather than buy AI hardware companies without a sufficient margin of safety, despite their dominance in major value indexes. The portfolio favors businesses such as Corebridge Financial while expecting holdings including Capital One, AIG, Alphabet and Amazon to benefit from AI adoption.2Q 2026 +2.5%AMZN$2.8TCOF$133BCRBG$14BGOOGL$4.2T
Why didn’t we do better when value outperformed? ↗Oakmark Fund argues that Russell Value’s advance was driven by expensive semiconductor-related growth stocks rather than the low-P/E rebound it expected. The fund added eight positions, sold energy and industrial holdings, and redeployed into software and financials, where it sees AI fears and unusually low relative valuations improving the risk-reward case.1Q 2026 -2.5%—
