Billionaire Warren Buffett Says He Sold His Favorite Stock "Too Soon." Should Investors Back Up the Truck?

Berkshire's Apple stake was 23% of its portfolio at end-2025 (down from ~40% peak) and Warren Buffett admitted he sold Apple "too soon." Buffett praised Tim Cook, said he's happy Apple is Berkshire's largest holding, and noted Berkshire could buy materially more if the stock keeps declining but "not in this market," while new CEO Greg Abel labeled Apple a core position with limited expected activity. Apple is down over 6% YTD, lacks a clear standalone AI strategy (but has partnered with Google's Gemini), and faces tariff/supply-chain risks despite strong iPhone 17 sales and ongoing large capital returns to shareholders. Analysts (e.g., Wedbush) see AI monetization and subscription opportunities for Apple, supporting the buy case alongside Buffett/Abel's public backing.

Analysis

Berkshire’s shift toward formally “core” holdings and controlled activity is a governance signal that changes market microstructure for large-cap liquidity: when a $200–300bn buyer commits to limited trading, it reduces a persistent marginal bid in drawdowns and raises the probability that steep sell-offs will be deeper and longer before other institutional buyers step in. That second-order effect elevates tail volatility for mega-cap consumer names while improving portfolio-level concentration controls at the parent — a net positive for BRK.B idiosyncratic risk but a potential negative for any stock that had relied on Berkshire as the marginal stabilizer.

Apple’s capital-light posture on AI infrastructure is not a sign of strategic absence but of optionality preservation; by avoiding heavy fixed-cost commitments it retains superior FCF optionality to buy compute, partner with model providers, or vertically bundle AI features into subscriptions. That approach accelerates vendor economics for cloud/AI infra providers (directly favoring datacenter compute vendors and cloud AI stack providers) while concentrating operational risk in China-based manufacturing and trade-policy vectors — a policy shock could compress volumes and margins quickly because Apple’s hardware margins are tight relative to its services flywheel.

Near-term catalysts to watch are product-cycle announcements, WWDC feature rollouts that convert device-level AI into recurring revenue, and large-cap fund rebalancing windows that could mechanically change flows into BRK.B and its holdings. Tail risks include a rapid shift in AI monetization expectations (either a “miss” in subscription uptake or a costly capex pivot) and escalation of tariffs or export controls; both can flip the current optionality value into an execution or geopolitical haircut within 6–18 months.

Key Decisions for Investors

• AAPL: Staged accumulation via buy-and-hold with covered-call financing — initiate on a 10–15% pullback from current levels, build to a 2–3% portfolio weight over 3 months. Sell 12-month OTM call spreads (sell ~10–15% OTM, buy farther OTM) to reduce carry; target IRR ~12–18% if services/AI monetization begins to show traction. Stop-loss: trim if iPhone unit comps decline >20% YoY or China manufacturing sanctions materialize.
• BRK.B: Tactical overweight (small-sized, 1–2% incremental weight) into the next 3–12 months to capture governance tailwind and lower volatility profile. R/R: low volatility, expect 100–200bps of alpha vs large-cap benchmark if management executes capital allocation; cut if cash redeployments become highly active (signals increased trading risk).
• NVDA vs AAPL pair: Long NVDA Jan-2027 2x call-spread (bull-call) to express AI infra upside funded by short AAPL Jan-2027 vertical call spreads (equal notional) to hedge device cyclicality. Time horizon 12–36 months; target asymmetric payoff where NVDA upside >3x downside from AAPL funding; unwind if AI server spend cadence decelerates for two consecutive quarters.
• GOOGL: Buy selective exposure to cloud/AI stack (long-dated calls or core equity) as a beneficiary of Apple’s partner-first AI route and broader enterprise AI spend. Horizon 12–24 months; risk: regulatory clampdown on data/AI monetization — size accordingly (1–1.5% notional) and use calendar spreads to cap premium decay.