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

Berkshire’s Apple position, which once approached ~40% of its portfolio, still represented 23% at end-2025; Warren Buffett said on CNBC he sold Apple “too soon” but is “very happy” it remains the largest holding and would consider buying more if the stock declines (Apple is down >6% YTD). New Berkshire CEO Greg Abel indicated Apple is a core position with expected limited trading; Buffett praised Tim Cook. Key drivers and risks: Apple has not made the large AI capex commitments seen in peers but has partnered with Google’s Gemini and may monetize AI via subscriptions, while tariff exposure from China manufacturing and questions about iPhone 17 innovation persist.

Analysis

Apple’s tepid public AI posture is a latent market bifurcation: it preserves margin profile today but cedes near-term monetization to cloud/AI vendors that power device-level features. That makes Google (Gemini/GCP) and Nvidia (inference/training silicon) the likely incremental beneficiaries of any Siri/AI feature rollout, shifting recurring service revenue and cloud GPU demand away from Apple-captive pathways and toward third‑party infra over the next 6–24 months.

Berkshire’s reduced trading activity in the position is a supply-side stabilizer — fewer forced large blocks hitting the tape compresses downside volatility and raises the bar for a sell-off-driven re-rating. The key catalysts that could reverse the mild positive bias are clear: tariff escalation (weeks–months, immediate margin pressure), an Apple announcement of aggressive in-house AI capex or M&A (3–12 months, large reallocation of supply chains), or rapid subscription ARPU proves unattainable (12–36 months) — any of which could swing returns by +/-20–40% versus consensus.

Second-order winners: GOOGL receives durable stickiness from device OEM partnerships (increases GCP procurement optionality), NVDA captures incremental datacenter inference volume via cloud partners rather than device OEMs, and Intel remains a marginal beneficiary unless it wins cloud design share. For portfolio construction, this argues for expressing AI exposure through cloud + gpu incumbents and using Apple as a defensive, cash-generative core rather than the primary lever for AI upside.

Key Decisions for Investors

• AAPL core overweight (2–4% net portfolio): buy stock or buy 12‑month LEAPS (~1yr) and sell 25–35% OTM calls to finance (~buy write). Timeframe: 6–18 months. R/R: downside tail ~15–25% if tariffs/innovation stall; upside 30–60% if subscription/AI monetization ramps — income cushions ~2–4% p.a.
• GOOGL tactical long (1–2%): buy stock or buy 18–24 month calls to capture Gemini/GCP monetization to Apple users. Timeframe: 12–24 months. R/R: asymmetric — limited capital expenditure risk vs potential 30–50% upside if cloud monetization accelerates; downside ~20% on ad softness or regulatory action.
• NVDA conditional convex play (0.5–1%): enter a 6–12 month call spread (buy 15–25% OTM call, sell 45–60% OTM call) to get leveraged exposure while capping premium spend. Timeframe: 3–12 months. R/R: good if datacenter inference demand continues; be explicit about IV crush risk and plan to exit on 30–40% rally or 25% drawdown.
• Pair hedge: long BRK.B / short AAPL (dollar neutral, 6–12 months) to express conviction in Berkshire’s capital allocation optionality while neutralizing device cyclicality. Timeframe: 6–12 months. R/R: trade profits if market rotates to active capital allocators; main risk is correlated move if both rally together driven by broad market multiple expansion.