He manages the best-performing stock mutual fund of the past 25 years—here's his top advice for investors

Baron Opportunity Fund delivered a 13% annualized return over the 25 years ended 2025, outperforming the S&P 500's 6.8% annualized return. Manager Michael Lippert, in his 20th year, credits a growth-focused, theme-driven process that targets durable, multi-S-curve companies and leans on deep analyst research (example: early Nvidia data-center/AI investment beginning in 2018). The piece stresses a fact-based, long-term (rolling five-year) stock-picking approach and institutional learning from mistakes rather than short-term trading.

Analysis

Think in terms of stacked S‑curves, not single-product stories: a company that can sequentially monetize multiple adjacent S‑curves earns a structural premium on revenue CAGR because each new curve leverages existing distribution, data and margins. That implies winners will be businesses that combine hardware-enabled moats with proprietary software ecosystems — the durable economic value is in the ecosystem capture more than raw silicon performance, and that persistence translates into multi‑year cash flow visibility (5–10+ year windows) rather than quarterly inflections.

Second‑order beneficiaries are underpriced and often overlooked: OSATs, HBM memory suppliers, power delivery and advanced cooling vendors face 6–12 month capacity tightness as cloud AI deployments ramp, creating pricing power upstream even if end‑market unit growth slows. Equally important are ML software stacks and model‑ops players that convert raw compute into recurring revenue; the cloud infra vendors that sell integrated AI stacks (infrastructure + managed services) win twice — as hardware customers and as sticky SaaS clients.

Key reversal risks are structural rather than cosmetic: rapid model efficiency gains (algorithmic or hardware-software co‑design) could meaningfully reduce training compute per useful output within 12–36 months, compressing demand; export controls or a sudden entrant with a genuinely superior accelerator architecture could reprice winners quickly. Monitor data‑center capex guidance, HBM lead times, and enterprise adoption metrics as primary catalysts — these are higher‑information, multi‑quarter signals that precede durable re‑rating or derating.

Key Decisions for Investors

• Long NVDA via a 12–24 month call spread (buy LEAP call / sell higher strike): size 2–3% portfolio. Rationale: capture multi‑year AI/data‑center curve while capping downside premium. Target 2x payoff; stop‑loss at -40% of premium if next two quarters’ data‑center order cadence and cloud capex commentary disappoint.
• Long AMZN (12–18 months) overweight into cloud infra exposure: size 3–4% equity. Rationale: benefits from increased cloud AI spend and ad/retail mix. Trim on 20–30% move and re‑evaluate vs actual AWS/advertising monetization metrics; consider pairing with short AAPL if consumer discretionary signals weaken.
• Pair trade: long GOOGL/GOOG (9–18 months) and short a basket of gas‑lamp AI small caps (size net market‑neutral ~1.5%): harvest platform monetization and ad/AI services upside while hedging headline volatility. Set target +25% on long leg vs capped downside via protective puts; exit if Google lags peers on AI product monetization by two consecutive quarters.
• Event hedge on commoditization risk: buy a 6–9 month MORN put spread sized 0.5–1% to protect against research‑as‑a‑service compression from generative AI tools. Risk/reward: limited premium outlay, asymmetric payoff if subscription ARPU guidance weakens or churn accelerates.