3 Monster Stocks to Hold for the Next 20 Years

The article highlights long-term bullish cases for Okta, ASML, and Enbridge, anchored by strong operating fundamentals: Okta’s 2025 EPS rose 208% to $1.31 on 12% revenue growth, while ASML reported Q1 fiscal 2026 revenue of €8.8 billion, up 13.3%, and 2026 revenue guidance of €36 billion to €40 billion. Enbridge posted 2025 EPS of CAD $3.23, up 38%, and continues to support investors with 31 consecutive annual dividend increases and a roughly 5.15% yield. Overall, it is a constructive, stock-specific long-term thesis rather than a market-moving catalyst.

Analysis

The common thread is not “quality compounders,” it’s infrastructure toll-booths whose pricing power comes from ecosystem complexity. OKTA’s real edge is not identity security per se, but vendor-agnostic integration at a time when enterprise stacks are becoming more fragmented across cloud, SaaS, and AI agents; that makes it a beneficiary of multi-cloud sprawl and a subtle loser for single-vendor IAM efforts, especially MSFT. The second-order risk is that hyperscalers increasingly bundle identity into broader security suites, compressing standalone valuation multiples unless OKTA keeps expanding into machine identity and privileged access.

ASML is the highest-conviction secular bottleneck, but the market may still underappreciate how much of the upside is already pre-committed and how little of the current cycle is actually discretionary. That shifts the trade from “AI demand beta” to “capacity scarcity plus execution,” which should support the stock through a macro slowdown; the bigger risk is not demand collapse but export controls, customer concentration, or a digestion period if leading-edge capex pauses after the current order book is delivered. Competitively, the real knock-on winner is the semiconductor tool ecosystem: optics, metrology, and specialty materials suppliers should see persistent content growth even if chip designers’ margins compress.

ENB is being priced too narrowly as a yield vehicle when the better framing is regulated energy data-center infrastructure with inflation-linked cash flows. The AI power buildout creates a multi-year call option on gas transport and utility earnings that is less headline-sensitive than upstream energy, while renewables add optionality without driving the thesis. The hidden risk is political/regulatory: pipeline scarcity is a moat until permitting becomes weaponized, and dividend durability matters more than growth if rates stay higher for longer.

The contrarian takeaway is that the market is likely overpaying for pure AI beneficiaries while underpaying for the picks-and-shovels and plumbing that every AI buildout requires. OKTA is the most vulnerable to multiple compression if growth decelerates into the high-single digits, whereas ASML and ENB have better downside support because their economics are tied to structural bottlenecks, not just sentiment. Over 6-12 months, the cleaner expression is long bottleneck suppliers and service providers, short the most crowded AI adjacency names that lack hard capacity scarcity.

Key Decisions for Investors

• Long ASML on pullbacks over the next 1-3 months; use any 8-10% drawdown as entry for a 12-18 month hold, with downside cushioned by backlog visibility and upside from continued AI capex.
• Add OKTA only on weakness and size modestly; treat it as a 6-12 month re-rating trade, but hedge with a short MSFT basket leg if you want to isolate vendor-neutral IAM share gain versus suite bundling risk.
• Overweight ENB for income-plus-growth over 12-24 months; pair it against a lower-quality yield name if you want to express “durable cash flow + AI power demand” rather than simply chasing yield.
• Buy ASML call spreads instead of outright stock for the next catalyst window; the order book supports a controlled upside trade, but export-policy headlines can create abrupt gap risk.
• Avoid chasing NVDA/INTC on this thesis alone; the cleaner second-order expression is the tools and infrastructure layer, where scarcity is more defensible and valuation is less dependent on perfect end-demand execution.