What is a moat?

When Jensen Huang dimmed the lights at Nvidia’s annual GTC conference this spring, the audience expected silicon. Instead they got software: one slide, eight minutes, and 900 brightly colored rectangles, each a library inside CUDA. “A long time ago,” Huang said, “this slide was all we had.” In 2004 it fit on a Post‑it; today it glows like a stained‑glass window of modern computing. For investors, that slide was a live X‑ray of a moat — an ever‑deepening ditch that keeps competitors on the far bank. Warren Buffett popularised the metaphor in the 1980s, but the idea predates medieval battlements: a company earns the right to durable profits only if something makes attacking it ruinously expensive.

The five engines that deepen the ditch

Intangible assets: the invisible wall
Brands, patents and regulatory licences can harden into stone. LVMH’s trunks and flutes may carry less gloss in 2025 — Moët Hennessy burned €1.5 billion of cash last year — yet the Louis Vuitton monogram still commands margins a generic bag can only dream of. In pharma, Novo Nordisk’s semaglutide patents don’t sparkle on Instagram, but they lock rivals out of a GLP‑1 market the Danish firm still dominates with 52 % share — protections that run to at least 2032.

Switching costs: ‘too sticky to quit’
In enterprise software the real invoice arrives the day you try to leave. Ask a Fortune 500 controller what it would cost to migrate decades of ERP data from SAP, or an iPhone user confronted with the green‑bubble purgatory of group chats.

Positive Network effects: the self‑reinforcing current
CUDA is a textbook case. The more researchers write to Nvidia’s language, the more indispensable Nvidia chips become, and vice‑versa. The result: gross margins that hover near 80 % and a graphics dynasty that once held 90 % of the AI‑GPU market. Even critics who warn the moat is “only 18 months deep” concede that momentum compounds faster than challengers can rewrite code.

Cost advantage: fighting on the low ground
Walmart’s scale lets it negotiate prices competitors can’t match; BYD’s integration — batteries, chips, even the ships that haul its cars — has punched a 15‑point cost gap against foreign EV brands in China. Here, the moat is measured not in patents but pennies: the widest ditch is often the thinnest margin.

Efficient scale: natural monopolies
Some industries are natural cul‑de‑sacs. A single water utility in a mid‑sized U.S. county can earn steady returns precisely because it would be madness to lay a second set of pipes. These moats are quiet, regulation‑heavy, and — in the age of climate stress — newly fashionable with pension funds hunting bond‑like yields.

Proof in the Performance: Why Wide Moats Pay

Multiple quantitative studies have tested whether companies designated “wide‑moat” or “high quality” go on to outperform the broader market:

• Morningstar Wide Moat Focus Index – Since its inception in February 2002 through May 2025 the index has compounded at 11.8 % annually versus 8.4 % for the S&P 500, delivering roughly 350 basis points of excess return with slightly lower standard deviation (15.6 % vs 17.2 %). The outperformance survives sector‑neutral tests and persists after transaction‑cost assumptions. Morningstar Indexes 2025 Fact Sheet
• “Quality Minus Junk” (Asness, Frazzini & Pedersen, 2019) – Although not branded as “moat” research, the study shows that firms with persistent profitability, low leverage, and stable earnings — attributes that overlap strongly with moat ratings — earned a global premium of ~4 % per year over five decades.

The common thread: wide‑moat companies are not just defensive; they recycle excess returns back into the moat, compounding advantage — and shareholder value — in tandem.

How moats erode

Across the last century, the identity of the world’s largest company has been a relay race, not a coronation. In 1901, U.S. Steel became the first billion‑dollar corporation; four decades later it had ceded the throne to General Motors. By 1967, IBM’s market cap eclipsed both. The oil shocks of the 1970s crowned Exxon; Japan’s asset bubble briefly pushed the Industrial Bank of Japan to the top in 1989; the dot‑com mania anointed Microsoft in 1999. Each reign looked impregnable — until a new technology, a new policy regime, or a fresh geography redrew the map. So moats DO erode over time.

The lesson is statistical as well as narrative: between 1960 and 2024, the median tenure at the peak of the global market‑cap rankings was barely six years — an observation echoed by multiple studies that track corporate longevity. Innosight’s 2021 Corporate Longevity Forecast shows the average tenure of companies on the S&P 500 has collapsed from 30–35 years in the late 1970s to just 15–20 years this decade, while a 2024 McKinsey analysis pegs the average life span at 21 years. Academic work by Wiggins and Ruefli, which followed 6,772 firms across 40 industries for 25 years, concludes that persistent superior economic performance beyond a single decade is “very rare.”

Even the Magnificent Seven are living on half‑life: regulators, AI commoditisation, and shifting consumer sentiment all lurk at their moat’s edge. Edge cases such as Berkshire Hathaway — which has clung to the top‑ten list for decades by owning a rotating portfolio beneath the surface — only reinforce the point: adaptability, not any single moat, is the ultimate defence.

And back at GTC, Huang ended his slide‑show sermon with a confession: “We are one open‑source repo away from irrelevance.” In that single sentence lies the paradox of the modern moat: the same forces that deepen it — community, data, scale — can just as quickly reroute the water.

Conclusion: Keep the Moat in Motion

The evidence is unmistakable: portfolios tilted toward companies with demonstrably wide moats have, across market cycles, delivered superior returns with a smoother ride. Yet the very data that confirm the payoff also warn against complacency. Longevity research shows that almost every moat—no matter how wide—erodes over the very long run. Patents expire, tastes pivot, regulators redraw the playing field, and technological leaps tunnel beneath yesterday’s walls.

For investors, wide‑moat ownership therefore demands vigilant stewardship rather than blind faith. Two disciplines matter most:

1. Re‑test the moat. At least annually, re‑examine the five classic defences—intangible assets, switching costs, network effects, cost advantage, and efficient scale. If you can’t explain in a sentence why the moat still holds, assume it doesn’t.
2. Mind the price. A moat justifies a premium, but not an unlimited one. Overpaying compresses future returns faster than any rival ever could.

Own the castle, patrol the perimeter, and be willing to redeploy capital when the drawbridge rots. Done well, wide‑moat investing becomes an adaptive process: Accumulate enduring franchises when they’re mis‑priced, replace those whose ditches have silted up, and let disciplined compounding raise the walls anew. In a landscape where advantage flows like water, the edge belongs to investors who keep moving with the current.