LVMH

LVMH is described as a portfolio and behaves like a pyramid. Seventy-five maisons sit inside five reported divisions, and the group's own accounts show how uneven the weight is: Fashion & Leather Goods produced 47% of 2025 revenue and 74% of profit from recurring operations. Inside that division two houses, Louis Vuitton and Dior, do most of the work. Louis Vuitton alone is generally estimated at over €20 billion of revenue, though LVMH has never published the figure and I am not going to pretend the sell-side consensus is a disclosure. Everything else — champagne, cognac, Sephora, Tiffany, Bulgari, Tag Heuer, the hotels — is real, large, and secondary.

The model at maison level is the one the Dior case describes: a brand owns its design, its stores and its price, subcontracts more of the making than it admits, and spends roughly two-fifths of every euro on selling. What the group adds is three things. It buys the real estate for every maison at once, which is why Louis Vuitton and Tiffany end up on the same corner in every city. It moves managers between houses, so the Dior chief executive came from Louis Vuitton and the Tiffany one did too. And it borrows against the whole to buy the next name, which is the actual business Bernard Arnault has been in since 1984.

The cost of the model is that the group is only as strong as its two engines. When Fashion & Leather Goods grew at double digits, from 2021 to 2023, the portfolio looked like genius. When it fell — margin went from 39.9% in 2023 to 35.0% in 2025, on a Chinese customer who stopped buying — nothing else in the group was large enough to hide it.

The moat is the names, and the names are not for sale. Louis Vuitton has been making trunks since 1854, Dior showed the New Look in 1947, Tiffany registered its blue in 1837. Heritage of that kind cannot be founded, only bought, and LVMH has bought more of it than anyone. The Hermès case on this site shows the alternative — one family, one name, one category — and the Hermès premium over LVMH tells you what the market thinks of the two approaches. But no other conglomerate has assembled anything like this collection, and Kering's difficulties with Gucci since 2023 show what a portfolio looks like when it has only one engine instead of two.

Scale is the second moat and it is more concrete than brand. LVMH is the largest tenant on every luxury street in the world, the largest advertiser in every fashion magazine, and the largest buyer of crocodile skins, tanneries and Italian workshops. When it wanted Tiffany it paid $15.8 billion in cash and borrowed the rest without strain. A single-brand competitor cannot match the rents LVMH pays for corners, and the corner is the marketing.

The honest caveat is that a moat made of names is only as wide as the names are desirable, and desirability has a cycle. From 2024 the aspirational customer — the person buying one €2,000 bag a year, not ten — left the category, in China first and then everywhere. LVMH's answer was to raise prices further, which protected margin for two years and then did not. The moat is intact. The customer it was dug for has gone quiet.

LVMH's capital allocation is Bernard Arnault's, and it has one consistent shape over forty years: borrow, buy the name, pay down, repeat. Boussac in 1984 for Dior. The Louis Vuitton–Moët Hennessy merger in 1987, and the stake-building in 1988 and 1989 that turned an invited investor into the controlling one. Celine, Loewe, Fendi, Sephora and Tag Heuer between 1996 and 2001. Bulgari in 2011 for €3.7 billion, Loro Piana in 2013, Christian Dior Couture in 2017 for €6.5 billion, Tiffany in 2021 for $15.8 billion. The failures are as instructive as the wins: Gucci in 1999, lost to François Pinault and ultimately sold back at a profit, and the Hermès stake of 2010, built through equity swaps the French regulator fined him €8 million for not disclosing, and unwound in 2014 with the family still in control.

The record on organic investment is nearly as good — the group has funded flagship rebuilds, workshops and Sephora's expansion through every downturn — and the record on returning cash is deliberately modest. The dividend has grown most years and buybacks are small and opportunistic, because Arnault would rather hold the powder for the next name. The one allocation I cannot score is the one that matters most now: how the group is divided among Delphine at Dior, Antoine at Christian Dior SE, Alexandre at Moët Hennessy, Frédéric at Loro Piana and Jean at Louis Vuitton watches. The board raised the chief executive's age limit to 80 in 2022 and to 85 in 2025. That is a decision about capital too.