Ferrari is the first company in this dataset where nobody has a majority of anything, and it is more instructive for that.
As of February 4, 2026, Exor N.V. held about 21.33% of Ferrari's outstanding common shares and about 32.32% of the votes. Trust Piero Ferrari held about 10.67% of the common shares and about 16.17% of the votes, with Piero Ferrari holding usufruct including the voting rights. Exor's wedge is roughly 11 points; Piero Ferrari's is roughly 5.
Add the two together and you get about 48.49% of the votes. Just under half, which is the sort of number that decides everything in practice and nothing on paper.
Loyalty voting is not dual-class, and the difference matters
Ferrari's extra votes come from special voting shares, issued one-for-one to common shareholders who register their shares and hold them continuously through a qualifying period. Each special voting share carries one vote, which doubles the voting weight of the common share it attaches to.
The mechanism is open to every shareholder. An index fund could register tomorrow and, after the qualifying period, vote twice per share exactly as Exor does. Nothing about the structure reserves the privilege to a founder or a family.
That is why this profile records the mechanism as loyalty voting rather than dual-class, even though the arithmetic looks similar. A ten-vote Class B is a privilege granted at incorporation to specific people and usually extinguishable only by them. Loyalty voting is a standing offer that most shareholders decline. The concentration is a consequence of who bothers, not of who was allowed.
It also inverts the usual durability question. A founder's Class B erodes when the founder sells. Loyalty voting erodes when anyone else starts participating, because every newly registered long-term holder dilutes the voting weight of those already registered.
Working control, and why the table says so
The decision-rights row below is marked as needing another party's agreement, which will read oddly to anyone who follows Ferrari and knows how its general meetings actually go.
The marking is deliberate. Exor cannot elect the board alone on 32.32% of the votes. In practice it does not need to, because turnout at a general meeting is never 100% and a third of the register is decisive against a dispersed float. That is working control: reliable, unremarkable, and legally contingent on other people not showing up.
The distinction is not academic. Working control is the kind that a concentrated activist position, an unusually high turnout, or a coordinated institutional vote can actually threaten. Majority control is not. Recording both as "controls the company" would erase the only difference that matters when something goes wrong.
What is not established here
Whether Exor and Trust Piero Ferrari act in concert. Their combined 48.49% is arithmetic performed here, not a disclosed bloc, and no filing cited above describes an agreement between them. If such an arrangement exists, the control picture changes materially and this profile is wrong to treat them separately.
Also the qualifying period's exact terms and the register mechanics, which determine how quickly the position could be diluted by others registering.