Atlante del Capitalismo Italiano (Atlas of Italian Capitalism)

Who owns the companies behind Italy’s largest private fortunes? The Atlante del Capitalismo Italiano (Atlas of Italian Capitalism) is an interactive attempt to make those chains visible: people and families connect to holdings, foundations and operating companies through explicit ownership links.

The current release is a stock analysis dated 31 December 2025. It contains nodes, relationships and source records. Its billion-euro coverage includes all 71 qualifying entries in the December 2025 Forbes Italy wealth ranking and all 83 Italian listed companies whose end-of-year market capitalization met the same threshold under the stated dataset definition.

Motivation

Rankings of wealthy people are easy to find, but they usually compress a fortune into a single number. Corporate reports contain the opposite problem: they describe individual shareholders and subsidiaries in detail, while the broader chain remains scattered across many documents and legal entities.

The atlas tries to connect these two views. Its purpose is to show how and why a person or family is wealthy: which holdings concentrate the ownership, which companies sit downstream, what percentage is held at each step, and which public source supports the link. Turning that structure into a graph makes complex chains easier to explore, question and correct collaboratively.

Explore the atlas

Use the filters to isolate a family or group, select a node to inspect its sources, and zoom or drag to navigate. If the embedded view is too small, open the full atlas ↗.

How to read the graph

Node area is proportional to an indicative value. Size is used to communicate order of magnitude and visual importance, not to produce a single accounting total for the network.

The graph uses four main visual categories:

• gold and coral nodes represent people and families;
• violet nodes represent holdings, trusts and similar vehicles;
• turquoise nodes represent operating companies;
• blue nodes represent the state and public administrations.

An arrow means owns and carries the disclosed percentage. Dashed lines describe family relationships. The geographical arrangement is deliberately soft: Turin, Milan, Rome and other cities orient the layout without claiming that ownership itself is territorial.

How values are chosen

The atlas needs a value measure so that a large fortune or company is immediately distinguishable from a smaller one. No single measure works for every kind of node, so the database records both the value and the basis used to obtain it:

• People and families: a published wealth estimate from a recognizable source.
• Listed companies: generally market capitalization near the snapshot date. Market capitalization (share price multiplied by shares outstanding) is a public, reproducible estimate of the market value of the company’s equity.
• Holdings and investment vehicles: net asset value when a credible estimate is available. NAV considers the value of the assets held minus liabilities and is often more meaningful for an entity whose main function is to own investments rather than sell goods or services.
• Private companies: a source-based indicative equity value when available; otherwise the value is left empty rather than manufactured from insufficient information.

Market capitalization and NAV are conceptually closer to each other than revenue because both attempt to describe value attributable to equity. They are nevertheless not equivalent: market capitalization is a live market judgement, while NAV depends on the valuation of underlying assets and may be measured less frequently. A listed holding can also trade above or below its NAV.

Revenue is deliberately not used for node size. It measures the volume of business conducted during a period, not what the business is worth. Two companies with the same revenue can have very different margins, debt, assets and growth prospects; a pure holding may own valuable stakes while reporting little conventional turnover.

These values must therefore be read as visual approximations, not interchangeable financial measurements. The value of an owner is not automatically calculated by multiplying every downstream company by the ownership percentage, and values along a chain must not be added together: the holding and the company appear as separate nodes precisely to make the ownership structure visible, even when their economic value overlaps.

Sources, limits and collaboration

Every node and every relationship carries a source. Primary materials such as annual reports, governance documents, official shareholder disclosures and investor-relations publications are preferred wherever available.

The EUR 1 billion boundary is reproducible rather than editorial: dollar wealth estimates are converted with the ECB reference rate closest to the Forbes observation date, while listed companies use end-of-2025 market capitalization. Related ranked people can be collapsed into one family node when they share the same ownership pool, avoiding visual double counting. Private-company coverage is relationship-driven rather than claimed as exhaustive because there is no complete, observable public universe of private valuations. The full aggregation map and the deliberate no-link cases are documented in the 2025 coverage register ↗.

The initial research was produced with the assistance of OpenAI Codex and checked on a best-effort basis. The atlas may contain errors, omissions or outdated interpretations and should not be used as financial, legal or investment advice. Company logos are shown only for identification and remain the property of their respective owners.

The database is structured as annual snapshots so that future contributors can propose corrections, add ownership chains and build comparable editions for other years.