Regulators decide a company is too powerful in two steps. First they define the market it competes in, then they measure how concentrated that market is. The first step usually determines the outcome, which is why most serious antitrust fights are arguments about market definition rather than about behavior.

The question matters more than it used to. A handful of firms now control infrastructure that everything else runs on, and the rules being applied to them were written for railroads, oil and steel.

Step One Is Defining the Market, and It Decides Everything

US antitrust agencies define a relevant market as an area of effective competition, made up of a product element and a geographic element. The boundary is drawn by asking what customers would switch to, using the reasonable interchangeability of the products and the cross-elasticity of demand between them.

The stakes in that definition are enormous. A company with 80 percent of premium electric sports cars has a trivial share of cars. Same firm, same sales, completely different case. Defense lawyers argue for the widest plausible market and agencies argue for the narrowest, because share follows definition.

The Hypothetical Monopolist Test

To stop that argument collapsing into assertion, agencies use a specific test.

Imagine one firm became the only present and future seller of a candidate group of products. Would it profitably impose a small but significant and non-transitory increase in price, known as a SSNIP? If yes, the group is a real market, because customers have nowhere better to go. If customers would escape to substitutes, the group is drawn too narrowly and the boundary widens until the test passes.

The threshold is usually 5 percent of the price charged by the firms involved, though agencies may use a larger or smaller figure.

Price is not the only dimension. The same test applies to any worsening of terms, including quality, service, capacity investment, product variety and innovative effort. That extension matters for services that are free at the point of use, where harm shows up as degraded quality rather than a higher bill.

Step Two Is Measuring Concentration

Once the market is drawn, concentration gets a number: the Herfindahl-Hirschman Index, calculated by squaring each firm’s market share and adding the results, so that a few large firms score far higher than many small ones.

Under the 2023 US Merger Guidelines, a market with an HHI above 1,800 is highly concentrated, and a change of more than 100 points is a significant increase. A merger that does both raises a presumption of illegality. A deal that creates a firm with a share over 30 percent while raising the HHI by more than 100 points is likewise presumed to lessen competition substantially.

Presumed is the operative word. These are starting points that shift the burden of explanation onto the merging parties, not automatic prohibitions.

Being Big Is Not Illegal, What You Do With It Can Be

This is the most widely misunderstood part of the field. Neither US nor EU law prohibits being dominant.

US law makes it illegal to monopolize or attempt to monopolize, which requires conduct, not merely size. EU law prohibits a dominant firm from abusing its position, through unfair pricing or limiting production for instance, while leaving the position itself lawful.

Courts also sort conduct by how clearly harmful it is. Price fixing and bid rigging are treated as illegal in themselves, with no inquiry into their actual effects. Most other conduct is judged under the rule of reason, weighing intent and consequences together, which is slower and far less predictable.

Notably, the 2023 guidelines set no automatic share threshold for dominance. It is assessed through direct evidence or through shares that demonstrate durable market power, with weight given to how long the power has persisted and how high the barriers to entry are. Durability is the test, not a number.

Why Technology Cases Are Harder

Every step above was designed around products with prices, rivals and stable boundaries.

A SSNIP test struggles when the price is zero. Market definition struggles when a firm supplies infrastructure that its own competitors depend on. And concentration measured today says little about durability when the relevant question is whether control of one layer forecloses the next.

That is the live argument in current enforcement, including deals structured as licensing rather than acquisition specifically to avoid triggering merger review. Whether the existing framework can reach those structures is not settled.