The American Concert Venue Capacity Atlas

Ask how big an American concert venue is and you will get a number that sounds like a fact: 1,200, 5,000, 20,000. Almost none of those numbers travel with the information that makes them usable. A theatre's 2,800 might be seats with the orchestra pit covered; a club's 1,100 might be bodies on an open floor; an amphitheatre's 20,000 folds in a lawn that is not a seat at all; and the same building can legitimately publish all three shapes of number for different configurations. This atlas — the article companion to the interactive venue atlas and the venue inventory dataset — is built on the premise that a capacity without a basis is not yet a fact, and it quantifies only what the Live Index register actually holds.

The scope statement comes first because everything else depends on it. The register currently holds 94 U.S. venue records spanning 21 states and 30 markets, of which 90 carry a capacity published by the venue or its operator, stored with the configuration that figure describes; the rest carry no sourced capacity and are excluded from every count rather than estimated. This is a research sample assembled record by record, reviewed to 2026-08-20. It is not a census of American rooms, of which no complete public list exists, and every figure below is a statement about the sample.¹

Why capacity is configuration-dependent

A room's capacity is the output of a negotiation between architecture, production and regulation. Seated capacity counts chairs a configuration actually deploys; a full end-stage setup can delete hundreds of them behind sightlines. General-admission capacity counts standing bodies per square foot of floor and changes when a barricade or a mix position moves. Combined figures add the two. “Maximum published” figures are the largest number the venue is willing to put in public, usually without saying which configuration produces it. Licensed or fire capacity — the legal ceiling — is frequently different from all of the above and is rarely published at all. The register stores whichever figure the venue publishes together with its basis, and where a venue publishes a range, the range survives instead of being collapsed to a midpoint. The measured tables below show the basis mix across the cohort, and it is not a footnote: figures that count different things are the single largest obstacle to any national capacity arithmetic.³

The bands Live Index uses

The atlas groups rooms into six bands: under 250, 250–499, 500–999, 1,000–2,499, 2,500–4,999, and 5,000 and above. The boundaries sit where routing decisions actually change rather than at round numbers that flatter arenas. Under 250 is the open-mic and first-headline tier where an act discovers whether it draws at all. The 250–499 club is the first room with real production and a real guarantee. Crossing 500, and again 1,000, changes the economics of staffing, security and sound wholesale — the tier examined in the venue ladder research. The 2,500–4,999 theatre tier is where the touring middle either finds rooms or stalls, and 5,000 and above is the amphitheatre, arena and stadium territory where capacity itself becomes a marketing number. A band is assigned only where a capacity is sourced; unknown is not zero.

What the held sample shows — and what it cannot

The histogram below counts the cohort's records per band, and its most important property is the one that disqualifies it as a market measure: the sample skews large. Arenas, amphitheatres and publicly financed buildings are heavily documented — ownership records, bond prospectuses, management agreements — while the small independent rooms that carry the bottom of the ladder are documented worst, precisely because nothing about their arrangements is filed anywhere. The register's under-250 and 250–499 rows are therefore thin in a way that says more about public documentation than about American nightlife. The regional coverage matrix makes the same point geographically: its low and zero cells mark where the register has not yet documented rooms, not where rooms do not exist. Treating either table as a national census would be exactly the error — an inventory number without an inventory — that the companion piece on venue capacity as an inventory problem diagnoses across the industry's own claims.

What the sample can support is structural reading. It shows the shape of documented supply across the four Census regions.² It shows how often published figures rest on maximum or variable-configuration bases rather than on a countable setup. And joined to the register's relationship layer — who owns, operates and tickets the same rooms, mapped in the U.S. venue graph — it shows that the buildings whose capacity is best documented are also the buildings whose control is most concentrated, which is not a coincidence: both facts are downstream of the same public paper trail.

What a real national census would need

A defensible national capacity census would need three things nobody currently publishes: a register of rooms that present live music, including the ones inside restaurants, breweries and churches; a per-room capacity with a stated basis and date, resurveyed as configurations change; and a way of counting rooms that close and open below the threshold of trade coverage. Federal statistics measure the industry's employment and prices, not its rooms; trade databases cover what promoters report; and municipal licensing data uses categories built for fire codes rather than for music. Until that changes, the honest form of the national question is a maintained sample with visible edges — which is what this atlas is, and why its counts are computed from the register at build time rather than frozen into prose that would quietly go stale.

How to use the atlas

• Never sum or average capacities across records without checking the basis column; a seated figure plus a GA figure is not a number.
• Read the bands as routing tiers, not as prestige tiers; the economics change at the boundaries.
• Treat every count as a statement about the Live Index cohort as of its review date, never as a national figure.
• Read zero cells in the regional matrix as register to-do items.
• Follow each record to its source; every capacity in the register names the document it was read from, and corrections with a source are accepted and dated.