AI Makes Recorded Music Cheap. What Happens to the Concert?

Recorded music has spent more than a century becoming cheaper to reproduce. Mechanical recording separated a performance from the room in which it occurred; radio and physical media multiplied distribution; digital files reduced reproduction cost toward zero; streaming converted ownership into effectively continuous access. Generative artificial intelligence extends that trajectory in a different direction. The scarce object is no longer only the recording. The process of producing a plausible new recording can itself become abundant.

The rate of change is measurable. Deezer reported in July 2026 that fully AI-generated tracks had exceeded 50 percent of new daily music deliveries at peak periods in June, with a monthly average of roughly 90,000 AI-generated uploads per day.¹ Three months earlier the company had reported approximately 75,000 such tracks daily, representing about 44 percent of deliveries.² Deezer's figures describe one streaming platform and its own detection methodology, so they should not be generalized to the entire music market. They nevertheless demonstrate that synthetic supply can grow at a scale that would have been impossible under human-only production.

Consumption tells a different story. Deezer reported in April 2026 that fully AI-generated tracks represented only 1 to 3 percent of streams on its service and that a large share of those streams were classified as fraudulent and demonetized.² The gap between supply and genuine demand is analytically important. AI has already made production abundant; it has not yet demonstrated equivalent audience preference.

Recorded music as an industry remains commercially strong. IFPI reported that global recorded-music revenue reached $31.7 billion in 2025, an increase of 6.4 percent and the eleventh consecutive year of growth.³ The rise of generative audio therefore should not be framed as evidence that recorded music is economically collapsing. The more useful question concerns scarcity. If millions of additional tracks can be produced at negligible marginal cost, what types of musical value become relatively harder to reproduce?

Human presence is one candidate. A live performance occupies a specific time and place, requires bodies to travel and cannot be consumed simultaneously by an infinite audience without changing its form. A recording of a concert can be copied; the shared event itself cannot. The crowd, weather, technical failure, improvisation and knowledge that a particular human being is physically performing create forms of scarcity that generative media do not eliminate.

This does not mean AI will make all human concerts more valuable. Supply of live events can exceed demand, artists can be interchangeable, and consumers may prefer synthetic music in some contexts. The claim is relative: when one category of cultural output becomes radically more abundant, scarce categories can acquire greater differentiation. Economic value often migrates toward what remains difficult to reproduce.

Authorship law reinforces the distinction conceptually. The U.S. Copyright Office's 2025 report on generative AI concluded that copyright protection for outputs depends on sufficient human authorship and that mere prompting does not itself establish the necessary expressive control.⁴ The legal rule is not an economic measure, but it reflects a social distinction between automated output and human creative contribution that may also influence audience preferences.

The live sector should be cautious about interpreting this as permission to premiumize human presence indefinitely. If authentic performance becomes more culturally scarce while ticket prices simultaneously move beyond ordinary audiences, the industry can convert increased relative value into reduced participation. Pollstar's long-run Top 100 ticket data already show substantial price growth since 2019, even with moderation in 2025 and 2026.⁵ A future in which synthetic recorded culture is abundant and human live culture is scarce can strengthen the case for preserving broad access rather than weaken it.

AI may also reduce costs inside live music. Generative systems can assist with marketing assets, translation, customer service, scheduling, research, production planning and analysis. The relevant distinction is between using AI to lower the cost of supporting a human event and using AI to replace the human event itself. Live Index's own research policy follows the same logic: automated systems can expand analytical capacity, but material factual claims and conclusions remain subject to human scrutiny.

There is also a provenance opportunity. As synthetic media increase, verified human performance may become a stronger signal. Ticketing identity, artist certification, live recordings, provenance metadata and direct fan relationships could help audiences distinguish human-origin work from industrial-scale synthetic supply. The economic value of "live" may therefore expand beyond the event into a broader authenticity layer around artists.

The central uncertainty is demand. Deezer's current data show explosive synthetic supply but limited genuine consumption.¹² If listeners ultimately become indifferent to origin, the scarcity argument weakens. If provenance and human identity remain important, live performance may become one of the clearest places where authenticity can be observed directly. This is a question for longitudinal data rather than aesthetic prediction.

Live Index will track AI music through supply, consumption, fraud, labeling practices and the relationship between synthetic proliferation and live-event demand. The hypothesis is testable: as synthetic recorded supply grows, do verified human artists experience stronger live conversion, greater willingness to travel, or higher repeat attendance? The answer should be measured before it becomes a slogan.