S11.1 · Media, Entertainment & Creative
Scripted content spend of ~$250B splits into high-margin finished libraries and low-margin physical production, with AI compressing the latter's crew-labor cost.
Global scripted content spend runs at roughly $250B annually (Ampere Analysis), off a 2022 "Peak TV" high near $260B after streamers cut order volume. The label covers two businesses that deserve separate underwriting. Finished content libraries are perpetual-licensing annuities — high margin, durable, largely indifferent to what happens on set. Physical production is thin-margin, working-capital-heavy and tax-credit-dependent. Generative AI is compressing the cost base of the second while leaving the first largely untouched, and every claim about the segment should be tested against which of the two it actually describes.
Where a show is commissioned and where it is shot are no longer the same question. Greenlighting concentrates in the US and UK, where Netflix, Disney, Warner Bros. Discovery, Paramount, NBCUniversal and Amazon control the great majority of scripted budgets, with thousands of independent production companies beneath them holding no comparable pricing power. The physical work itself migrates to Canada, Central and Eastern Europe, and other tax-incentive jurisdictions — a labor and subsidy arbitrage, not a demand signal. The chain runs from talent and IP ownership upstream, through production companies, to distributors and platforms downstream, and position determines economics: a finished library throws off syndication and perpetual-licensing revenue at high margin, while the production layer that creates new titles carries thin, working-capital-heavy, incentive-dependent margins and no annuity value at all.
Spend was flat to declining through 2023-2025 as streamers tightened order books after the Peak TV years. The recovery case rests less on demand than on AI-driven reductions in per-title cost — cheaper titles change the greenlight math at the margin, but they do not expand total addressable spend. That is more volume inside a flat budget, not a bigger market.
The cost compression is landing in three specific places: previsualization, dubbing and localization, and visual-effects labor. What is being eroded is scarcity-priced below-the-line crew labor — the specialized technical work that has historically set the floor on production budgets. Dubbing and de-aging applications are already in production use; AI-generated background and visual-effects work at scale sits on a 2-5 year horizon. The savings accrue first to whoever holds the greenlight, not to the crews and vendors being displaced.
Alongside traditional production, fully synthetic "AI content studios" are forming — video generated without a conventional physical production process at all. They do not yet threaten licensed-library economics; distribution and IP relationships still gate access to audiences. But they compete for the same greenlight budgets, with a cost structure that has no analog in crew-based production.