Industry View · Media & Entertainment
Generative AI has moved from demo reel to production pipeline. Netflix put AI-made VFX on screen for the first time in 2025, Disney agreed to take a $1 billion equity stake in OpenAI and license 200-plus characters to Sora, and AI-native tooling startups now carry $4-11 billion valuations — even as the music labels and SAG-AFTRA fight over who gets paid when the machine does the work.
Falling inference cost and faster VFX, but return mostly qualitative
AI handles script breakdown, shot previsualization and concept art; studios report meaningful concept-art cost reductions.
LED volumes plus AI logistics shorten location shoots; vision systems track actors and manage set data in real time.
Generative VFX, AI rotoscoping, de-aging and rendering compress post cycles from months to weeks; Netflix cut some VFX render costs via on-demand GPU scaling.
AI dubbing and voice cloning re-monetize archives in Hindi, Portuguese and Arabic at a fraction of prior re-dubbing cost; fastest-growing agentic segment.
GenAI powers conversational search, localized promo assets and real-time ad creative; Netflix is rebuilding recommendations and ad formats around it.
01 · The thesis
The industry's center of gravity is shifting from who can fund content to who can generate it cheaply and distribute it at scale. The clearest tell of 2025 was Netflix using generative AI to render a building collapse in The Eternaut substantially faster than conventional VFX — a sequence its co-CEO Ted Sarandos said simply 'wouldn't have been feasible' on that budget. Cost curves like that compound: a sharp fall in inference costs between 2024 and 2025 put studio-grade tools in the hands of teams that could never afford a VFX house. The strategic question is therefore vertical integration. Netflix's agreement to acquire Warner Bros. — an ~$72B equity-value deal announced in December 2025 — is read by analysts partly as a bid for premium catalog and training data; Disney's OpenAI deal sought to convert IP-lawsuit risk into an equity position. The winners will be platforms that own both a vast catalog (the fuel) and global distribution (the moat). The losers are the middle layers — boilerplate VFX, localization, stock music, background performance — where the marginal cost of a 'good enough' synthetic asset is collapsing toward zero.
AI handles script breakdown, shot previsualization and concept art; studios report meaningful concept-art cost reductions.
LED volumes plus AI logistics shorten location shoots; vision systems track actors and manage set data in real time.
Generative VFX, AI rotoscoping, de-aging and rendering compress post cycles from months to weeks; Netflix cut some VFX render costs via on-demand GPU scaling.
AI dubbing and voice cloning re-monetize archives in Hindi, Portuguese and Arabic at a fraction of prior re-dubbing cost; fastest-growing agentic segment.
GenAI powers conversational search, localized promo assets and real-time ad creative; Netflix is rebuilding recommendations and ad formats around it.
02 · The two clocks
The cost clock. Inference costs for generative AI fell sharply between 2024 and 2025, and hyperscalers cut cloud-GPU prices materially over the same period. Every drop pulls studio-grade VFX, dubbing and music generation further into reach of teams with no studio budget — eroding the cost advantage that scale used to confer. The litigation clock. Scores of lawsuits worldwide challenge unlicensed AI training. The music majors are converting that risk into revenue — UMG-Udio, WMG-Suno and WMG-Udio — but Sony is still litigating, and the musicians' union (AFM) is suing the labels themselves, arguing performers see none of the settlement money. The legal architecture for who-owns-what is being written in real time. The trust clock. With AI-generated content set to flood feeds in 2026 (Deloitte), the scarce asset becomes verified human quality. Disney's deal pointedly excluded talent likenesses and voices, and OpenAI abruptly moved to wind down Sora in early 2026 — a reminder that platform strategy here can reverse on a quarter's notice.
03 · Public players & exposure
We plot the listed players on two editorial axes — how exposed each is to AI disruption, against how ready its data, brand and position are to be the answer. The figures in the table are sourced; the placement is our read.
04 · Private flagships
The companies attacking this industry AI-first, with disclosed funding where available:
NY-based frontier lab selling text-to-video on subscription and per-seat enterprise tiers; partners span film studios to fintech.
London-based voice synthesis and dubbing leader; launched Eleven Music built with labels and cleared for commercial use, and is eyeing an IPO.
Enterprise text-to-video with lifelike multilingual avatars for training, comms and marketing; positioned around consent and misuse controls.
Consumer AI music generator at the center of the label lawsuits; settled with Warner and is rebuilding on licensed models for 2026.
Builds world models for video and 3D capture; one of the larger 2025 video-AI raises as funding to the category surged year-on-year.
Creator of 'AI actress' Tilly Norwood, pitched to talent agents as a 'hyperreal digital star' — and the lightning rod for the performer-replacement debate.
05 · Signals
06 · The exposure read
AI rewards clean, structured advantage and punishes friction. The line runs through who owns the data, the brand and the customer — and who is merely a step the technology can route around.
Sources
AI-native M&E tooling: disclosed valuations, 2025-26
Latest disclosed post-money valuations. Sources: ElevenLabs/CNBC, Runway/TechCrunch, Synthesia/SiliconANGLE, Luma/CNBC, Suno/Forbes. ($B valuation)
Netflix puts GenAI VFX on screen
The Eternaut becomes the first Netflix original with final generative-AI footage, completed substantially faster than traditional VFX (BBC).
Labels pivot from suing to licensing
UMG settles with Udio; WMG settles with both Udio and Suno; Sony remains among the majors still litigating (Chartlex, THR).
Disney agrees $1B OpenAI deal
Disney agrees to license 200+ characters to Sora and take equity, while excluding talent likenesses and voices (CNBC, Disney).
ElevenLabs hits $11B, eyes IPO
A $500M round more than triples its valuation in a year on $330M+ ARR, signaling AI audio's breakout (CNBC, ElevenLabs).
OpenAI moves to wind down Sora
Within months of the Disney agreement, OpenAI signals it will refocus away from Sora — underscoring how fast platform strategy can flip (reported).
NY-based frontier lab selling text-to-video on subscription and per-seat enterprise tiers; partners span film studios to fintech.
London-based voice synthesis and dubbing leader; launched Eleven Music built with labels and cleared for commercial use, and is eyeing an IPO.
Enterprise text-to-video with lifelike multilingual avatars for training, comms and marketing; positioned around consent and misuse controls.
Consumer AI music generator at the center of the label lawsuits; settled with Warner and is rebuilding on licensed models for 2026.
Builds world models for video and 3D capture; one of the larger 2025 video-AI raises as funding to the category surged year-on-year.
Creator of 'AI actress' Tilly Norwood, pitched to talent agents as a 'hyperreal digital star' — and the lightning rod for the performer-replacement debate.
| Company | Stance | The sourced fact |
|---|---|---|
| Netflix NFLX | AI-native incumbent | Used generative AI for final on-screen VFX in The Eternaut, completed substantially faster than traditional methods; Q3'25 revenue grew about 17% to roughly $11.5B (Q3'25 shareholder letter, BBC). |
| The Walt Disney Company DIS | Hedging via equity | Agreed to a $1B equity stake in OpenAI and to license 200+ characters to Sora in a multi-year deal; the arrangement was later thrown into doubt when OpenAI moved to wind down Sora in early 2026 (Disney, CNBC, Dec 2025). |
| Warner Bros. Discovery WBD | Being acquired | Its studio and streaming business is being acquired by Netflix under a December 2025 agreement (~$72B equity value); its film library is valued by analysts partly as generative-video training data (Netflix, Hollywood Reporter, Dec 2025). |
| Universal Music Group UMG | Litigate-then-license | Settled with Udio in Oct 2025 to co-build a licensed 'walled-garden' AI music platform targeted for 2026, while continuing to litigate Suno (Chartlex, MBW). |
| Warner Music Group WMG | First label to settle Suno | First major label to settle with Suno (Nov 2025) in a first-of-its-kind partnership, also acquiring Songkick; settled with Udio the same month; new licensed models due 2026 (Chartlex, MBW). |
| Sony Group (Music) SONY | Lone holdout | Among the majors still litigating both Suno and Udio rather than settling, as the legal architecture for AI-music rights is written in real time (Chartlex, THR). |
| NVIDIA NVDA | Picks and shovels | Omniverse is used to accelerate VFX rendering across major studios and ILM-class pipelines; NVIDIA also participated as an investor in Runway's Series E (Crunchbase, TechCrunch). |
| Adobe ADBE | Creative-tool pivot | Firefly generative AI has produced tens of billions of assets cumulatively since 2023 and is embedded across inpainting and compositing workflows (Adobe). |
| Runway PVT | Frontier video model | Raised a $315M Series E at a ~$5.3B valuation (Nvidia, Adobe, Fidelity among investors) (Crunchbase, TechCrunch, 2026). |
| ElevenLabs PVT | Voice/audio leader | Raised $500M at an $11B valuation in Feb 2026 (Sequoia-led), closing 2025 with $330M+ ARR; launched Eleven Music cleared for commercial use (ElevenLabs, CNBC). |