Earlier this week, Disney announced another round of layoffs, this time targeting Pixar, National Geographic, and the NFL Network (part of ESPN). This latest round of cuts follows a round in April that impacted Marvel Studios , among other divisions. During that round, layoffs were attributed to “general cost-cutting and workforce reductions” along with a reduction in the slate of Marvel films. But it also follows a broader industry trend of cutting staff roles and using freelance contractors to do the work – reporting in April claims that “much of the entire department has been let go with only a small team of full time production staff and artists remaining in place to coordinate the hiring of resources on a per-project basis.”

Companies laying off staffers to save the costs of insurance and retirement is nothing new, but the adoption of AI into the entertainment industry and post-production workflows adds another element to the mix. After adopting the technology at a blistering pace last year, companies have been starting to reckon with the high cost of token usage and, in some cases, beginning to pull back. Video production tends to be especially costly – OpenAI’s Sora was estimated to cost $15 million per day in compute resources, which dwarfed any reasonable revenue model, especially with the collapse in engagement numbers and copyright litigation. OpenAI wound up killing the platform earlier this year, unraveling a billion dollar pledge from Disney.

So the smart math, then, is to start outsourcing those big AI costs. Freelancers have long had to pay for their own tools, and AI services are becoming just another cost of doing business. But the economics of that start to fall apart fairly quickly, and that could have a negative impact on both the studio and the AI companies.

Already, Hollywood is facing a jobs crisis. A nationwide downturn in U.S. film and TV production caused a 30% employment drop from late-2022 that continues to this day. Screenwriters are now doing construction or training AI systems ; many jobbing actors are struggling due to AI replacing roles in industrial training films . There will always be people who head to LA to make it big, but there will also be plenty of rational actors who decide to pivot to another industry when the cost of doing business becomes too high.

Then the future looks like this: studios cut to the bone and outsource production to freelancers and small shops. Those shops either have to charge much higher rates due to the cost of AI-powered production, obviating any cost savings, or just do things the old-fashioned way. That would be a wash for the studios, but could have real impact on the AI companies that need to get to scale in order to cover their compute costs and satisfy their investors. If AI doesn’t save time or money, what’s the point of it?

The companies building these products face a delicate balancing act. Some of the larger firms in the space, like Microsoft and Google, have the runway to keep discounting their products as they build data center capacity, although eventually the bill will come due. They have to strike partnerships with post-houses and freelancers by explaining that they will solve their problems and save them money without replacing them altogether – it’s an absolutely plausible and probably correct take, but the backlash against AI, especially in creative industries, is very real.

I’ve personally spent ten years producing and directing immersive content, with an Emmy nomination and CAA representation under my belt, and have used plenty of AI tools as part of my workflow. Some are great ( Google Genie , in particular, has been excellent for scoping virtual worlds), but many of them are simply too expensive to justify the time saved or require too much upskilling to be a seamless replacement. This tension was played out in an ad for Coca-Cola that was produced with AI last year – it didn’t seem to save any time or money for the brand and the ad itself was just fine.

AI companies need to thread the needle of communicating value and keeping costs low across a vast swathe of small shops and freelancers, which is no easy feat. To succeed, they need to make sure they provide real value that can be integrated across the entire workflow, not just individual parts and pieces, and do so at a cost that is less than just doing it the old-fashioned way.