AI makes pixels abundant. Atoms stay scarce.
AI is pushing the marginal cost of digital production toward zero. I can feel that shift in real time. Every week, another tool shows up that makes generating images, edits, scripts, and variations cheaper and faster. That is not a small change. When pixels become abundant, the advantage stops being about who can generate the most content and starts being about who can reliably capture something real, somewhere specific, with the right permissions and the right people.
In film and branded content, pixels are cheap but permission to point a camera somewhere is not. That gap is where the new leverage lives. Access matters. Permits matter. Relationships with venues, landowners, and municipalities matter. Unions, safety requirements, insurance, and compliance matter. Gear is more available than it has ever been, but crews are still built through trust and repetition. Digital production scales quickly. Physical production scales slowly, and that friction is exactly why it becomes strategically valuable.
Studios and producers that invest in sets, stages, sensor networks, and distribution fleets will outlast teams that only optimize prompts. The reason is simple. Owning or controlling physical infrastructure compounds. A permanent set amortizes across projects. A stage becomes a pipeline, not a one off expense. A calibrated workflow with a known crew becomes a schedule advantage. Meanwhile prompt optimization is real, but it is not defensible for long. The hard parts are gated by time, trust, and capital, and those gates do not disappear just because software gets smarter.
Real time virtual production is incredible, but it rides on power, bandwidth, LED volumes, and operators. I love what it unlocks creatively, especially when you can pre visualize, iterate, and shoot with confidence. But it is still physical at the core. Someone has to run the volume. Someone has to manage the color pipeline. Someone has to keep the stage running reliably. Own those capabilities and you do not just own a tool. You own the schedule, the certainty, and the ability to say yes to tight turnarounds when others have to pass.
For independent creators, the move is to partner with venues, municipalities, and suppliers. When budgets are tight, relationships become infrastructure. If you can lock in recurring access deals, you have something close to inventory. You are not just hoping a location is available when the client calls. You have a reliable option. That is hard for models to replicate because it is not data. It is a set of agreements, a track record, and a level of trust that reduces risk for everyone involved.
AI will flood channels with near infinite variations of the same idea. We are already seeing it. The first wave looks impressive, then it starts to blur together. In that environment, what cuts through is footage from a place you cannot spoof or a performance you had to be there to direct. Not because AI cannot imitate reality, but because audiences can feel when something has real constraints behind it. Real weather, real crowd energy, real stakes, real access. Physical production adds texture, and texture is becoming the scarcity.
Expect rates to diverge. Commodity post goes down while day rates for specialized grips, drone pilots, and location managers go up. That is a rational market response. If AI compresses the value of tasks that are purely digital and repeatable, then the premium shifts to roles that combine taste with physical execution and responsibility. Safety, planning, and logistics are not glamorous, but they are increasingly the difference between a shoot that happens and a shoot that collapses.
Even robotics points the same way. OpenAI working with Figure on embodied systems reinforces that value accrues where software meets factories, warehouses, and safety frameworks. Digital intelligence is impressive. Embodied intelligence is expensive. The minute software has to deal with the messiness of the physical world, you enter a domain defined by sensors, supply chains, regulation, and operational excellence. It is the same theme, just at a different scale.
In an AI rich world, build or rent the bottlenecks that touch the ground. That might sound unromantic compared to talking about new models, but it is the practical play. If everyone can generate concepts, storyboards, and variations at speed, then the constraint becomes execution in real environments. The bottlenecks are locations, crews, power, transport, permissions, and the operational muscle to coordinate them.
If I were greenlighting projects this year, I would fund permanent sets, mobile field kits, and permit pathways before another seat of a new editing tool. Editing tools will continue to get cheaper and better. Access does not. A mobile kit that can deploy fast turns into a promise you can make to clients. A permit pathway, built through repeat collaboration and a clean track record, becomes a defensible advantage. Those investments show up as speed, exclusivity, and leverage at the negotiating table, which is where a lot of margin is actually decided.
Bet the farm on the physical. The models are a force multiplier, but atoms are the moat. I do not read that as anti AI. I read it as a reminder that the most resilient businesses and careers will sit at the intersection. Use AI to compress the digital steps. Then put the saved time and money into the parts that cannot be copied: relationships, infrastructure, access, and real world production capacity.
Joshua Campbell
Director