Tencent forecasts AI in up to 30% of long-form film and animation
Tencent projected that 10 to 30 percent of long-form film and animation could be dominated by or deeply involve AI within two years, framing generative tools as a supply-side shift in premium screen content.
Treat the number as directional rather than precise. The useful signal is that a major platform is planning capacity around it, and animation is named first — the discipline where generative work has the shortest path to finished frames and the fewest likeness and clearance problems.
Forecasts from platform executives usually say more about the platform than the industry, and this one is useful for exactly that reason. Sun Zhonghuai, a Tencent vice president and CEO of Tencent Video, was describing what one of the largest buyers of premium content in China expects to be commissioning, not making a neutral prediction. When a buyer says the shape of its supply will change, producers selling into that market should listen to the details as much as the headline.
The details are more careful than the headline. Speaking at a forum at the 7th Hainan Island International Film Festival, Sun put the share of film, television, and animation works “dominated by or deeply involving AI” at between 10 and 30 percent over the next one or two years. The one-third figure is the top of a range, not a point estimate. He was also blunter about the obstacles than most vendors are: AI video still cannot resolve visual consistency and controllability, character details drift visibly over extended shot durations, human micro-expressions are not reproduced, and most mainstream output tops out at 1080p, short of 4K or DCI theatrical standards.
Short-form content, in his account, has already gone further. Sun compared short dramas and short animation to fast-moving consumer goods that ruthlessly pursue efficiency, and estimated that AI-native companies now account for nearly 30 percent of the short-animation market. The long-form forecast reads as an extrapolation from that, tempered by the quality gap.
The most practical part of the talk is the least quoted. Tencent Video says it is already using AI in early-stage concept design, previs, and rotoscoping, and in LED virtual production to build digital sets for background scenery. Those are not generated films. They are AI inside conventional stages of production, the same pattern Netflix described in July when it reported generative AI on roughly 300 titles across the pipeline. For a VFX or post vendor, roto and previs are where the work is changing first, and that is where bids and staffing plans should be revised before anything else.
We would treat the resolution point as the real constraint for long-form. Drift over long shots can be worked around with editorial choices, shorter takes, references, and comp. Delivery specs cannot be negotiated away: a platform that requires 4K masters needs upscaling, finishing, and quality control steps that generative output does not yet supply on its own. Anyone quoting AI-heavy long-form work should price that finishing stage explicitly rather than assuming the model output is the deliverable.
Sun also argued that lower costs would revive genres previously too risky to greenlight, naming grand science fiction and “Age of Discovery” narratives, and said the future talent model needs “art engineers or tech artists.” The first claim is worth testing against what actually gets commissioned over the next two years. The second is already true in any hybrid pipeline: the scarce skill is the person who can judge a frame and also fix the system that made it. What to watch is whether Tencent’s own commissioning data, rather than its forecasts, starts to appear.