Microdrama platforms are back in the headlines — and this time they’re promising to do what Quibi couldn’t: make vertical, short-form storytelling profitable by leaning on AI-assisted production. In this episode of Denoised, hosts Addy and Joey break down MicroCo’s plan to revive the “vertical shorts” model, show examples of VFX artists using local AI workflows to produce film-quality results, and map the shifting landscape of physical production from Texas and London to a resurging Los Angeles.
MicroCo and the rebirth of vertical microdramas
MicroCo — a new U.S. start-up led by veteran executives including Lloyd Braun — is betting on the same vertical microdrama model that has taken off in China: short, vertically framed episodic content designed for TikTok-style funnels, with paywalls to unlock later episodes. Their big headline is a dramatic cost claim: using an AI-assisted pipeline (via an acquired company called Intelligent Animation) MicroCo says it can produce content for as little as $1,500 per minute.
That price is important to contextualize. Quibi’s failure is often attributed to very high production costs (reportedly $100k–$125k per minute for live-action, A-list talent-driven content). Chinese microdrama factories, by contrast, produce thousands of episodes at extremely low per-episode costs — seasons can range from dozens to over a hundred episodes, with episodes often 1–3 minutes long and whole seasons reportedly produced for figures vastly lower than western live-action budgets.
The bottom line: the MicroCo number likely mixes animation-first workflows with AI automation. Animation is presently the easiest genre to compress aggressively with AI; true live-action parity is harder and still evolving. The comparison between Quibi’s live-action spending and MicroCo’s quoted “per-minute” AI/animation cost is not apples-to-apples.
Where efficiency comes from: virtual production, semi-permanent sets and vertical framing
For live-action vertical shorts to scale profitably, the most plausible path is combining several efficiency levers:
Virtual production (LED/IBL) and pre-built background systems so crews don’t build new environments for every scene.
Semi-permanent sets or multi-floor studio buildings that mimic the “factory” model used by some Chinese studios.
AI plate generation and scene-extensions focused on narrow vertical compositions — wide shots are less important for vertical-first storytelling.
Tools like V1-style capture systems and emerging Vu Studio features (integrated lighting + background tablets) make it easier to generate tall plates and consistent lighting quickly. For producers, creating a small, well-rigged warehouse with LED walls and repeatable grids can dramatically shorten turnaround and crew time.
AI VFX experiments worth watching
Some of the most interesting work today isn’t coming out of big studios — it’s coming from seasoned VFX artists experimenting locally. Greg Teegarden, a Digital Domain VFX supervisor, has been posting experiments built on local instances of ComfyUI and the Wan models (2.1 / 2.2). Two demos stand out:
Will Smith opera-driven sequence: Teegarden used a combination of LoRAs and ControlNet-style conditioning to drive a multi-second character performance (roughly 873 frames) locally — a continuous shot longer than the typical public-tool duration limits.
Alien/Ian Holm pass: Teegarden took a single reference frame and used an image-to-video workflow to reduce uncanny-valley artifacts from a practical prosthetic used in a recent Alien instalment. The AI pass smoothed realism in ways that made the shot feel more lifelike than the original practical or the studio VFX replacement in certain respects.
Those demos demonstrate a recurring theme: expert human taste + AI tooling beats generic AI output. Experienced VFX artists know how to push tools, compose imagery, and fix imperfections. When they apply that judgment to models and pipelines (ComfyUI workflows, LoRAs, ControlNet), the results become genuinely useful for production.
Wan 2.2, ComfyUI templates and practical VFX workflows
Juan (WAN) 2.2 and ComfyUI templates have improved speed and temporal control. Tests that once took tens of minutes per short clip can now be prototyped in minutes at lower resolution (examples: ~2 minutes for a 5-second low-res clip; ~5 minutes for 720p). That means feasible iteration for directors and VFX supervisors at the previsualization and first-look stages.
Recommended practical workflow the hosts discussed:
Prototype motion and timing at low resolution using Juan 2.2 + ComfyUI templates (fast iteration).
Use a detail-insertion model (for example, enhancer-style tools) to add photoreal skin and material detail to key frames or passes.
Use the enhanced outputs as references for a higher-resolution generation, then upscale and composite final plates (Topaz or studio upscalers/jitter-stabilization as needed).
This staged approach decouples “what moves and when” from “how photoreal it looks,” letting teams lock performance before investing heavy render time.
Where production is moving: Texas, Marvel's relocation, and LA's rebound
Physical production continues to redistribute globally and within the U.S. — incentives, labor costs, and anchor clients determine where large projects land.
Texas: a new studio hub for serialized content
Taylor Sheridan (Yellowstone and its spinoffs) and Paramount are building a new studio complex in Texas as part of the AllianceTexas campus. The build is expected to include multiple stages capable of hosting several large-scale productions simultaneously and to serve as a home base for Sheridan’s Paramount-produced slate — a good example of how tentpole TV/IP owners are vertically integrating production locations.
Marvel and the Atlanta to London shift
Recent reporting has flagged a partial shift of Marvel production from Atlanta (Trilith) to London. Factors cited include tax incentives, wage and benefits math (where healthcare and employer contributions differ by country), currency effects, and studio buildouts overseas. Two important nuances:
Marvel is producing fewer theatrical titles than during its peak output years, so some of the Atlanta slack may simply reflect lower volume.
Being overly dependent on a single anchor client is risky; when a major client shifts or reduces output, local infrastructure feels the impact. Trilith and regional hubs are responding by diversifying client bases and investing in local talent incubators to reduce exposure to any single “Lucky Strike”-style client dependency.
Los Angeles: incentives and a possible comeback
Los Angeles has been taking steps to win production back. The state and city recently boosted film incentives (the updated figure discussed during the episode was $750 million) and changed permitting processes to reduce red tape — a reported uptick in approvals suggests LA is trying to make on-location production once again efficient and predictable. For local production teams, that’s promising: easier permitting and larger incentive pools help tip the scale for projects with above-the-line talent and complex logistics.
Key takeaways for filmmakers, VFX artists and producers
Animation is the short-term AI frontier. If you want to test AI cost savings today, start with animation or hybrid 2.5D approaches where AI can replace many hand-drawn or CGI steps.
Prototype cheap, then refine. Use fast, low-res generation to lock timing and performance. Apply enhancer/detail-insertion passes and upscales only after the edit and motion are agreed.
Invest in repeatable production rigs for vertical work. Small VP stages, LED walls, and a few semi-permanent sets can dramatically reduce location costs and set-up time for microdrama pipelines.
VFX artists who understand framing, light, and fixes will win. Human decision-making — knowing how to tweak and composite AI output — remains the most valuable skill in these pipelines.
Watch incentives and diversify location risk. Studio buildouts, tax incentives, and anchor clients can shift quickly. Plan for multiple revenue streams and relationships rather than one dominant client.
Collaborate with short-form creators. YouTube and short-form talent understand attention mechanics and cliffhangers; their input is valuable when reformatting genres like horror or action into micro-episodic structures.
Conclusion
AI is lowering the barrier to producing vertical and animated content, and experienced VFX hands are already turning those tools into production-ready assets. At the same time, physical production is rebalancing: new studio investments in Texas, shifting big-buck franchises overseas, and targeted incentives in California are reshaping where work gets done. For filmmakers and producers, the practical route forward is hybrid: prototype fast with AI, keep creative oversight tight, invest in repeatable production infrastructure, and stay nimble about location and client risk.
