Liu Haonan and Qiao Yuhang have been finishing each other’s stories since high school, where they drew together, talked about animation and film, and traded ideas for comics and stories of their own. More than a decade later, they’re doing something a two-person team couldn’t have pulled off a few years ago: building a science-fiction universe spanning 30 light years, almost entirely by themselves, using AI production tools instead of a film crew.
Hollywood has largely controlled who gets to tell stories at scale, largely because scale itself required budgets, crews and infrastructure most creators never had access to. A newer wave of AI production tools is starting to loosen that grip, and the companies building it are converging on Venice this year to make their case.
Chief among them is Tamar AI, a Palo Alto company building what it calls the creative infrastructure of the AI era: production tools, streaming platforms and IP incubation programs meant to carry an idea from concept through to an audience. Its flagship product, TapNow, is pitched as an AI-native creative operating system, bundling multimodal AI, intelligent agents and production workflows into one environment that the company says people use for everything from short films to marketing campaigns and presentations. TapNow says more than 3 million people use the platform daily, spread across 150 countries and regions, and runs a creator community, TapTV, alongside an events arm, TapTV Arena, that has staged more than 30 international hackathons, residencies and creator summits.
“AI doesn’t replace the filmmaker — it closes the gap between what you imagine and what you can see,” says Tianshu, head of global AI product solutions at Alibaba Token Hub, one of TapNow’s technology partners. “For us, that changes everything about how stories get made. This aligns well with Tamar AI’s belief that AI amplifies human creativity.”
That philosophy fuels 10,000 Parallel Universes, TapNow’s global creator and AI-native IP incubation program. It‘s framed as an alternative to a traditional producer-led development pipeline, built to discover new voices and original stories from around the world. The program’s first season drew more than 2,000 submissions and offered a prize pool topping $500,000. Selected teams received computing support to produce their projects, and in some cases funding so creators weren’t working under financial pressure. TapNow says it didn’t take copyright in exchange for that support, and it helps teams fill gaps in sound or art and connects them to festival and distribution contacts.
What that support looks like on the ground shows up in how Liu and Qiao actually built “Primordia,” one of the program’s standout projects. Liu worked as a concept designer on film and television productions, including several theatrically released Chinese films, translating screenplays into images so a production designer would have something concrete to discuss with the director. He and Qiao stayed creative partners the whole time, at one point hauling equipment to an island together to shoot a student film. “Primordia” confines its story to a cluster of star systems the pair call the Ore, home to dozens of inhabited planets and a war over resources mined from ancient creatures. A frontier boy named Jesse rescues a fugitive military officer named Rook, and the bargain between them pulls both into empire, war and the secret of Jesse’s own origins.
Still from “Primordia,” created by Liu and Qiao
Image courtesy of “Primordia”
Qiao traces the appeal of AI filmmaking to scale as much as tools. “In the past, even a relatively small location shoot for a short film might involve 30 or 40 people,” he says. “The director still had to handle a huge amount of organization, management and communication. Everyone would come to you asking what happened next, and your creative train of thought was constantly being interrupted.” On “Primordia,” he says, it’s mainly the two of them, working through writing, shots and detail before moving into execution.
Liu puts it in terms of where his time actually goes. “The reason I first wanted to move into AI filmmaking really was ‘people,’” he says. “In traditional production, you sometimes spend less time with the work itself than you do dealing with people and logistical matters. With AI filmmaking, the work becomes the thing I spend the most time with.”
The pair chose a series format over the AI shorts that have become common online, betting that “Primordia’s” characters and history needed more than 10 or 15 minutes to land. Liu says he wants the project to have the texture of something “that could genuinely live on a platform and continue to develop.” Production runs through TapNow’s own tools, including an agent system Liu uses to turn conversational shot descriptions, camera moves and blocking into instructions the underlying models can act on. Qiao says the creators lean on the same system to organize the world and screenplay, using it less as a decision-maker than a collaborator that offers structure and pushback on writing that still starts with them.
“Primordia” is the clearest evidence for a broader claim: that AI tools don’t just make filmmaking cheaper, they change who gets to run the experiment in the first place. It’s an argument that 10,000 Parallel Universes is building its own streaming platform around, betting that AI can trade the traditional model of fewer, bigger-budget bets and longer production cycles for more creators, more experiments and more finished worlds moving through the pipeline at once. The company has also argued that a project doesn’t need to appeal to everyone to work on a platform where audiences pay to follow serialized stories, and that a distinct authorial voice holding onto a smaller, committed audience is worth something on its own.
Some of that infrastructure runs through HappyHorse, the AI video generation model housed inside Alibaba Token Hub (ATH), the independent core business group Alibaba Group set up in March to bring together its foundation models, model services and AI applications under one roof. Rather than functioning as a finishing tool, HappyHorse is built to help filmmakers visualize and test ideas earlier, spanning text-to-video, image-to-video, reference-to-video and editing workflows that carry a project from concept exploration through previsualization. HappyHorse served as technology partner on three films from the 10,000 Parallel Universes program, all selected as winning works headed to Venice.
Tamar AI and HappyHorse will co-host a program at the Venice Production Bridge built around that slate. The panel “AI + Cinema: The Next Generation of Storytelling and Creativity,” runs Sept. 4 at Spazio Incontri in the Hotel Excelsior, with the president of CILECT Europe, a senior executive from Italy’s Rainbow Group and representatives from both companies. A curated screening of selected films from the 10,000 Parallel Universes initiative, “New Voices in AI Cinema,” including the three HappyHorse-backed titles, follows Sept. 7 at Sala Pasinetti in the Palazzo del Casinò.
Sabrina Qin, brand lead at TapNow, has framed her own optimism about AI filmmaking as a response to Guillermo del Toro’s perspective on AI at this year’s Cannes Film Festival. “Wait, till AI blows your mind,” she says.
The clearest sign of where TapNow wants all of this to go is Glanze, a streaming platform the company is preparing to launch out of the 10,000 Parallel Universes initiative and power with TapNow’s production tools. Every title on Glanze starts as what the company calls a Glance, a first-look preview that can grow into full episodes, seasons and an expanding world audiences return to; viewers can press play immediately, reserve what comes next or dig into the places and artifacts behind a story. TapNow describes the model as pairing the serialized, ongoing relationship between creators and audiences familiar from Japanese manga with AI production tools that cut the cost and complexity of building out an ambitious world over time.
The web version of Glanze launches Sept. 4 with more than 40 series Glances live and five series beginning official serialization; a mobile app follows in late October with more than 30 series serialized and over 100 series Glances to browse.
Glanze is ultimately a bet that there are more Liu-and-Qiao-type stories out there, waiting on tools rather than budgets. “Primordia” carries forward a project the two have been building toward since their teenage years, now made with tools that didn’t exist when they started drawing together. Asked what advice they’d give the next creator trying to build a “parallel universe,” Liu keeps it short: “Just do it. Take the first step.”