Ben Lerer on AI film studios, Hollywood's resistance, and why the short-form drama boom hasn't crossed the Pacific

Summary

Ben Lerer, managing partner at Lerer Hippeau, sees AI-driven film production colliding with Hollywood's entrenched resistance. The industry is moving slower than almost any other sector facing automation.

Lerer's portfolio company Staircase is building an AI-first film studio using SAG-affiliated actors and motion capture to produce content entirely without on-location shoots. Talent provides voice and performance data via capture, then AI handles the rest. Everything, not select scenes, gets produced this way. Technology has advanced dramatically in the past six months, and Lerer sees the current trajectory making traditional location shooting absurd within the near term.

Hollywood's resistance is structural. Unions lock in production workflows. Moguls and old-money operators move slowly. Unlike most industries facing disruption, Hollywood has regulatory scaffolding that slows adoption. Lerer acknowledges the friction directly: "At some point, it's just gonna get steamrolled." But frontline talent like Brad Pitt will likely extract enormous premiums even as production costs crater. The real question is what happens to the middle class of acting talent in LA, the waiters trying to break in.

Emerging actors could bypass traditional studio channels by building personal brands on social media. If they demonstrate pull and audience size, they gain pricing power independent of whether a studio films them on location or synthesizes their performance. That flips the leverage.

On short-form drama, Lerer is skeptical despite seeing a pitch once a week. Chinese platforms like Reel Short have built enormous audiences with vertical drama content. The US equivalent keeps failing to materialize. Live shopping followed the same arc—massive in China, never took off here. Lerer forced himself to watch some content and got hooked for two hours on a story about a woman meeting a man in an elevator. The addictiveness is real. The venture scale isn't.

Lerer draws a firm line: "The difference between an interesting space and a venture scale space is the Grand Canyon." Most of media has turned out not to be venture scale. The TAM for short-form drama in the US, even with AI, looks constrained. The competitive moat is nonexistent. Meta, Amazon, Apple, and Netflix each own a social network. Every major-seven company except Nvidia has a media platform. Building an independent short-form drama play against that lineup is futile. Quibi, backed by Jeffrey Katzenberg with hundreds of millions, proved the point. Even with better technology today, the competitive dynamics haven't changed.

Lerer is hunting for categories that AI can now unlock for the first time. His thesis finds industries that never digitized, not because they're backwards, but because the problem was previously impossible to automate. Aftermarket automotive parts exemplify this. Millions of SKUs exist with no standardization. Customers still fax orders. Most inventory sits with endemic players like AutoZone, which cannot answer whether a spark plug fits your 1984 Miata after you've made other modifications. That represents tens of billions in annual US TAM, one of the biggest hobbies in America, still largely offline. AI agents buying on behalf of humans could mimic what traditional e-commerce never achieved. Pair that with an embedded AI mechanic teaching buyers how to install what they buy, and the result differs fundamentally from a DoorDash clone.

On labor marketplaces, platforms like Mercos positioned as marketplaces feel more like enterprise staffing products. Growth has been astounding. Lerer is wary. Some smart investors think the whole category is a race to the bottom heading toward zero. Founders might be wise to take secondary sales now while momentum is hot. The broader market is flooded with categories growing at multiples that look unsustainable. Inference companies, training data platforms, and companies with only four or five long-term customers are everywhere. Massive capital is pouring in because of FOMO and because huge funds need to deploy. That creates self-fulfilling prophecies. Lerer expects fewer winners and lower final multiples than current valuations suggest.

On OpenAI versus Anthropic, the vibe shifted hard in three months. Anthropic is "the main character" now, dominating social media and legacy coverage. OpenAI looks defensive by comparison. Lerer notes this could be optical. Google Trends tell a different story. Meta hasn't shipped yet. The market is in a bubble, and people are invested in drama. It's the F1 race metaphor, where audiences want passes and upsets, not Max Verstappen winning every time. It's too early to say whether OpenAI's multiproduct, consumer-plus-enterprise-plus-hardware bet will outperform Anthropic's focused strategy.

On LA venture capital, Lerer describes it as a drought. He's not hunting companies in LA. His advice to founders not in hard tech is stark: don't start your raise in LA. It signals you're not serious. The money is in San Francisco. The geography has shifted decisively.