Key Takeaways
- Google DeepMind is investing approximately $75 million in A24 as part of a long-term research partnership
- A24 filmmakers will test AI tools during actual production, giving DeepMind real-world feedback
- No concrete products or timelines have been announced; both sides describe the collaboration as exploratory
Google DeepMind and A24, the studio behind Everything Everywhere All at Once and Moonlight, have entered a long-term research partnership on AI filmmaking tools. Google is backing the deal with roughly $75 million in investment, according to the Wall Street Journal.
The arrangement is unusual. Most AI-to-Hollywood deals involve licensing existing tools or acquiring training data. This one puts working filmmakers inside a feedback loop with one of the world's leading AI labs. A24 directors and editors will test DeepMind's tools during real productions. DeepMind gets something harder to buy: professional critique from people who ship films, not demos.
What's actually in the deal?
Eli Collins, VP of Product at Google DeepMind, outlined the partnership in a blog post. The details are thin. Both sides describe a multi-project collaboration where A24 filmmakers integrate AI tools into their day-to-day workflows. In exchange, DeepMind receives feedback on what works and what doesn't.
There are no announced products. No release dates. No feature lists. Collins and A24 say they want to discover together how AI can be useful in film production. That's either refreshingly honest or frustratingly vague, depending on your tolerance for announcements that are mostly about intent.
The $75 million investment gives Google a financial stake in A24's success. A24 was valued at around $2 billion in its last funding round. The studio reportedly generates over $500 million in annual revenue and has accumulated more than 35 Academy Award nominations since its founding in 2012.
Why A24 specifically?
A24 sits at an interesting spot in the industry. It's not a major studio with legacy infrastructure and union complexities. It's not an indie scrapper working on shoestring budgets. It's a mid-sized operation that backs directors with distinctive visions and gives them unusual creative control.
That profile matters for AI research. A24 filmmakers are more likely to experiment with unconventional tools if those tools help them realize a specific creative idea. They're also more likely to push back hard when the tools don't work. That's the feedback DeepMind needs.
The studio's recent hit Backrooms, along with its track record of visually ambitious films, suggests an openness to production techniques that larger studios might resist.
The bigger picture for AI video generation
This partnership arrives as AI video generation tools mature rapidly. Runway, Pika, and OpenAI's Sora have all demonstrated text-to-video and image-to-video capabilities that were implausible two years ago. But professional adoption remains limited. The gap between impressive demos and production-ready tools is wide.
Hollywood's relationship with AI is also tense. The 2023 writers' and actors' strikes centered partly on AI concerns. Studios have experimented with AI for pre-visualization, de-aging, and voice synthesis, but full integration into creative workflows remains rare and controversial.
DeepMind's approach sidesteps some of that friction by partnering with a studio known for protecting creative autonomy. If A24 filmmakers adopt AI tools, they'll do it on their terms. That could produce more interesting results than top-down mandates from corporate production offices.
What this means for AI builders
For teams building AI products, the DeepMind-A24 structure offers a model worth studying. Instead of building tools in isolation and hoping for adoption, DeepMind is embedding itself in an active production environment. The research happens alongside real work, not before it.
That's expensive. It requires the kind of capital only Google can deploy casually. But the principle scales down: get your tools into the hands of actual practitioners early, not just friendly beta testers. The feedback from someone under deadline pressure differs fundamentally from the feedback you get in a sandbox.
Logicity's Take
The vagueness of this announcement is, paradoxically, its most credible feature. If DeepMind had promised specific products or timeline, skepticism would be warranted. Instead, they're admitting they don't know what will work yet. That's a more honest starting point than most AI-to-creative industry deals. For AI product teams, the real lesson is about feedback loops: DeepMind is paying $75 million to get professionals to tell them what's broken. Most startups can't write that check, but the principle, embedding in real workflows rather than building for imagined use cases, applies at any scale.
Open questions
Several things remain unclear. Will DeepMind build film-specific models, or adapt existing Gemini capabilities? How will intellectual property work when AI tools contribute to A24 productions? What happens to this research if it produces tools that threaten jobs, as unions fear?
Neither side has addressed these questions publicly. That's understandable at this stage, but the answers will determine whether this partnership produces usable tools or just research papers.
Frequently Asked Questions
How much is Google investing in A24?
Google is investing approximately $75 million in A24 as part of this research partnership, according to the Wall Street Journal.
What AI tools will A24 filmmakers use?
No specific tools have been announced. The partnership is exploratory, with A24 filmmakers testing and providing feedback on AI tools as they're developed.
When will products from this partnership be available?
No timeline has been provided. Both Google DeepMind and A24 describe the collaboration as a long-term research effort without concrete product announcements.
Which A24 films might use AI tools?
Neither company has specified which productions will involve AI tools. A24's recent work includes Backrooms and the Oscar-winning Everything Everywhere All at Once.
Need Help Implementing This?
Building AI tools that require real-world feedback loops? Logicity works with product teams on research strategy and go-to-market planning. Reach out to discuss how embedded user research can shape your development process.
Source: The Decoder / Maximilian Schreiner
Manaal Khan
Tech & Innovation Writer
Produced with AI assistance and reviewed by the Logicity editorial team. Learn more in our Editorial Policy.
