Google DeepMind enters A24 with 75 million: the bet isn't on generating movies, but on the tools that will make them

Seventy-five million dollars and an equity stake: with those two figures, as The Verge reported based on the Wall Street Journal, Google DeepMind seals its first entry into the capital of a film studio. And not just any studio, but A24, the house behind 'Everything Everywhere All at Once', 'Hereditary' and 'Midsommar', a byword for prestigious auteur cinema. The choice of partner is no accident: it gives Google creative credibility before a community historically wary of Big Tech.

The stated goal is to develop tools that help filmmakers 'expand their storytelling possibilities'. The company insists in its statement that the tools of the future must be 'shaped by the creators who use them'. It is a carefully chosen, almost defensive formulation, and it is easy to see why: in a chastened industry, language matters as much as investment. Scott Belsky, an A24 partner and former chief strategy officer at Adobe, was more explicit to the WSJ: what they are building 'will look nothing like' the prompt-based generation that unsettles so many people, but rather like uses that 'preserve creative control'. The entire narrative of the deal pivots on that distinction.

Two details of the agreement are worth underlining because they mark the limits. The deal is multi-year and non-exclusive, so A24 retains the freedom to ally with other tech companies. And, above all, Google will not have access to A24's film and television library. Amid a climate of copyright litigation —the article recalls the legal battles of Disney, Universal and Warner Bros against AI companies— shielding the catalogue looks like a carefully negotiated condition. A24 gains capital and technical capacity without ceding the asset that most exposes it legally.

The crack in the narrative has a name: Kane Parsons, creator of 'Backrooms', is among the artists the studio hopes to include. However, according to an interview in The Australian in early June 2026, Parsons himself stated that generative AI strikes him as 'less an innovation than a symptom of a broader cultural and economic rot' and that he gets 'no enjoyment' from using it. The article does not clarify whether he has changed his stance or whether his inclusion took place with his prior knowledge. That unknown, far from anecdotal, measures the distance between the corporate rhetoric of creative buy-in and the reality of the workshop.

Beyond the specific case, the move illustrates an underlying trend: the big AI companies no longer want to be just infrastructure providers, but partners in content creation and in the design of specialised tools. It is a shift of position in the entertainment value chain. The reasonable unknown is execution: for now there is no film announced, no tools described, no public timeline. The financial commitment is real and the signal powerful. But in cinema, as in software, the script is only judged when it is shot. We will have to wait for the first projects to know whether this preserves the artist or simply works around them.