<- Back to Blog

Canal+ Signs with OpenAI and Google Cloud — and Turns Creative Production into an AI Pipeline

Canal+ Signs with OpenAI and Google Cloud — and Turns Creative Production into an AI Pipeline
Posted 3/17/26
7 min read

Canal+ just announced dual partnerships with OpenAI and Google Cloud, embedding generative AI into both its streaming app and its production workflows. With 40 million subscribers and a target of 100 million by 2030, the first major French media group to make this move is signaling something larger than a product update: the production infrastructure behind premium content is being rebuilt around AI — and the question for every creative operations team is what that means for how work gets made.

  • Canal+ will deploy natural language search via OpenAI and multimodal indexing via Google Cloud starting June 2026
  • Production partners get access to Google's VEO 3 for scene previsualization and archive reconstruction
  • The move follows Disney's $1 billion OpenAI deal and accelerates AI adoption across the European media industry

On March 11, Canal+ CEO Maxime Saada announced two multi-year agreements — one with OpenAI, one with Google Cloud — during the group's annual results presentation. The timing was deliberate. Canal+ had just closed its acquisition of South African giant MultiChoice, reported 42.3 million subscribers across 70+ countries, and outlined a path to 100 million by 2030. The AI partnerships aren't an experiment. They are infrastructure for the next phase of the business.

What Canal+ Is Actually Building

The two deals serve different layers of the same system, and Saada described them as « très complémentaires »: Google extracts and structures data, OpenAI exploits it for search and recommendation.

Starting in June 2026, the Canal+ app will replace its keyword-based search with a natural language interface powered by OpenAI's frontier models. Subscribers will describe what they want to watch by mood, genre, or vague impulse — « a heartwarming romantic comedy » or « something light and entertaining » — and the system will return personalized results. Ashley Kramer, VP of Enterprise at OpenAI, framed it as making entertainment « simpler, more intuitive, and more personal. »

On the Google Cloud side, Canal+ will accelerate the indexing of its entire content library into a multimodal database combining sound, video, and text. The enhanced classification feeds directly into the recommendation engine, enabling more granular personalization based on viewing behavior — not just titles and genres, but what's actually happening inside the content.

The deployment will cover all European and African markets where the Canal+ app operates. For a group that has quadrupled its subscriber base in ten years, the AI layer is less about novelty and more about operational necessity: no traditional search engine can serve a catalog this fragmented across this many markets.

VEO 3 Enters the Production Pipeline

The less-publicized part of the announcement is the one that matters most for creative production.

Canal+ will give its production partners access to Google's VEO 3 generative video model. The stated use cases: previsualize scenes before shooting, or reconstruct historical moments from a single archival photograph. Canal+ CTO Stéphane Baumier described the goal as enabling « AI-generated video scenes, impossible to produce using traditional methods. »

Canal+ was careful to frame this as a creative tool, not a replacement. Intellectual property remains protected within Google Cloud's secure environment. Editorial control stays with production teams. No one at Canal+ is claiming VEO 3 will write scripts or direct episodes.

But the operational signal is clear. Pre-visualization — which traditionally requires storyboard artists, rough cuts, and weeks of back-and-forth — can now be accelerated to hours. Archive-based reconstruction, which required manual compositing and expensive VFX, gets a generative shortcut. The bottleneck shifts from asset creation to asset evaluation: which generated output is good enough to move forward, and who decides.

This is the same dynamic reshaping creative project management across the industry. When production tools generate more variants faster, the critical capability becomes structured comparison, traceable approvals, and version control that doesn't depend on someone remembering which file was the latest. The volume problem that generative AI creates is real — and it doesn't solve itself.

The Pattern: Disney, Canal+, and What Comes Next

Canal+ is not the first major media company to embed generative AI into production workflows. In December 2025, Disney signed a $1 billion deal with OpenAI, licensing 200+ characters for Sora and deploying ChatGPT APIs across Disney+ and internal operations. The deal positioned AI not as an add-on but as foundational infrastructure for content creation and distribution.

Canal+'s approach is structurally different. Where Disney focused on consumer-facing content generation (letting fans create with characters), Canal+ is embedding AI directly into the production pipeline — indexing, pre-visualization, archive reconstruction — while also transforming the subscriber-facing experience. The combination of upstream production tools and downstream recommendation systems in a single strategic push is unusual. It reflects the reality of operating a super-aggregator across 70+ markets: the operational complexity demands infrastructure that traditional production tools cannot provide.

For creative teams and agencies that serve media groups, the implication is direct. When a client like Canal+ hands you VEO 3 for pre-visualization, the deliverable changes. You're no longer submitting storyboards — you're submitting generated scenes that need to be reviewed, compared, annotated, and approved within a structured creative workflow. The asset volume increases. The review cadence accelerates. And the infrastructure to manage that volume becomes the differentiator, not the production talent alone.

What This Means for Creative Ops

The Canal+ announcement crystallizes a shift that has been building across the media and entertainment industry. AI is no longer a tool that sits alongside production. It is becoming the environment in which production happens — from content discovery to pre-production to delivery.

For Creative Ops leaders, three things change immediately. First, asset volume will increase — not gradually, but structurally. When VEO 3 generates scene variants in hours instead of weeks, the review pipeline needs to absorb that volume without breaking. Second, multimodal indexing — cataloging content by sound, image, and text simultaneously — creates new requirements for asset management and organization that most teams aren't equipped for today. Third, the coordination between AI-generated outputs and human editorial decisions needs a governance layer that is currently missing from most production workflows.

Canal+ claims its creative teams retain full editorial control. That's a statement of intent, not a description of infrastructure. Making it true at scale — across 70+ markets, with generative tools in the hands of external production partners — requires centralized coordination, traceable approvals, and structured version control that go far beyond what a shared drive or email thread can provide.

The question for the industry is no longer whether AI enters the production pipeline. Canal+ just answered that. The question is whether Creative Ops teams have the infrastructure to keep up with what comes out the other end.

FAQ

What exactly did Canal+ announce on March 11? Canal+ signed two separate multi-year partnerships: one with OpenAI to power natural language search and content recommendations in the Canal+ app, and one with Google Cloud for multimodal content indexing and access to the VEO 3 generative video model for production use. Both roll out starting June 2026 across European and African markets.

How does this compare to the Disney/OpenAI deal? Disney's deal focused primarily on consumer content creation — licensing characters for Sora and ChatGPT Images. Canal+ is embedding AI deeper into the production pipeline itself, combining subscriber-facing recommendation with upstream tools for pre-visualization and archive reconstruction. The scope is narrower in budget but broader in operational integration.

What is VEO 3 and how will production teams use it? VEO 3 is Google's generative video model. Canal+ will make it available to production partners for two primary use cases: previewing scenes before they're shot, and reconstructing historical moments from archival photographs. Canal+ maintains that editorial control remains entirely with human creative teams.

Does this affect how agencies and production partners work with Canal+? Yes. If production partners receive AI tools that generate scene variants at scale, the review and approval workflow changes fundamentally. The deliverable shifts from static storyboards to generated video outputs that need structured comparison, annotation, and traceable sign-off — a capability that requires dedicated production infrastructure.

Sources

Other Posts

The Missing Generation: Who Will Run Creative Production in Three Years?
DEEP DIVE

The Missing Generation: Who Will Run Creative Production in Three Years?

When Every Team Member Uses Their Own AI, the Brand Loses Its Thread
DEEP DIVE

When Every Team Member Uses Their Own AI, the Brand Loses Its Thread

AI Content Labeling: Why Voluntary Disclosure Wins
IA & AUTOMATISATION

AI Content Labeling: Why Voluntary Disclosure Wins