The news: Meta is in discussions with Google to use Gemini as a benchmark for its own content understanding systems, per a company source.

The social media giant wants to test its systems against Gemini, not integrate the AI model, to help support its ad targeting and recommendation systems. Findings could show Gemini is stronger, or that Meta’s own systems already match or surpass it.

What Meta says: Meta communications director Andy Stone emphasized on Threads that Meta builds its own “industry-leading, proprietary” targeting and recommendation systems. Benchmarking with third-party tools is “part of the work we regularly do.” It’s unclear if Meta has previously used Gemini in this context.

Zooming out: The company is also pushing into providing its own full-fledged, AI-powered ad creation suite for brands, which could both threaten independent agencies and democratize ad creation for smaller businesses.

Sharpening how ad targeting and recommendation systems understand content and context could boost the efficacy of these tools and make them an even stronger offering for larger businesses with their own internal ad departments.

Why it matters: At first glance, this looks like an unlikely partnership between two tech and advertising giants. Beyond that, it shows how crucial AI models are becoming in the plumbing of ad systems.

For Meta, this isn’t about teaming up with a rival, but about saving time and resources on evaluation as AI competition accelerates. It could also suggest that Meta sees benchmarking partnerships as a faster path for catching up in the AI race than trying to close the gap through solo research and testing.

What this means for marketers and the industry: Building AI infrastructure on the same level as Google’s is a long and costly road, and collaboration could save Meta time. Meta is also easing rivalry here—it can use Google’s models without openly challenging its dominance in the field.

Stronger content understanding, even if only through benchmarking, could yield more nuanced insights and richer ad tooIs, enabling better campaign planning, targeting, and measurement. It highlights that AI in ads is less about flashy features and more about the invisible infrastructure that shapes outcomes.