The Data Delusion: Meta's European Slump Exposes the AI Ad Myth

For years, the pitch from Silicon Valley has been the same: AI will optimize everything. The promise is an ad engine so sophisticated it can predict consumer desire with surgical precision, regardless of the inputs. But as I look at Meta's latest numbers coming out of Europe, that narrative is hitting a wall of reality.

According to reporting from Wired, as the outlet first reported, Meta is seeing a clear deceleration in European advertising revenue. In the three months ending in September, growth in the region sat at 18 percent year-on-year. While that sounds positive in a vacuum, it is a stark underperformance compared to the 33 percent growth seen in the US and Canada. This isn't a fluke; Wired reports that growth in the region had previously slid from approximately 40 percent down to 24 percent over the preceding two quarters.

***Opinion:*** *The market is finally seeing through the 'AI-driven' smoke screen. Meta's struggle proves that no matter how advanced the recommendation engine is, it cannot function without the raw, invasive data harvest. When you starve the engine of personal data, the 'intelligence' vanishes, leaving behind an inefficient machine that advertisers aren't willing to pay a premium for.*

This crisis is the direct result of regulatory pressure. Wired reports that EU regulators fined Meta €200 million after ruling the company breached Digital Markets Act obligations by failing to give users a choice of a service that uses less personal data. Meta had previously attempted to satisfy this by introducing an ad-free paid service in November 2023, but the European Commission ruled this insufficient. Consequently, Meta rolled out the 'Less Personalized Ads' (LPA) model at the start of this year.

Meta's own admissions reveal the fragility of the model. The company states that LPA uses 90 percent less data than its personalized counterparts, relying instead on session-based activity and basic demographics like gender, age, and location. In a response to the EU ruling, Meta warned that this leads to poorer returns for advertisers, noting that these ads are less efficient, result in fewer transactions, and are more frequently closed by users for being repetitive or irrelevant. This is reflected in the P&L: Wired notes that the price advertisers are willing to pay per ad in Europe is not growing at the same pace as it is in North America.

Meta's leadership is attempting to pivot the conversation toward future AI capabilities. During a Wednesday earnings briefing, CFO Susan Li told investors the company has "a lot of runway" to improve ad conversions and the performance of its recommendation engine. Li also highlighted the commercial potential of Muse, Meta's personal AI agent launched last month, noting that users are utilizing the tool for product research and recommendations.

Dominic Ball, equity research director at Rothschild & Co Redburn, told Wired he believes Meta can eventually use AI to improve targeting even with fewer data signals. But for now, the evidence is clear: the 'AI revolution' in advertising is entirely dependent on the data pipeline. When the pipeline is constricted, the revenue follows.