AI Ad Optimization Fuels Revenue Growth Without Bigger Audiences
Meta’s latest results highlight a decisive shift in digital advertising economics. The company reported strong top-line gains, with advertising revenue rising on the back of 19% growth in ad impressions and a 12% increase in average ad pricing, even though daily active users grew only modestly. The message for marketers is clear: AI ad optimization is allowing platforms to squeeze more value from the same attention pool. Meta is using AI to refine targeting, campaign optimization, pricing efficiency, inventory utilization, and creative recommendations in tandem, creating a flywheel where more effective ads justify higher prices and denser ad loads. This efficiency does not necessarily translate into lower costs for advertisers; instead, the platform captures a greater share of the performance upside. As AI becomes the operating layer of Meta’s commercial engine, performance that once depended on human media buying is increasingly governed by opaque machine-driven decisions.
Infrastructure Arms Race and Deepening Platform Power
Behind these gains is a massive infrastructure bet. Meta has sharply raised its capital expenditure forecast to as much as USD 145 billion (approx. RM667 billion), channelling spending into data centres, AI models, and automated advertising systems. With operating cash flow of USD 32.23 billion (approx. RM148 billion) and cash and investments of USD 81.18 billion (approx. RM373 billion), it can afford to pursue AI scale that few rivals can match. This scale turns infrastructure into a competitive moat: more compute supports more advanced AI, which drives better programmatic advertising performance, which in turn funds further investment. For advertisers, this concentration of capability means that top-tier performance is increasingly tied to a handful of platforms able to sustain such spending. As AI-optimized auctions grow more sophisticated, the bargaining power tilts toward the platforms, raising questions about how much of the value created by optimization flows back to marketers versus being absorbed as ad revenue growth.
Agentic AI Marketing: From Automation to Algorithmic Dependency
Meta’s introduction of AI connectors, along with tools like Meta AI for Business and AI-driven campaign optimization, points toward a world of agentic AI marketing where software agents actively manage campaigns end to end. These systems automate tasks once handled by performance teams, from bid adjustments to creative rotation and cross-channel learning. At the same time, new research from Taboola shows that 76% of advertisers are already seeing uplift from agentic AI solutions, mainly in search and social walled gardens. Yet many feel trapped: 80% say they would immediately increase open-web spend if comparable agentic options were available. As agentic AI becomes the de facto interface to media buying, advertisers enjoy "always-on" optimization but grow dependent on proprietary engines whose inner workings they cannot fully audit or replicate. The more these agents learn from platform-specific data, the harder it becomes to recreate their performance elsewhere.
Advertiser Autonomy, Workflow Friction, and the Open Web Gap
The rise of agentic AI is not frictionless. Taboola’s study finds that integration into existing workflows is the top barrier, especially for large advertisers. Only 9% of mid-sized advertisers cite integration as a challenge, compared with 74% of the biggest spenders who say it is their dominant roadblock. This suggests that as AI ad optimization deepens, scaling it across complex global teams, legacy tools, and multi-channel strategies becomes harder. Meanwhile, marketers are eager to extend AI-driven performance beyond dominant platforms, but the open web lacks comparable, turnkey agentic AI. This imbalance reinforces platform lock-in: to access best-in-class automation, advertisers must accept the rules, data policies, and black-box logic of a few major ecosystems. The strategic tension is growing between short-term performance gains and long-term autonomy over data, bidding strategies, and channel mix.
The Next Frontier: Balancing Performance With Power Dynamics
As AI and agentic systems accelerate programmatic advertising, digital marketing is entering a new phase where economic power flows to whoever owns the optimization engines. Meta’s ambition to deliver “personal superintelligence” underscores that AI is not just a feature; it is the substrate of its commercial model. For advertisers, the key challenge will be to harness AI ad optimization without ceding all strategic control. That may mean demanding more transparency around algorithmic decision-making, experimenting with AI solutions on the open web, and building internal expertise to interpret platform-generated recommendations. The next frontier will be less about whether AI can improve campaign performance—it already is—and more about who ultimately controls the levers of spend, targeting, and measurement. The platforms that define those levers are poised to capture outsized ad revenue growth, unless marketers find ways to rebalance the relationship.