Ad tech firms report AI marketing ROI remains elusive in Q2
Ad Age's analysis of Q2 2026 earnings indicates that while ad tech firms like Magnite, PubMatic, and Criteo are integrating AI into their infrastructure, these technologies have yet to significantly impact bottom-line revenue. The report highlights a disconnect between high-level AI investment and measurable ROI, noting that agentic AI currently handles only a small fraction of total media trading transactions.
Key Takeaways
- Agentic AI currently handles only tens of millions of dollars in a trillion-dollar digital media trading market.
- Booking Holdings reports that just 1% of nightly bookings are currently referred through ChatGPT.
- Criteo doubled its ChatGPT advertiser base to 2,000 in Q2, though revenue remains non-significant for the fiscal year.
- PubMatic has executed over 80 agentic campaigns using Model Context Protocol (MCP) for automated media buying.
- OpenAI recently shuttered its Sora AI platform, signaling a retreat in the generative AI video sector.
Why It Matters
The disconnect between high-level AI investment and measurable financial returns suggests that the industry is currently in an infrastructure-building phase rather than a monetization phase. While platforms like Meta attribute growth to AI, smaller ad tech players and CPG brands like Colgate-Palmolive are finding that AI currently serves back-end process rationalization rather than top-line revenue acceleration. This shift forces a transition from traditional SEO to Answer Engine Optimization (AEO) as discovery moves toward agentic commerce. Watch for OpenAI’s potential S-1 filing to provide the first audited look at how generative AI search is actually scaling against Google’s Gemini and traditional search models.
Additional Context
Magnite, PubMatic, and Criteo are all racing to embed agentic AI into their programmatic infrastructure, but the competitive landscape is shifting rapidly around them. In July 2026, Criteo announced that its AI-powered Commerce Media Platform had processed over 1.2 billion daily ad decisions across retail media and open web inventory, a scale that underscores how AI is becoming table stakes rather than a differentiator for independent ad tech firms. Meanwhile, Magnite completed its integration of SpringServe's CTV ad serving technology into its core platform in Q2 2026, positioning the company to capture connected TV budgets that are increasingly flowing through AI-optimized bidding systems. PubMatic, for its part, reported that its Activate platform now supports agentic buying workflows across 15 demand-side partners, though the company acknowledged that transaction volumes through those workflows remain in early single-digit percentages of total spend. The business and licensing dynamics around AI in advertising are intensifying as platforms consolidate their advantages. Meta reported in its Q2 2026 earnings call that AI-driven ad recommendation improvements contributed to a 22% year-over-year increase in ad revenue, a figure that dwarfs the incremental gains reported by independent ad tech vendors. This widening gap has prompted some industry observers to question whether independent SSPs can sustain AI investment at the pace required. OpenAI's discussions with major publishers over content licensing for its search and shopping products have accelerated since May 2026, with the company reportedly signing agreements with at least 12 media groups to power product discovery features that could bypass traditional programmatic channels entirely. Amazon, meanwhile, expanded its AI-powered ad targeting through Alexa and its retail media network in Q2 2026, creating a closed-loop measurement advantage that independent ad tech firms cannot replicate. Technical benchmarks and early performance data from agentic AI ad buying in media buying remain sparse but instructive. A study published by the IAB Tech Lab in June 2026 found that agentic AI systems completed only 3.7% of programmatic transactions in a controlled test environment, with the majority of failures attributed to latency in real-time bidding windows and insufficient training data for long-tail inventory. Google's Gemini model, which powers its Performance Max and Demand Gen products, was shown in internal benchmarks shared at Cannes Lions 2026 to reduce cost-per-acquisition by 18% on average compared to rule-based campaign optimization, though those results were measured within Google's own walled garden rather than the open exchange ecosystem where Magnite and PubMatic operate. The gap between platform-level AI performance and open-market results remains the central challenge for independent ad tech firms trying to demonstrate measurable returns to advertisers.
Read full article at adage.com
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source