PubMatic

Agentic Advertising Rebuilds, Not Removes, Intermediaries

This AdExchanger opinion piece (The Sell Sider) argues that 'agentic advertising'—the use of AI agents to plan, package and execute media buys—risks recreating the same intermediary complexity publishers experienced with SSP proliferation. The article notes publishers historically integrate dozens of SSPs to chase incremental demand and warns buyer and seller agents will encode incentives that favor favored supply paths, potentially biasing spend. It recommends a disciplined publisher strategy: define a primary sales agent, build a small set of specialized agents, standardize payment/clearing, and participate selectively in external ecosystems. The piece highlights that settlement and payment control (clearinghouses) may determine influence more than the agent implementations themselves.

PubMatic’s AgenticOS Doubles Reach Per Dollar

Level Agency and PubMatic published a head-to-head case study where PubMatic’s AgenticOS outperformed the agency’s incumbent DSP across multiple metrics for an education-sector campaign. Running simultaneously with identical budgets and objectives, AgenticOS delivered more than 2x reach per dollar for qualified audiences (351 impressions per $1 vs 164), cut weighted CTV CPMs by 53%, and improved video completion rates by 54% overall. The test attributed gains to PubMatic’s agentic AI operating inside the supply chain via its Activate supply activation platform, faster audience building and retargeting (days vs. typical 30–60 day ramps), and an integrated stack including publisher-side signals and owned GPU infrastructure. Level Agency expanded open-internet spend mid-flight based on the results.

Containerization Explained: Impact on Ad Tech

This Digiday explainer defines containerization in ad tech as deploying self-contained bidding logic and required data inside exchange or SSP infrastructure rather than running it remotely. Proponents say moving decisioning closer to the auction reduces latency, cloud egress costs and signal loss; critics call some of the hype a rebranding of older solutions. Large, co‑located DSPs/SSPs see limited performance gains, while smaller algorithmic specialists and signal providers benefit most. Examples include PubMatic’s Decision Fabric, Index Exchange’s Index Cloud and startup SWYM, which is containerized across multiple SSPs (Index Exchange, Magnite, Media.net, OpenX, PubMatic) with others in process. The piece highlights unresolved infrastructure gaps (global pacing, budgeting, frequency caps, real‑time sync), the potential for consolidation and opportunistic middlemen, and frames containerization as a prerequisite for advanced outcome‑based bidding and other AI-driven adtech capabilities.