Observed Signal · Oct 14, 2025 · Partnership · Source: ExchangeWire · Impact: 4/5 · Sentiment: Positive
PubMatic and NVIDIA Revolutionize AdTech with Lightning-Fast AI
PubMatic announced the results of a multi-year integration with NVIDIA to accelerate AI processing for open internet programmatic advertising. The collaboration delivers AI processing up to five times faster than traditional systems and reduces inference latency to about 1 millisecond, cutting auction timeouts by 85% and boosting ROI for publishers and advertisers. The initiative utilizes NVIDIA L40S GPUs for real-time decisioning across connected TV, Triton Inference Server for serving ad requests to DSPs, and RAPIDS Accelerator for Apache Spark for faster analytics. PubMatic’s AI-enabled cloud infrastructure provides real-time insights, twice-as-fast live campaign processing, and a 30% reduction in energy consumption. In 2025, PubMatic launched AI-powered platforms for buyers and publishers with generative and agentic AI, including an open spec for agent-to-agent communication and generative AI deal management that reduces deal creation time by 87%.
Significant AI acceleration and platform enhancements via NVIDIA integration; substantial improvements in latency, efficiency, and new AI-powered capabilities affecting the programmatic advertising ecosystem.
Track PubMatic Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- PubMatic and NVIDIA completed a multi-year integration delivering AI processing up to five times faster than traditional systems for open internet programmatic advertising.
- NVIDIA L40S GPUs power real-time ad decisioning and streaming across connected TV, cutting latency to 1 millisecond and boosting ROI for publishers and advertisers.
- NVIDIA Triton Inference Server serves the right ad requests daily to DSPs via traffic shaping for higher DSP spend and more efficient integrations.
- RAPIDS Accelerator for Apache Spark enables faster analytics, delivering quicker insights at a lower cost.
- In 2025 PubMatic launched AI-powered platforms for buyers and publishers with generative and agentic AI capabilities, including an open spec for agent-to-agent communication and deal management that reduced deal creation time by 87%.
Connected Companies & Entities
2 Entities mappedRelated Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
PubMatic Deploys AI Agents to Simplify Programmatic Buying
PubMatic is deploying a network of AI agents, via its AgenticOS infrastructure (built with NVIDIA) and its Activate buying platform, to automate planning, execution and optimization of programmatic media campaigns. The rollout is part of a partnership with Untapped Growth, a collective of 20 independent agencies, which allows agencies to input campaign preferences and access negotiated commercial terms. PubMatic says AgenticOS replaces manual DSP/SSP orchestration with specialized bots for tasks such as audience discovery and execution; early tests reportedly showed an 82% reduction in buy-side costs and nearly a 90% reduction in campaign setup time for one agency. Two Untapped Growth agencies are expected to go live shortly, while others are testing the system and exploring direct LLM-driven execution (e.g., via Claude) that can bypass traditional DSP layers.
PubMatic Unveils Fast-Track OS for Agent-Based Advertising
PubMatic has launched an "Agentic OS," a Nvidia-accelerated operating system for agent-based programmatic advertising that aims to enable "Agent-to-Agent" workflows where specialised AI agents plan, execute, optimise and monitor campaigns autonomously. Early partners including WPP Media and MIQ will test agentic campaigns under real market conditions. The three-layer architecture (infrastructure, application/agentic functions, transaction layer) claims up to fivefold faster decisioning, sub-millisecond reaction times and fewer auction timeouts. Agentic functions operate via protocols such as the Ad Context Protocol (AdCP) and the Model Context Protocol (MCP) and automate planning, pacing, optimization, troubleshooting and measurement. PubMatic integrates Agentic OS with existing programmatic buying models (Programmatic Guaranteed, Private Marketplaces) and routes media via its Activate buying platform. PubMatic positions the approach as particularly relevant for complex environments like CTV, while noting adoption will depend on control, transparency and scalable governance in practice.
AI Insights Revolutionize Ad Tech: PubMatic's Vision Unveiled
Punit Shah, Director of Product Marketing at PubMatic, discusses the company’s focus on AI-driven analytics and agentic workflows in advertising infrastructure. Shah describes PubMatic’s recent AI Insights launch, which restructures signals to enable reasoning across data, real-time benchmarking, demand-shift detection and surfacing next-best actions inside workflows. He outlines how agentic AI can improve decision quality for publishers and buyers, gives an election-cycle example where AI benchmarking aids pricing and packaging, and predicts CTV will evolve into performance-grade programmatic media by 2026 through AI optimization, outcome-based evaluation, and intelligent curation. Shah advises publishers to build structured data foundations, operationalize transparency, and adopt AI with intent. The interview frames AI as architectural — reducing reporting friction, amplifying intelligence, and reshaping market design when built on governed, structured data.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.