#Japan #AI Technology #Tokyo #PubMatic #AgenticOS

PubMatic Unveils AgenticOS: A Leap Forward in Ad Technology

On January 13, 2026, in Tokyo, the leading digital advertising technology company, PubMatic, announced the launch of their new operating system, AgenticOS. Designed to seamlessly manage autonomous inter-agent advertising across premium digital environments, AgenticOS integrates system-level layers that empower agents to control programmatic advertising in a rapid, consistent manner, from planning to execution and optimization.

Through collaboration with advanced advertisers, agencies, and publishers, PubMatic has actively engaged in testing and developing agent-driven workflows, pushing forward actual market validations and innovations that lead to the next phase of AI-native advertising.

In a world where advertising has become increasingly complex, characterized by new devices, formats, and a multitude of advertisers reaching billions of users globally, traditional programmatic systems face mounting pressure. Thus, it's essential to navigate operational complexities while ensuring consistent and effective decision-making. Addressing this opportunity demands a high-level system capable of real-time detection, learning, and adaptation—providing superior strategic control and clarity for human operators. AgenticOS is tailor-made to meet this challenge, enabling intelligent execution without delays or operational burdens.

Advertisers utilizing AgenticOS can define goals, safety standards, brand safety requirements, and creative parameters within a familiar large language model (LLM) interface. PubMatic’s platform translates these intentions into action via AI-agent groups, continuously planning, executing, and optimizing campaigns within the defined safety standards. This allows teams to focus on high-value strategies, creative development, and performance analysis without being bogged down by manual settings and troubleshooting.

From Agent Innovation to Operating System

The launch of AgenticOS is the result of PubMatic's long-standing investment in agent-based AI and interoperability. From standardizing agent-to-agent communication to optimizing revenue and conducting audience and inventory discovery and planning, as well as troubleshooting deals, a wide array of innovations has been consolidated into AgenticOS. This creates an integrated end-to-end operational environment that allows agent-driven campaigns to be executed and monitored at scale, while continuously improving results.

Unlike point-execution layers or single-protocol connectors, AgenticOS operates on NVIDIA accelerated computing, integrated into PubMatic’s global infrastructure. This foundation supports high-performance, low-latency inference and coordination across millions of ad transactions per second, ensuring reliable execution in real-world markets where performance, scale, and accountability are paramount.

In December 2025, PubMatic collaborated with the independent agency Butler/Till and Geloso Beverage Group's Clubtails to implement early agent-based campaigns on AgenticOS. Based on natural language input from Claude, PubMatic's AgenticOS autonomously proposed campaign strategies, executed buys, and optimized performance in real-time within pre-defined parameters. This automation allowed the platform to handle ongoing execution and optimization, enabling agency teams to concentrate on value-driven strategy, creative development, and effectiveness measurement.

The Three-Layer Architecture of Agentic AI

AgenticOS is grounded in PubMatic’s Architecture of Advertising Intelligence, which is a three-layer framework designed to support responsible, autonomous decision-making at a global scale.

Infrastructure Layer

This layer leverages NVIDIA accelerated computing to enable microsecond-level inference and real-time data integration for tens of millions of auctions per second, achieving decision-making speeds up to five times faster while significantly reducing auction timeouts and ensuring responses in milliseconds.

Application Layer

The agent capabilities integrated into the entire PubMatic platform interpret intentions expressed through protocols like Ad Context Protocol (AdCP) and Model Context Protocol (MCP). These agents automate and fine-tune core advertising functions—such as planning, forecasting, pacing, yield management, troubleshooting, and measurement—working together to continually improve outcomes within defined constraints. Initial tests of agent-driven campaigns report an 87% reduction in setup time and a 70% improvement in issue resolution.

Transaction Layer

This layer translates agent-based AI into real-world execution. The interoperable execution environment connects agent decision-making directly to PubMatic's Activate buying platform, integrating premium supply, signals from live campaigns, and direct bidding infrastructure to execute programmatic guaranteed (PG) and private marketplace (PMP) transactions in real-time. This layer ensures that agent-driven workflows maintain compatibility with existing buying models while enabling more intelligent and adaptive execution in the long term.

Accelerating Agent-Based AI Deployment Across the Ecosystem

To support real-world implementation, PubMatic is launching the Agentic AI Acceleration Program. This intensive initiative aims to help advertisers, agencies, publishers, and partners transition to real agent-driven workflows within weeks of initial tests, focusing on execution, governance, and measurable outcomes, with support ongoing through the first quarter of 2026.

Currently, PubMatic is seeking additional ecosystem partners to contribute to the development, management, and expansion of agent-based workflows across the open internet.

Partner Testimonials

Amanda Grant, Global Head of Data & Technology Partnerships at WPP Media, remarked, “We are using agent-based AI to enhance WPP Open and Open Intelligence. The test of AgenticOS in partnership with PubMatic demonstrates our commitment to advancing agile innovation and scaling autonomous advertising.”

Scott Ensign, Chief Strategy Officer at Butler/Till, expressed excitement about integrating AI to bring speed and agility to programmatic advertising through agent-based media buying, stating that PubMatic is a crucial partner for rapid market entry and testing of emerging technology benefits.

Tom Leone, VP of Media Services at Brkthru, noted that adopting AgenticOS fits naturally within their operations, turning manual settings into real-time optimization, thus improving client service.

John Matthews from Foxtel Media shared his anticipation for the automatic reactions that AgenticOS facilitates, particularly in addressing real audience movements effectively. He believes this agent-based workflow can enhance revenues for live sports and entertainment environments significantly.

Jules Minvielle, Co-founder and CEO of Olyzon, stated that AgenticOS represents a significant step towards ensuring interoperability and scalability across the open internet, changing the dynamics of CTV planning and optimization.

John Goulding from MiQ emphasized how integrating AgenticOS with MiQ Sigma reinforces that the future of programmatic buying lies within the ecosystem as a whole.

Finally, Skyler McGill from Wpromote highlighted that the swift transition from theory to practice in agent-based media buying is astonishing, as campaigns deploy and autonomously optimize in real-time, marking a new frontier in personalization and scalability through collaboration with PubMatic.

For more details about PubMatic’s AgenticOS and participation in upcoming agent-based AI pilot programs, visit PubMatic's website.

About PubMatic

PubMatic is an AI-powered ad tech company that drives digital advertising performance. Through a connected and intelligent platform serving buyers, publishers, data partners, and commerce media networks, it delivers superior performance with heightened transparency, control, and efficiency. Since its founding in 2006, PubMatic has led significant evolutions in programmatic advertising, from early OpenRTB transactions to incorporating AI-driven optimizations and privacy-focused innovations across its platform. With omnichannel scale, proven reliability, and continuous innovation, PubMatic is committed to building a more intelligent, profitable, and sustainable open internet. Built to Connect. Powered to Perform.