pubX Raises $5M to Build Agent-to-Agent Advertising

On October 6, 2026, an advertising startup most marketers have never heard of announced that the future of media buying will be negotiated by machines talking to machines. According to a company announcement carried by PR Newswire, pubX closed a $5 million Series A led by Chicago Ventures and simultaneously acquired Compliant, a media-quality and governance company, to build what it calls the first infrastructure designed from the ground up for agent-to-agent advertising.

The pitch is strikingly direct. pubX builds the agents that plan, buy, sell and optimize advertising, and it also builds the marketplace those agents transact on. Its buyer agents and seller agents work directly with one another, from the initial campaign brief through to final reporting, across every channel, so that more of every advertising dollar reaches the media and the publisher behind it. The company says governance, transparency and accountability are incorporated into that infrastructure from the start, including checks on purchases and traceable decision-making by the agents.

Governance as a founding principle, not a patch

The Compliant acquisition is the most interesting part of the deal. Compliant was founded by former Unilever and Procter & Gamble executives and built its business on the demand side, delivering media-quality and governance services to global brands and their agencies. pubX, by contrast, began on the publisher side, spending years helping publishers price and sell their inventory. According to Pulse2, the acquisition unites both sides of the advertising transaction inside a single company, and Compliant's quality signals now feed directly into the agents' buying decisions, letting them weigh factors such as data integrity and connected-TV quality before committing spend.

Compliant will keep its name and operate as a pubX company under chief executive Jamie Barnard, who also takes the role of President of Global Demand Growth at pubX, leading expansion across agencies, advertisers and the wider buy side. The combination matters because the trust problem is the central objection to machine-to-machine commerce. Advertisers will not hand budgets to autonomous agents unless every decision is explainable and every purchase is checkable. pubX's answer is to put the governance layer inside the transaction rail rather than bolting it on afterwards, and to let clients set the level of autonomy themselves, keeping approval rights at the stages they care about.

The rails are open, not proprietary

pubX says its infrastructure for agent-to-agent advertising is built on the open Ad Context Protocol, known as AdCP, and is designed to work for buyer and seller agents alike. That open-protocol choice is significant. The emerging agentic adtech market is dominated by vendors adding AI layers onto existing programmatic infrastructure, which mostly means buying through the same pipes with smarter bidding. pubX is attempting something different: a market where software agents negotiate directly with each other on shared protocols, closer to how trading venues work in finance than to how display advertising works today.

The company is already running in the field. As reported by Marketing-Interactive, pubX launched in Australia in September, appointed former Afterpay and Yahoo executive Andrew Gilbert as its ANZ country manager, and positioned itself as one of a small number of companies globally already running fully agent-to-agent advertising campaigns. The $5 million round is earmarked for United States expansion, building out the demand-side business for brands and agencies, developing the buyer and seller agents further, and integrating Compliant's signals across the platform.

Why machines buying from machines matters

Advertising is an ideal first proving ground for agent-to-agent commerce because it is already a fully digital, rule-bound, high-volume market where transactions happen in milliseconds, and agent-to-agent advertising pilots can be measured against decades of programmatic benchmarks. If agents can negotiate media buys with auditable trails and human approval checkpoints, the same patterns can extend to procurement, travel booking, cloud resource purchasing and any other domain where software already spends money on a schedule.

The deeper story is that agent commerce needs financial plumbing that matches its speed. Experimental work is already underway on machine-speed settlement rails, including tokenized deposit trials for AI agent payments that move money at the pace software operates. pubX is attacking the other half of the problem: the negotiation and governance layer that decides how the money moves in the first place. Both halves have to work before agents can transact with each other at scale.

That convergence is drawing sustained investor attention. Enterprise software is racing to embed agents into existing products, but the larger economic question is what happens when agents start buying from other agents without a human in the middle of every transaction. Startups like FIS with its agentic finance suite are already building systems that act on behalf of CFOs rather than merely advising them. pubX extends the same logic to media, where the agent is not an assistant to the media buyer but the media buyer.

pubX chief executive and co-founder Andrew Mole has argued that agentic technology could finish a shift in media buying that programmatic started two decades ago: advertisers describe how they buy, sign off where it matters, and every page of the brief is read every single time. That framing deliberately lowers the drama. The company is not promising to replace media agencies with superintelligence. It is promising to remove the operational friction that makes every campaign a small bureaucracy, by letting agents handle the repetitive negotiation while humans keep the decisions that matter.

The round itself is modest by the standards of the current AI funding boom, and the company is honest about what the money is for: expansion, product development and integration. But the structural claim is large. If agent-to-agent advertising works in media, one of the most fraud-sensitive and measurement-obsessed markets in existence, it will work almost anywhere commerce is digitized. The investors funding pubX are betting that the agents buying ads today will be buying everything else tomorrow.