Every decade or so, a technical standard emerges that quietly rewires who holds power in advertising. OpenRTB did it in 2010. The TCF did it for consent. Now, two competing frameworks for agentic AI advertising are fighting for the same prize — and the outcome will shape whether the open web gets more open or less.
This is not an abstract standards nerd debate. This is about who controls the transaction layer of the next era of advertising.
The Contenders
On one side: the Ad Context Protocol (AdCP), a ground-up open-source framework built on Anthropic's Model Context Protocol. AdCP proposes a world where AI agents can autonomously discover inventory, negotiate deals, and execute campaigns — talking directly to publishers without the usual cast of intermediaries. It is ambitious, technically elegant, and explicitly designed to disintermediate the existing programmatic stack.
On the other: the IAB Tech Lab's agentic roadmap — the Agentic RTB Framework (ARTF), User Context Protocol (UCP), and umbrella initiative AAMP. The IAB's argument is essentially pragmatic: OpenRTB already works at scale, is already compliant (or capable of compliance), and is already implemented by thousands of platforms. Layering AI orchestration on top of it is faster and safer than fragmenting the ecosystem with a parallel protocol built by people who have, until recently, been in the business of building large language models rather than ad servers.
Both Digiday and AdExchanger have described this as potentially as consequential as the original launch of OpenRTB. That framing is not hyperbole.
The Case for AdCP
AdCP's appeal is structural. If you believe — as many do — that the programmatic supply chain is irreparably bloated, opaque, and extractive, then incremental reform feels like rearranging deckchairs. AdCP offers something different: a clean break. Publishers expose their inventory directly via machine-readable context. AI agents discover it, evaluate it, and transact against it. No exchange. No SSP taking 30% for unclear reasons. No mystery delta between what the buyer paid and what the publisher received.
For independent publishers who have watched their programmatic yields erode for a decade, that proposition has genuine appeal. A world where an AI media buyer can discover and purchase a contextually relevant placement on a niche editorial site as easily as it can on a major platform is a world with a more level playing field.
The challenge is that AdCP requires the ecosystem to meet it. Publishers need to implement the protocol. Buyers need to build agents that use it. And someone needs to handle the messy realities that OpenRTB has spent fifteen years solving: fraud detection, brand safety, frequency capping, consent signal passing, billing reconciliation. The protocol's elegance is also its vulnerability.
The Case for AAMP
The IAB Tech Lab's counter-argument has the force of incumbency behind it, which is both its strength and its problem. ARTF and UCP build on infrastructure that already exists, already processes billions of auctions a day, and already has compliance frameworks bolted on. For most buyers and sellers, the path of least resistance runs through AAMP.
But the history of ad tech standards is not reassuring here. Standards developed under the auspices of industry bodies are supposed to serve the whole ecosystem. In practice, they tend to reflect the engineering priorities and commercial interests of the largest players in the room. Smaller publishers, independent ad servers, and mid-market agencies have repeatedly found that 'open' standards are open in theory but demanding in practice - requiring implementation resources they simply do not have, while benefiting platforms that can absorb those costs at scale.
The question worth asking of AAMP is not whether it will work technically. It probably will. The question is whether it will work for everyone - or whether it will work primarily for the entities that already dominate OpenRTB.
The European Variable Everyone Is Underweighting
There is a dimension to this debate that the US-centric trade press keeps treating as a footnote: GDPR and the Transparency and Consent Framework.
Any agentic advertising protocol that processes user data — and both AdCP and UCP do, in different ways — must handle consent signals compliantly in European markets. That is not optional. It is not a detail to be sorted out in a future working group. It is a legal requirement that has already resulted in significant regulatory action against programmatic infrastructure that failed to get it right.
AdCP's current specification does not provide a clear, auditable mechanism for TCF consent signal passing. The IAB Tech Lab's approach at least starts from a framework that has existing consent architecture — though UCP's handling of user context data will require careful scrutiny as the specification matures. European publishers and vendors cannot afford to run experiments with consent compliance. The regulatory risk is asymmetric and very real.
This is not a reason to dismiss either protocol. It is a reason to insist that consent handling is treated as a first-class requirement, not an afterthought, by whichever working groups are finalising these specifications. European ad tech players — and there are significant ones - need a seat at that table, not a retrofit six months after adoption begins.
What This Means for Publishers and Independent Vendors
At Adnuntius, we sit in an instructive position in this debate. We are an independent ad server and marketplace with no stake in preserving the opacity of the existing programmatic stack - our business model is built on transparency and direct buyer-seller relationships, not on margin extraction in a black box. We also serve publishers who lack the engineering bandwidth to implement multiple competing protocols on an experimental timeline.
After careful evaluation, we have made our choice: we are building on AdCP.
That decision was not taken lightly. It carries real implementation risk, and we are not naive about the scaling challenges that come with a new protocol that has yet to achieve critical mass. But we believe those challenges are solvable — and more importantly, we believe they are worth solving.
The core reason is structural. AAMP builds on OpenRTB, and OpenRTB is not a neutral foundation. It is infrastructure that has, over fifteen years, concentrated margin and data in the hands of a small number of intermediaries while steadily eroding publisher yields. Layering AI orchestration on top of that architecture may produce a faster, smarter version of the same system — but it does not change what the system fundamentally does. We are not interested in making the existing supply chain more efficient. We are interested in replacing the parts of it that do not serve publishers or buyers.
AdCP offers a genuine alternative: a protocol designed from the ground up for direct, transparent, machine-readable transactions between buyers and sellers. That aligns precisely with how we have always believed advertising should work.
Our Answer to the Scaling Challenge
The most legitimate criticism of AdCP is reach. An AI buyer using the protocol can only access inventory from publishers that have implemented it — and for a small or mid-sized publisher, that implementation cost is a real barrier.
This is where our position as both an SSP and a DSP becomes directly relevant. Because Adnuntius operates a marketplace - a single system serving both the buy side and the sell side - we can absorb that implementation cost on behalf of our publishers. A buyer sending an AdCP request to the Adnuntius marketplace does not need each publisher to have individually implemented the protocol. They make a single request and receive responses back from across our publisher network. The marketplace acts as the AdCP-compatible layer, and publishers participate through it without needing to build anything new.
In practice, this means the scaling challenge that makes AdCP risky for an individual publisher becomes manageable through a marketplace model. Smaller and independent publishers - precisely the ones who stand to gain most from a more transparent supply chain - are not priced out of the agentic era before it begins.
Our Position
We are committed to AdCP, and we are building the infrastructure to make it work for publishers who cannot wait for a standards war to resolve itself. Our implementation will be available to publishers on our platform, and we will contribute what we learn back to the open-source development of the protocol.
We remain engaged with the IAB Tech Lab's agentic roadmap - not because we are hedging, but because the GDPR and TCF compatibility questions that both frameworks must answer are best addressed through industry-wide dialogue rather than in isolation. European publishers and vendors need compliant infrastructure from day one, not a retrofit. We will continue to advocate for that in every working group we participate in.
The agentic era of advertising is coming. Our bet is that it should arrive as a genuinely better system; not as the same supply chain with a chatbot bolted on the front. AdCP, built on a foundation of transparency and direct relationships, is the architecture most likely to get us there. We are building accordingly.