There is a reasonable argument that the open web is slowly losing relevance.

Budgets continue moving towards platforms offering simplicity, scale and measurable performance. Meta, Google, Amazon and newer creator ecosystems make buying easier than navigating hundreds of independent technology companies.

But I think focusing on whether the open web is “dead” misses something more interesting.

Look at what people are buying.

In August, Nielsen agreed to acquire DoubleVerify for approximately $2.15 billion, explicitly describing the combination as an independent “media intelligence platform”.

Publicis agreed in May to acquire LiveRamp for $2.2 billion enterprise value. Publicis described data collaboration as an important capability for AI and “agentic business transformation”.

Integral Ad Science was taken private by Novacap in December 2025 in a transaction valuing IAS at approximately $1.9 billion.

That is roughly $6.25 billion of transaction value across three companies associated with identity, measurement and media intelligence.

You can explain each deal individually. Perhaps the standalone economics of verification are becoming less attractive. Perhaps public markets no longer reward these businesses appropriately. Perhaps LiveRamp saw consolidation coming and decided scale mattered more than independence.

Each explanation could be true, but it only addresses why the companies might sell. It does not explain why sophisticated buyers want to own them.

Follow the buyer

Publicis provides the clearest example.

Before LiveRamp, it acquired Lotame, giving Publicis additional identity technology, data and more than 1.6 billion IDs. Publicis explicitly connected the acquisition to CoreAI and its ability to connect data with consumers.

Publicis then acquired AdgeAI in March 2026. AdgeAI analyses creative engagement and conversion data, identifies which creative elements work, then converts those signals into recommendations. Publicis described it as predictive measurement designed to improve business outcomes.

So consider the sequence:

Identity → data → measurement → prediction → decisioning → activation.

The result looks less like an advertising company collecting miscellaneous agencies and more like an operating system. Publicis isn't alone in moving in this direction.

Stagwell acquired BERA, whose technology predicts brand performance and connects brand investment with business value. It subsequently acquired UNICEPTA, adding large-scale media monitoring and analytics to its marketing technology platform.

Havas acquired TED Consulting for AI-powered automation, then DMPG for data, analytics and marketing-technology implementation. Havas explicitly positions those capabilities within its wider Converged strategy.

Different companies are following different routes, but the destination looks increasingly similar.

Intelligence is moving closer to the decision

Advertising historically distributed intelligence across the supply chain. Agencies planned. Data companies identified audiences. Verification companies judged inventory. Demand-side platforms optimised bidding. Supply-side platforms understood publishers. Measurement companies determined whether anything worked.

Each company took a piece, creating an enormous AdTech ecosystem and an equally large amount of complexity.

Artificial intelligence changes the economics of that model by making it easier to connect datasets, interrogate information and turn signals into decisions. The value therefore moves away from simply possessing another piece of information towards combining information into a decision. Owning the underlying capabilities becomes more valuable: an agency holding company with identity, customer data, measurement, creative intelligence and execution can increasingly make its own judgement about where a client should spend money, with fewer independent intermediaries telling it what to do.

This may be the real consolidation story

This doesn't mean the open web disappears. Consumers will continue visiting independent publishers. Advertisers will continue wanting those audiences. Large publishers will continue building valuable businesses outside the major platforms.

The more interesting question concerns the technology sitting between advertiser and publisher. The open web can remain large while the independent AdTech layer around it gets considerably smaller.

The companies under pressure may therefore be the intermediaries whose primary value was providing information, measurement or decisioning that somebody else can now incorporate directly, rather than the publishers themselves.

This also explains the investment in agency technology. Holding companies don't need to recreate Meta; they need to recreate enough of the advantages it provides: integrated data, simple decision-making, measurement, optimisation and execution.

In other words, they are building something that behaves more like a walled garden without necessarily owning the media.

The next battle isn't primarily about execution

Much of AdTech still discusses competition through familiar categories.

DSP versus DSP.
SSP versus SSP.
Verification provider versus verification provider.

Those boundaries increasingly obscure where value is moving. The more important competition may be over who supplies the intelligence that determines what gets bought. Once that decision has been made, execution can become comparatively interchangeable, although scale, latency, auction technology, optimisation and financial infrastructure remain difficult and valuable problems.

Infrastructure receives the instruction; the more defensible position may be owning the intelligence that determines it.

That is why the recent M&A activity interests me more than another debate about whether programmatic is growing three or seven per cent.

The market's largest participants appear to be buying the ability to understand, value and ultimately decide.

And when several sophisticated companies independently start buying the same capabilities, it is usually worth paying attention.