The end of the eyeball economy

For much of my lifetime, the internet has been the foundation upon which an extraordinary amount of economic value has been created.

The commercial internet took shape in the 1990s, ushering in a period of technological and economic transformation unlike anything we’d experienced before. Entire industries emerged, fortunes were made and lost, and businesses that barely existed thirty years ago became some of the most valuable in history.

But for all its sophistication, the internet has largely operated around one fundamental assumption: humans would be the ones using it.

We search, browse, compare, evaluate and transact. And over the past three decades, we’ve built an enormous economic model around that behaviour.

Search advertising, programmatic display, retail media and countless other forms of digital monetisation ultimately depend on capturing and monetising human attention, intent and traffic.

I think we’re approaching the end of that era.

As artificial intelligence becomes increasingly agentic, more of the activity we currently perform online will be delegated to machines. Agents will search for information, evaluate alternatives and execute transactions on our behalf, operating at a speed and scale that humans simply cannot match.

The internet itself isn’t going anywhere. Neither are humans. We’ll continue to communicate, consume content and participate in digital communities.

But I believe an increasing proportion of the internet’s economically valuable activity will be conducted by agents rather than directly by humans.

And that creates a fundamental problem for an internet economy built around monetising human traffic.

For decades, businesses have paid to position themselves in front of human eyeballs. We’ve developed increasingly sophisticated ways of measuring and monetising that attention, whether through impressions, clicks, sponsored placements or marketplace advertising.

But what happens when the human no longer needs to visit the website, view the advertisement or click the sponsored result?

The underlying demand hasn’t disappeared. The economic activity hasn’t disappeared. But the opportunity to monetise the human interaction increasingly has.

The internet could become more economically active than ever while the value of human traffic as its primary monetisation mechanism begins to erode.

And I suspect the economic value will increasingly accrue elsewhere.

Companies like Cloudflare are already experimenting with mechanisms that allow websites to charge AI systems for accessing and using their content. Rather than monetising human attention, the opportunity becomes monetising machine activity itself.

There’s an interesting parallel with blockchain networks such as Ethereum, which could provide some of the underlying infrastructure for machine-to-machine payments.

I’m not convinced we yet know which companies or networks will capture the greatest share of this emerging economy.

But I think the direction is becoming clearer.

For thirty years, much of the commercial internet has been built around the economics of human traffic.

The next thirty may be defined by the economics of agents.