X does not work here anymore.

Every morning, I'd pull up yesterday's campaign performance for i-motors - a successful bricks-and-mortar company that had gone all-in on the dot-com boom.  

Sales data streaming in from across the United States. Real purchases. Real revenue. I loved it. 

Publishers pitched us every few days with new platforms, new ideas. When agency.com acquired i-traffic and we became part of an 800-person Omnicom network, it felt like validation. 

We were building the future.

I came back to Ireland to pursue an MBS in eCommerce. Stayed in touch with colleagues. Planned my return. Then the email bounces started.

“X does not work here anymore.”

One by one. i-motors went under. Digital Channel Partners, a digital consultancy in Dublin and others that had grown rapidly began shutting their doors. 

I watched the unraveling on both sides of the Atlantic.

The pattern became clear: when companies couldn't make their revolutionary business models work, they pivoted to advertising. 

"Eyeballs" replaced actual transformation.

Fast forward to October 2024

It’s been incredible to watch the evolution of AI and the transformative effects it's having on our daily lives, especially our workplaces. I’ve been struck by the powerful mission statements of these companies, for example, OpenAI’s mission is "to ensure that artificial general intelligence-AI systems that are generally smarter than humans-benefits all of humanity."

However, given my previous experience of that spectacular bust, I fear there are parallels. The numbers don’t stack up: the new pioneers now find themselves resorting to those old hands of porn and advertising. 

Take for example,  the recent complete reversal on the Open AI’s position on advertising. In 2021, Sam Altman said: "I have a strong dislike for ads… I think advertising is just not a great business model." By early 2025: "I love Instagram ads… they've added value to me… I actively like Instagram ads.”

Now, OpenAI expects to launch ads for ChatGPT's free user base in 2026, projecting roughly $1 billion in ad revenue that year, scaling to nearly $25 billion by 2029. Those internal projections appear to be revealing a new business strategy. 

This, along with the news that ChatGPT would allow adults to generate erotica and access "emotionally responsive companion-style interactions" as well as Sora, a text-to-video generator with no clear professional monetisation path but obvious potential as a content-generation engine for a social network play - we’re starting to see how OpenAI is now planning to see its mission come to fruition. 

The infrastructure numbers tell the parallel story.

McKinsey projects AI infrastructure investment could reach $7 trillion globally by 2030. Actual AI product revenues in 2024: around $100 billion annually, projected to reach $126 billion in 2025.

Even with aggressive 35% annual growth, no plausible adoption curve closes that gap in the relevant timeframe. It’s going to be driven by something else than just people signing up to using AI products. 

As marketing professionals, we recognise this pattern.

I'm not dismissing advertising - it's been my career after all, and not only that, it does work and definitely has its place. Google proved that advertising at scale with sophisticated targeting can be extraordinarily profitable. However, what makes Google a little different is that its founders built advertising into their original business model because search intent made it valuable from the start.

OpenAI is pivoting to advertising onlyafter promising AGI would transform civilisation. When a company shifts from "ads are a last resort" to projecting $25 billion in ad revenue within five years, they're not adding a complementary revenue stream. They're rewriting the business model because the original vision isn't scaling as promised.

I've seen this before. Money flowing into companies based on transformation narratives rather than revenue realities - but then pivots to proven revenue generators when the revolution doesn't materialise. Cue email bounces.

The question isn't whether AI will produce useful tools - it already has. The question is whether trillion-dollar valuations and infrastructure build-outs match the reality emerging in the market.

When OpenAI moves from "we're building AGI that benefits all humanity" to internal projections showing $25 billion in advertising revenue, alongside monetising synthetic intimacy, they're telling you something about the distance between their promises and their path to profitability.

As someone who trafficked those ads in 1999 and watched what came next, I'm paying attention to the signals. 

The email bounces haven't started. Yet….