Walmart Gave the Internet Its Own CEO. It Took Sixteen Years to Undo That.

In the last column I told the story of IBM and its “e-business” campaign: half a billion dollars to convince the world that the internet needed its own department. But IBM wasn’t the only one that fell into that logic. Walmart did something even more extreme, and their story is a better warning for anyone building an “AI office” today.

In 2000, Walmart decided that the internet was so different from its store business that it needed to live outside the company. It spun Walmart.com off as a separate company, with its own CEO, its own board, and a Palo Alto venture capital partner, Accel Partners. The executive who took charge described the approach in that year’s annual report: Walmart was treating the internet “like a foreign country,” with completely different rules of commerce from those of a physical store.

The experiment was, in the company’s own words at the time, a disappointment. Not because selling online was a bad idea. It was a disappointment because separating the internet from the rest of the business guaranteed it would never truly integrate with the logistics, inventory, and store network that were, in reality, Walmart’s biggest competitive advantage. It took the company sixteen years — and the acquisition of Jet.com in 2016 — to finally undo that separation and build, only then, an e-commerce operation that worked integrated with the rest of the business.

Sixteen years of a separate department to end up doing what could have been done from day one: treating the internet as part of the business, not as an annex with its own zip code.

Now let’s translate this to 2026, almost word for word. “E-business” became “AI-powered.” The independent web division became the AI Lab or the AI Center of Excellence. The “Head of Internet” title became “Chief AI Officer.” The parallel is so precise it stops feeling like coincidence.

I’m not saying that naming someone in charge of AI is a mistake in itself. What Walmart’s story teaches is something else: separating the technology from the rest of the business — as if it were a separate initiative with its own budget and territory — delays precisely the integration that ends up being the only real source of value. Walmart didn’t lose those sixteen years from lack of investment. It lost them because of structure. While the internet lived in a satellite company with its own board, the rest of the organization kept operating as if the internet didn’t exist.

The question then isn’t whether your “AI office” is going to dissolve. It will dissolve, sooner or later, the same way every “e-business team” and every internet spin-off dissolved twenty-five years ago. The real question is whether you decide when and how it dissolves — integrating AI inside every department now — or whether you let the market impose it on you after years of operating at two different speeds, the way it happened to Walmart.

Does your company today have an AI that lives “like a foreign country” apart from the rest of the business? Or have you already started bringing it home?