The zeroth world is forming. Which side is your organisation on?

"Every week there is a new model, a new competitor doing something I have not heard of, and underneath the fatigue is a quieter fear: that a small group somewhere is pulling so far ahead we will never close the gap." I have heard a version of that sentence in three board conversations in the last month. It is usually said quietly, near the end, after the strategy slides are back in the drawer. And here is the part that should reassure you and unsettle you at the same time: the people building the frontier are describing the same divide, out loud, by name.

So let me give you the answer before the worry hardens. The gap that frightens you is real, but it is not opening because of any one tool you missed. It opens when an organisation builds its strategy around the noise of the hype cycle instead of the signal of the steady capability curve. That is the choice that decides which side of the divide claims you. And it is a choice about how you lead the work, not how much you spend.

What did Dario Amodei actually mean by the "zeroth world"?

Asked whether the world is preparing enough for what is coming, his answer was one word. "No. I'll explain the longer version now." The longer version was this. Dario Amodei, in his Wall Street Journal interview at the World Economic Forum in Davos in January 2026, described what he called, as transcribed, "this emerging zeroth world country of like 10 million people", roughly 7 million in the Bay Area and 3 million scattered elsewhere, "forming its own economy" and "becoming decoupled or disconnected" from everyone else.

Sit with the framing for a moment. Not a first world and a developing world. A zeroth world, a tier above the top tier, compounding on its own. He pointed to his own company's revenue to show the slope: from zero to roughly a hundred million across 2023, to roughly a billion across 2024, to roughly ten billion across 2025. Anthropic's own Series F announcement corroborates the shape, a run-rate near one billion dollars at the start of 2025 and over five billion by August. And he described roughly 50% gains inside that small group against around 10% nationally. That is what decoupling looks like in a single chart.

Why is the hype cycle the thing that actually puts you on the wrong side?

Because the hype cycle and the capability curve are two different objects, and most strategy is built on the wrong one. Amodei put it cleanly: "we basically have a Moore's law for intelligence", and "the march has just been constant" over many years. The lurching, the sense that everything changed last Tuesday and nothing changed the Tuesday before, that, in his words, "is a public perception phenomenon". The capability is a smooth line. The noise is what you feel.

When you steer by the noise, you buy the tool that trended this quarter, you run the pilot that demos well, and you mistake motion for progress. When you steer by the curve, you assume capability keeps climbing on a known slope and you build the organisation that can absorb it whatever it is called next year. One is reactive. The other is conscious. The bottleneck is no longer the technology; it is whether your organisation is built to receive it.

And the divide this produces is already in the data, not just the rhetoric. MIT's 2025 study of enterprise AI found that about 95% of generative AI pilots delivered no measurable impact on profit and loss, while about 5% achieved rapid revenue acceleration. The decisive variable was not model quality. It was an integration and learning gap. The 5% had rebuilt how the work flows around the capability. The 95% had bolted a clever tool onto an unchanged process and waited for a result that never arrived.

What does the uncomfortable macro picture mean for my organisation?

Name the possibility he put on the table, because most leaders feel it without having words for it. Amodei described a future that could combine very high growth with very high unemployment and inequality, the possibility of roughly 5% to 10% GDP growth sitting alongside something like 10% unemployment. He was careful to call it a possibility, not a forecast, and he flagged why it unsettles even him: "That's not a combination we've almost ever seen before." Growth and pain have historically moved together. This would pull them apart. The early signs are quieter than the headline fear suggests: in the Federal Reserve Bank of Atlanta's late-2025 survey of corporate executives, firms expected the share of routine clerical roles to fall only modestly through 2026 while skilled technical roles rose, a reshaping of work rather than a sudden collapse of it.

You cannot set national policy from your boardroom. What you can do is decide where your organisation stands when software becomes cheap, abundant and good. Amodei is candid that the shift is already inside his own work: "I don't write any code anymore. I just let Opus do the work and I edit it." When the production of software collapses toward free, the things that hold value are the things software cannot copy. That is the strategic question hiding under the weekly-model fatigue, and it is more about your inner readiness to lead through it than your procurement budget. I have written a companion piece on exactly that inner question, on why almost no one is actually ready, and it pairs with this one.

So how do I move to the right side of the divide this quarter?

Put your attention on what compounds when software is cheap, and lead the work toward it. Walk your team through these in order.

1. Steer by the curve, not the quarter. Assume capability keeps climbing on a steady slope. Build for the level of intelligence arriving in eighteen months, not the model that trended this week.
2. Own your proprietary data. The model is a commodity available to your competitor by Friday. The data only you hold, your operations, your history, your context, is the moat. Map it and protect it.
3. Anchor in the physical world and the relationship. Operations that touch the real world and customer relationships built on trust are the slowest things to copy. Amodei's "Moore's law for intelligence" does not have a Moore's law for trust.
4. Redesign how work flows. The 5% in the MIT data won here. Rebuild the process around the capability rather than bolting the capability onto the process. As Federal Reserve Governor Waller put it, the technology is no longer the hard part; the change management is where the value now sits.
5. Read the real signal. Point your leadership at the Anthropic Economic Index, real anonymised usage data showing where AI augments versus automates, and steer by that rather than by the loudest demo of the month.

The leaders who land on the right side of this divide are not the ones who spent the most. They are the ones who upgraded themselves first, then led the work toward what only their organisation can hold. The zeroth world is forming whether you watch it or not. The question is whether you build the kind of organisation that compounds alongside it, on data, on trust, on the redesign of the work, or the kind that keeps buying this quarter's tool and wondering why the gap never closes.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.