There is a comparison that the artificial intelligence industry had kept out of the public conversation until now. Satya Nadella brought it up this Sunday in a post on X that garnered over a thousand responses in just a few hours. The metaphor he used is "industrial offshoring".
Just as the first wave of globalization hollowed out industrial economies, wiping out factory jobs and decades of competitive advantage with consequences we still feel today, artificial intelligence threatens to do the same to corporate knowledge.
The mechanism Nadella describes is concrete. If an organization hands over its workflows, its domain knowledge, and the accumulated judgment of its teams to external AI models, those models absorb it.
What was once a unique advantage, now could become a generic capability available to everyone.
There are no layoffs or plant closures. The hollowing out happens silently, in every usage cycle, in every operation the model records and leverages. Where there was once exclusive know-how, there is now a standard resource.
"You can outsource a task, or even a job. But you can never outsource the learning."
This warning is not directed at employees but at executives. The risk Nadella identifies is not the loss of an individual position, but the erosion of the organisation's collective intellectual property, its processes, and the judgment a team spends years building.
To name this, he introduces a term that was not in the business management vocabulary until now: "Token Capital".
This represents the layer of agentic capability that a firm builds and owns when it connects its real workflows with the AI models it uses. It is not software or a database. It is a system that learns with every use and improves with every cycle.
"I think of it as a machine climbing a hill," Nadella wrote. "And unlike most assets, it compounds." Every optimised process generates better training signals, which accelerates the accumulation of the firm's own tacit knowledge.
The argument is that this learning cycle, and not just isolated data or talent, is the company's new intellectual property. Whoever builds it first will have a difficult advantage to replicate, regardless of what new model arrives on the market. And whoever gives it away, never gets it back.
We have seen this movie before. During the peak years of manufacturing offshoring, macroeconomic indicators in industrialized economies looked good. Companies that outsourced production showed cleaner margins. The system seemed to work on paper.
What did not appear on the spreadsheets was the structural cost: economies lost their technical know-how, supplier networks, and decades of accumulated memory. The hollowing out was progressive but became visible years later, when it was already difficult to reverse.
Nadella applies the same logic to corporate knowledge. Quarterly metrics may look healthy while the company cedes its intellectual capital to external models. The problem arrives when that knowledge is no longer exclusive.
What makes this diagnosis notable is who is making it. Nadella is the CEO of Microsoft, the company that sells the most AI infrastructure as a service globally. Azure hosts AI models from other providers. Microsoft 365 Copilot automates workflows for organizations around the world. The alliance with OpenAI, renegotiated in April, remains active.
The same Company who Microsoft cut its engineers off from AI because the bill got too big — why AI might not take your job after all
Nadella describes the risk of ceding corporate knowledge to external models and proposes building a proprietary learning layer: each firm's token capital. Naturally, the infrastructure to build that is something Microsoft also sells.
There is a deeper argument in the post that goes beyond business and into political economy.
"There is no social permission for an AI future that hollows out entire industries," he suggests. If all value ends up concentrated in a few models that absorb the knowledge of everyone else, the political system will not tolerate it.
The first line of the post summarized it clearly:
"A frontier without an ecosystem is not stable."
The comparison to manufacturing is not just rhetoric. What happened during the first globalization was not only an economic problem but a redistribution of power that took decades to become visible and still has no clear solution. Nadella applies that diagnosis to corporate cognition. And he does it as the executive who stands to gain the most if organizations decide to build that proprietary layer on his cloud, using his tools.
He might be right about the diagnosis, but it does not change the fact that he is also the one selling the remedy.