When the Restructuring Decision Goes Wrong

The executives making the biggest workforce bets of their careers are working with a judgement gap they have not yet noticed.

The numbers are in for the first half of 2026.

Global technology companies cut nearly 154,000 jobs between January and July. Oracle led with over 21,000 reductions. Amazon, Meta, Microsoft and Cognizant followed. The stated rationale in almost every case pointed toward AI adoption, restructuring, and leaner operating models.

The analyst community is beginning to look more carefully at what actually happened.

Stanislava Savisheva of TradingPlatforms, reviewing the data, drew a distinction that the headlines have missed. Two things are happening simultaneously. Some cuts represent genuine AI repricing. A company assesses what a task costs with AI assistance versus without, and closes the gap. Others represent routine restructuring or fiscal-year adjustments, dressed in AI language because that is the story investors want to hear right now.

The companies are not distinguishing between these two things in their public communications. Neither are the executives driving the decisions.

The decision underneath

What looks like a workforce strategy is, in every case, a judgement call about what AI can actually replace.

That judgement requires deep command of what the affected roles actually produce. It requires knowing where the human contribution is genuinely replicable and where it is not. It requires weighing the second and third order consequences of removal before the decision is made, not after. None of this can be assembled from a briefing document or an AI-generated summary of the options.

This is Domain Mastery applied to the most consequential category of organisational decision. The evidence from the first half of 2026 suggests that in a significant number of cases, it was absent.

Some companies are already rehiring. Others are reporting productivity shortfalls the restructuring plans did not anticipate. The gap between the projected outcome and the actual one is not a communication failure or an implementation problem. It is the consequence of a judgement made without sufficient command of the terrain.

An equilibrium already collapsing

The Collapsing Equilibrium describes a specific structural condition. The information asymmetries that once gave senior experts authority over complex decisions are being eroded. AI tools are distributing analytical capability more widely, now available directly to executive decision makers. The expert counsel that would once have challenged a flawed restructuring assumption is itself under pressure, in many organisations already removed.

The result is a compounding problem. Executives are making decisions of greater consequence at precisely the moment when the internal mechanisms that would have stress-tested those decisions have weakened. The expert whose domain knowledge could have flagged the flaw was not necessarily removed from the organisation. Their standing to be heard had already eroded, long before the decision reached the table.

This is not a hypothesis about what could happen. It is the Collapsing Equilibrium doing exactly what the thesis describes.

What the comparison reveals

Set against this pattern is a contrasting signal.

SAP’s CEO Christian Klein recently took direct personal oversight of the company’s AI platform and autonomous systems capability. Rather than delegating the most consequential technological transformation in the company’s history, he pulled it into his own hands.

That is not a restructuring story. It is a Domain Mastery story. A CEO decided that the judgement required to navigate this transition was too consequential to be mediated by a layer of management, and too complex to be resolved by an AI summary of the options.

The executives who got the restructuring wrong and those who are getting it right are not separated by boldness or caution. They are separated by the quality of their command of the terrain they are operating in.

The open question

The analyst reviewing the 2026 cuts noted that in many cases, AI is one factor among several, alongside cost control, investor expectations, and management restructuring that would have happened regardless. The AI label is doing work that the underlying decision does not always warrant.

That is the diagnostic question the data raises. Not whether AI is disrupting the workforce, which it plainly is. It is whether the executives making these decisions command enough of the terrain to know where the disruption is genuine and where it is cover for something else.

Whether any given executive belongs in that group is a question only the depth of their domain command can answer.

Colin Gautrey, July 2026