Amazon has cut roughly 14,000 corporate management roles this year. Oracle has cut somewhere between 20,000 and 30,000, close to 18% of its workforce, concentrated in middle management. Korn Ferry’s global workforce data shows 41% of employees worldwide, 44% in the US, now say their organisation has eliminated a management layer entirely. Gartner is reportedly predicting that a fifth of organisations will use AI to remove more than half their current middle-management positions by the end of this year.
That is a real trend, backed by real numbers, and it is rational for the businesses doing it. Large enterprises spent two decades building management layers to coordinate work at a scale no individual could hold in their head. AI has now given them a socially acceptable reason to admit some of that structure was excess, and to cut it.
None of that is your problem. If you run a founder-led business with 15 to 150 people, your problem is the opposite one, and the flattening headlines are giving you a very convenient reason to keep ignoring it.
A trend that doesn’t transfer the way it looks like it should
It is worth conceding the obvious objection up front: a business this size rarely has enterprise-style management bloat to begin with. There was no army of middle managers at your company to cut, so on the surface the comparison looks like a stretch.
That is exactly the point. The flattening story isn’t dangerous to founder-led businesses because they share the disease. It is dangerous because they are exposed to the same proposed cure, fewer layers, leaner structure, more direct reporting, without ever having had the layer that cure is meant to remove. Amazon and Oracle had a middle to cut. Most businesses in the 15-to-150 range never built one.
Look at what actually happens inside a business that size as it grows. The founder hires functional heads, a head of sales, an operations lead, someone running delivery, and those people report straight up to the founder. There is no executive layer sitting between them translating day-to-day functional decisions into coordinated business decisions. The founder is still personally approving pricing exceptions, hiring calls, vendor terms and prioritisation trade-offs across every function, because no one else in the business has been given the mandate to own those decisions.
That is not lean. It is under-managed at exactly the level where a business this size should have built real management capacity, and it has nothing to do with how many managers Amazon just laid off.
Picture a composite, illustrative example rather than a named business: an operations-heavy company of around 80 people, three or four strong functional leads, and a founder who still signs off on every pricing exception and every hire above entry level. Growth has stalled not because the market shifted, but because every decision above a certain size has to wait in the founder’s queue. Read against the flattening headlines, that founder could easily conclude the business is already appropriately lean. Read correctly, it is missing the one layer that would let those functional leads actually run their areas.
Cutting a layer you never built isn’t possible
Some of the commentary around this trend treats AI as the clean explanation for all of it. It isn’t quite that simple, and it’s worth saying so: several analysts argue a meaningful share of these cuts is ordinary cost discipline relabelled as AI strategy, not AI genuinely removing coordination work. Whatever the true split, the effect on the reader is the same either way. Enterprise businesses are removing a layer. Founder-led businesses are being told, implicitly, that fewer layers is the goal, when what most of them actually need is the layer they skipped building the first time.
The risk isn’t abstract. A founder who reads flattening is progress as a general truth has a new, very reasonable-sounding justification for continuing to avoid the executive hires, the delegated decision rights, and the management cadence the business needs to grow past its current ceiling. Forbes has already flagged the downstream version of this problem at the enterprise end: aggressive management cuts today risk producing a shortage of trained, decision-ready leaders tomorrow, because you cannot build executive judgement in people who were never given real decisions to make. Founder-led businesses that use the flattening narrative to justify never building that layer in the first place are choosing to start from that shortage, not to avoid it.
What to build instead of what to cut
None of this is an argument against efficiency, and it isn’t a case for adding headcount for its own sake. It’s a case for precision about which structural problem you actually have. A business that has genuinely over-built its middle should simplify it. A business that never built an executive layer at all needs to build one, and no amount of staying lean changes that.
Building it looks less like hiring a layer of new managers and more like deliberately handing specific, named decisions to specific people, and then holding them to a cadence that lets the founder stop being the default approver for everything. It is a structural investment, not a headcount increase, and it is the opposite move to what the flattening headlines are celebrating.
Two questions worth asking this week
Before drawing any lesson from what Amazon or Oracle just did, ask two questions about your own business.
First: name the last five meaningful decisions made in your business this month. How many of them still went through you personally, even in functions where someone else holds the title?
Second: if you removed yourself from the business for four weeks, which specific decisions would simply stop being made, because no one else currently has the mandate to make them?
If the answers point to you, the fix is not fewer layers. It’s building the one you don’t have. Blaque Software works with founder-led businesses on exactly that shift, from a founder who owns every decision to an executive layer that can actually carry them. If those two questions are uncomfortable to answer, that’s the conversation worth having next.