INTEL'S CFO CALLED ITS HUGE MANAGEMENT STRUCTURE A “CULTURAL PROBLEM”
Intel did not just have too many managers. According to CFO David Zinsner, it had built an organization where too many people could slow down a decision, too many layers could change the information, and a chip could go through several development cycles before anyone finally said: ship it.
INTEL HAD A MANAGEMENT PROBLEM BEFORE IT HAD A CHIP PROBLEM
Intel's turnaround under CEO Lip-Bu Tan is often described through factories, chip designs, AI competition and manufacturing technology.
But CFO David Zinsner has described another problem that sits underneath all of them:
the way Intel was organized.
At the Deutsche Bank Technology Conference, Zinsner said Intel had reduced its management layers from 12 to 6.
The company had also reduced its number of vice presidents from roughly 450 at its peak to around 200.
Zinsner's argument was unusually direct.
He said many of Intel's problems over the previous decade could be “boiled down to culture.”
The issue was not simply that Intel employed too many people.
It was that too many people had the ability to participate in decisions.
WHEN EVERYONE CAN VETO, NOTHING MOVES FAST
Zinsner described a company in which there was a lot of veto power spread throughout the organization.
That creates a very different economic problem from simply having a large payroll.
Imagine a product decision that should require:
engineering
→ product leader
→ final approval
Now add several additional layers:
engineering
↓
manager
↓
director
↓
senior director
↓
VP
↓
another VP
↓
executive
↓
CEO
Every additional layer creates another place where a decision can be questioned, modified or delayed.
The result is not necessarily one giant mistake.
It is thousands of small delays.
And in semiconductors, delays compound.
THE PRODUCT COULD GO AROUND THE LOOP FOUR OR FIVE TIMES
This is the detail that makes Zinsner's comments particularly interesting.
He said Intel had situations where products went through four or five development spins before being ready, when the goal should have been to get the product right the first time.
That is not merely an engineering problem.
It is a capital problem.
Every additional development cycle can mean:
→ engineers spend more time
→ testing continues
→ manufacturing resources are consumed
→ schedules move
→ customers wait
→ revenue arrives later
→ competitors get more time
A management structure can therefore affect the economics of a semiconductor product without changing a single transistor.
A CHIP COMPANY CANNOT AFFORD SLOW DECISIONS
Intel operates in an industry where product cycles are measured in years, but competitive advantages can disappear much faster.
A chip has to move through:
design
↓
validation
↓
manufacturing
↓
yield improvement
↓
customer qualification
↓
volume production
If a decision is delayed at one stage, the delay can propagate through everything below it.
That makes organizational speed surprisingly important.
A semiconductor company can spend billions developing a technology and still lose economic value because the organization takes too long to turn it into a product.
THE STRANGEST PROBLEM WAS INFORMATION
Zinsner also described another organizational issue: information could change as it moved upward.
The people closest to a project could know that something was going wrong.
But by the time the information reached senior leadership, the situation could appear different.
According to Zinsner's account, this was not necessarily deliberate deception.
Managers simply did not always want to report that a project was failing.
That creates a dangerous corporate feedback loop:
bad news
↓
manager softens it
↓
next manager receives softer version
↓
next manager softens it again
↓
executive receives incomplete picture
↓
decision is made using distorted information
The bigger the hierarchy, the more opportunities there are for information to change.
THIS IS WHY “FLATTER” CAN MEAN MORE THAN CHEAPER
Flattening a company is usually presented as a cost-cutting exercise.
Fewer managers can mean lower salaries.
But Intel's stated objective is broader.
CEO Lip-Bu Tan said in 2025 that organizational complexity and bureaucracy were slowing decisions, while unnecessary silos were contributing to poor execution. Intel said it wanted smaller teams to make decisions faster and reduce the number of layers between employees and leadership.
So the financial equation is not simply:
fewer managers → lower payroll
It is:
fewer layers → fewer handoffs → faster decisions → fewer delays → potentially better execution
The second equation is much more valuable.
THE 450 VP NUMBER IS NOT REALLY ABOUT 450 PEOPLE
A vice president at a company like Intel is not simply an expensive employee.
The title often represents a piece of organizational authority.
When Intel had roughly 450 VPs at its peak, the company had created a large number of senior decision-making positions.
Reducing that number to around 200 changes something deeper:
who gets to say no.
That is why Zinsner focused so much on veto power.
A company can have hundreds of brilliant employees and still move slowly if every major decision requires consensus from too many senior stakeholders.
The bottleneck isn't intelligence.
It is authority.
INTEL IS TRYING TO REDEFINE WHO OWNS THE DECISION
One of the most important organizational changes under Tan has been moving major functions closer to the CEO.
Intel said in 2025 that critical product, manufacturing and general-and-administrative functions that had previously been spread across multiple layers were brought more directly under Tan's leadership.
That sounds like centralization.
And in one sense, it is.
But there is a paradox:
Intel is centralizing leadership while trying to decentralize execution.
The CEO wants fewer layers between himself and the technical organization.
At the same time, smaller teams are supposed to have more freedom to make decisions without waiting for several levels of approval.
The objective is not “the CEO makes every decision.”
It is:
the CEO sees the real situation faster, while teams make more decisions themselves.
THIS IS ALSO AN ACCOUNTING STORY
Organizational complexity eventually appears in financial statements.
Not as a line called “too much bureaucracy.”
It appears indirectly.
For example:
Longer product development
→ more engineering expense
Repeated development cycles
→ additional R&D spending
Delayed launches
→ delayed revenue
Excess management
→ higher operating expense
Slow manufacturing decisions
→ capacity and inventory inefficiencies
Poor information flow
→ worse capital allocation
This is why CFOs care about organizational design.
A management structure may look like an HR issue.
Eventually it becomes a margin issue.
INTEL HAS ALREADY BEEN CUTTING COSTS
The organizational changes are part of a much larger restructuring.
Intel had 108,900 employees at the end of 2024 and around 88,400 by late September 2025, according to reporting on the restructuring.
But there is an interesting detail.
In Oregon, only about 8% of the employees laid off had “manager” in their job titles, according to the same report.
Most were engineers and technicians.
That means “flattening the organization” should not be interpreted as simply firing managers.
The broader program has involved reducing the workforce while changing how the remaining organization operates.
The goal is not merely to make the org chart smaller.
It is to make the organization behave differently.
THE REAL COST OF A MANAGER IS NOT ALWAYS THE SALARY
Suppose a manager costs a company €200,000 per year.
The obvious cost is €200,000.
But imagine that manager is involved in approving 100 projects.
If every approval takes another week, the economic cost could be much larger than the salary.
The manager becomes a decision bottleneck.
That is why corporate restructuring can be economically valuable even when the direct payroll savings are relatively modest.
The important question becomes:
How much time does this layer add to the business?
STARTUPS HAVE AN UNFAIR ADVANTAGE HERE
Zinsner made an especially striking comparison.
He said startups with around 50 people were sometimes getting products out faster than Intel.
Obviously, a 50-person startup and Intel do not have the same technical or operational complexity.
Intel manufactures sophisticated processors at enormous scale.
A startup might only have one product.
But that is exactly what makes the comparison useful.
A small company can operate like:
engineer
↓
founder
↓
decision
A giant company can operate like:
engineer
↓
manager
↓
director
↓
VP
↓
senior executive
↓
committee
↓
another review
↓
decision
Scale creates capabilities.
It can also create friction.
INTEL IS TRYING TO BUY BACK SPEED
The restructuring can therefore be understood as an attempt to recover something Intel once had:
organizational velocity.
The company cannot become a 50-person startup.
It still has fabs, customers, manufacturing operations, engineers, suppliers and thousands of employees.
So the objective is not to eliminate complexity completely.
It is to make complexity invisible to the people who need to move quickly.
That means:
fewer approvals
↓
clearer ownership
↓
smaller teams
↓
faster decisions
↓
fewer repeated cycles
THE TEST IS ALREADY MOVING TO 18A AND 14A
This is where the organizational story meets Intel's technology roadmap.
Zinsner said Intel's 18A yields were ahead of internal milestones, while progress on 14A defect density was also tracking better than its target curve. Intel plans risk production for 14A in 2027 and high-volume manufacturing in 2028.
That does not prove that the management restructuring caused those results.
But it creates a measurable test.
If Intel's new organizational structure is supposed to improve execution, future product development should show fewer delays, fewer repeated cycles and clearer accountability.
In other words:
the org chart eventually has to produce a better chip business.
THE CFO IS USING “CULTURE” AS AN OPERATING VARIABLE
“Culture” can sound like a soft corporate word.
Zinsner is using it differently.
For Intel, culture means things employees actually do:
Who can block a project?
Who reports bad news?
Who owns the decision?
How many times does a product get redesigned?
How quickly does information reach leadership?
How much authority does an engineer have?
Those are operational questions.
And operational questions eventually become financial questions.
THERE IS A RISK TO FLATTENING TOO MUCH
There is another side to the experiment.
Removing management layers can make an organization faster.
But management layers also exist for reasons.
Large semiconductor companies need coordination across:
engineering
manufacturing
quality
supply chain
sales
finance
legal
customers
Too little coordination can create a different kind of failure.
So the challenge is not simply:
more managers = bad
or
fewer managers = good
The real question is:
where should decision-making live?
Intel is attempting to move that boundary.
THE NEW ORGANIZATION HAS A DIFFERENT ECONOMIC LOGIC
The old structure can be imagined as:
more control
→ more reviews
→ more consensus
→ fewer individual decisions
→ slower execution
The new model is supposed to become:
clearer ownership
→ smaller teams
→ fewer veto points
→ faster decisions
→ more accountability
That final word matters.
When fewer people participate in a decision, the person responsible for the outcome becomes easier to identify.
Speed and accountability can therefore reinforce each other.
THIS IS WHY THE STORY IS BIGGER THAN INTEL
Many large companies eventually develop the same problem.
Success creates growth.
Growth creates departments.
Departments create managers.
Managers create processes.
Processes create committees.
Committees create more processes.
Eventually, the company spends more time coordinating work than doing it.
This is one of the strangest costs of becoming large:
the organization starts producing administration about the work instead of the work itself.
Intel's restructuring is an unusually visible example because the consequences can be measured in physical products.
A delayed meeting is abstract.
A delayed chip launch is not.
MAACAT PERSPECTIVE
Intel's CFO did not describe the company's old management structure as merely expensive.
He described it as a cultural problem because the structure changed how decisions were made.
About 450 vice presidents at the peak became roughly 200.
Twelve management layers became six.
And according to Zinsner, fewer people now have the ability to stop a decision simply because they sit somewhere in the chain of command.
The business logic is simple:
too many decision-makers
↓
too many veto points
↓
slower product cycles
↓
more development iterations
↓
higher cost
↓
later revenue
Intel is trying to reverse the chain:
fewer layers
↓
clearer ownership
↓
faster engineering decisions
↓
fewer repeated cycles
↓
better execution
The interesting part is that Intel's turnaround is therefore not only about making better chips.
It is also about making the company capable of deciding what to build before the market moves on.
