The same interview link circulated through founder group chats for days: Jensen Huang still receives work updates directly from employees across the company. Had an unknown SME executive said the same thing, it would have passed as a one-line item. Attached to the person running the most valuable company in the world, it landed differently.
Founders who had spent years suppressing the urge to look at everything suddenly had cover. But that urge was rarely about trust — it was usually anxiety. Because "good leaders delegate rather than micromanage" circulates as settled doctrine, many held back to look like good executives. Now one sentence — "well, Jensen Huang does it" — supplies the rationalization.
Before copying it, some sorting is required.
Delegation and trust matter. But the word gets used romantically. Delegation transfers not just authority but the capacity to exercise that authority responsibly. It requires the recipient to have the expertise and experience to judge and execute independently.
Early-stage startups rarely have that person. Most of the team is junior; the senior hire has not happened yet. Throwing authority into that gap is not delegation — it is abandonment. When something goes wrong, someone has to own it, and you handed authority to someone with no capacity to.
It is not only people that are missing. The systems are too. Process is not the same as bureaucracy. Even Netflix, which built its brand on "no rules," maintained minimum process for expense handling and internal approvals. The skeleton existed.
So a founder handling things directly at a stage with no seniors and no systems is not micromanaging. It is simply managing. There is nobody else to do it, and at that stage not doing it is the larger failure.
You need a fence before you can let the cattle out. Releasing them without one is not free range — it is neglect.
This does not last forever. As the company grows, experienced seniors arrive, and structures form that function under looser supervision, handing off becomes possible. Even then, a senior hire is not immediately trusted with everything. They need time to absorb context and risk, and small gates that let that be verified.
Delegating is not switching off. At points where large resources commit, where execution is irreversible, or where cost is high, the founder must stay involved — or at minimum structure things so that flow stays visible.
Delegation is not withdrawing attention. It is designing where you look and where you do not.
That sounds grander than it is. In hiring: the team lead runs screening and interviews end to end, and the founder confirms only the final offer amount and start date. In spending: below a threshold the owner executes directly and logs one line in the weekly report. Not inspecting the process every time — checking once at the end of the flow.
Which brings us back to the original story. Huang does manage gates. They just do not look like an approval chain.
As reported: no fixed reporting template, no approval line. Employees summarize the five things they consider most important in their own work and email them. More than 60 executives plus staff below them send these weekly or biweekly, and he reads close to a hundred a day himself.
The critical detail is that this is not an approval process. Nobody waits for confirmation before sending. They have already acted on their own judgment; the email shares status and open questions. Huang does not issue instructions item by item — he reads across many emails looking for the same signal recurring. In his own framing, he is trying to catch weak signals early.
So the word "gate" covers two different objects:
How much either feels like micromanagement is not remotely comparable.
Even information gates strain at scale. Reports indicate the email system began to be limited once headcount passed 30,000. Past a certain size, even the lightweight version needs rework.
The moment a gate stops guarding something operationally important and starts guarding the founder's personal taste, the character changes.
A designer brings three landing page directions, and the founder responds not on direction or message but on button color tone and font weight, sending it back. Repeat that and it stops being direction and becomes a guess-the-founder's-taste game.
One question resolves the ambiguity:
Does this decision actually affect a core metric or the company's direction, or is it a matter of my personal preference?
The first, stay involved. The second, step back.
More telling: founders who micromanage in this distorted way frequently do not manage the gates on genuinely consequential decisions. Those get handed to another leader or a senior — because ownership is unpleasant. Small things watched obsessively, big things avoided. Companies where the founder appears to be across every detail while nobody owns any outcome are more common than you would expect.
What is worth copying is not Huang's reporting frequency but the direction that reporting points. Copy the frequency without the direction and you are not managing — you are checking font preferences daily.
Ten or twenty gates is not a gate system, it is an approval maze. Pick only the few irreversible or capital-heavy points: spend above a share of monthly revenue, external contracts beyond a certain term, hiring and exiting people. If you cannot count them on one hand, half are habits, not gates. To add one, remove one.
"Just show me this one" as a standard that shifts with mood reads to employees as luck, not process. Publishing the threshold and conditions somewhere durable converts the same gate from control into procedure. The more numeric, the better.
Gates built when the company was small often survive unchanged for two years. If seniors have arrived and systems exist but the criteria have not moved, that is inertia, not management. Once a quarter or half, lay out the list and ask which ones still need you. Headcount and revenue inflections are good triggers. The trap is the same one in why company assessments change nothing: having reviewed something is not the same as having changed it.
If you erect a gate and then push blame elsewhere when a decision that passed it goes wrong, the gate loses credibility immediately. A checkpoint you created because you wanted to see it is one whose results you carry. Watch who gets blamed and you learn who really owns the gate.
Sometimes you apply the criteria correctly, pick the right gates, and employees still experience it as micromanagement. There is one more layer: whether anyone explained why the gate is there.
The same approval step reads completely differently to someone who knows the reason. To a person who knows a spend approval exists because of cash flow, it is management. To someone who only sees the founder blocking things again, it is control. The same person often reads it as control in month one and as management once they understand the company's position.
You do not have to reduce the number of gates. But the moment a "why" is attached to each one, the word micromanagement stops sticking. Writing that line is faster and more effective than agonizing over which gate to remove.
Why founders have to do the work themselves early on has plenty of precedent. The better question is not whether to break the seal, but whether you write the reason behind it.
As reported, there is no fixed template or approval line. Employees summarize the five things they consider most important in their work and email them. More than 60 executives and staff below them send these weekly or biweekly. It is not an approval mechanism — the work has already been done, and the email shares status.
Ask one question: does this decision actually affect a core metric or the company's direction, or is it a matter of personal taste? Stay involved in the first case, step back in the second.
Not if there are no seniors and no systems yet — that is managing, not micromanaging. Delegation transfers both authority and the capacity to exercise it responsibly, so handing authority to someone unprepared to own outcomes is abandonment rather than delegation.
Few enough to count on one hand. Keep only the irreversible or capital-heavy decision points, and require removing an existing gate before adding a new one.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
Assessments fail for two reasons: purposeless design and unexecuted results. Here is how to design an organizational diagnosis and actually use what it produces.
Cohort retention analysis exposes what growth rate hides. Here are the three curve shapes, benchmarks by category, and what investors actually read off the chart.
Airbnb won investors with structure, not novelty. Here are the five layers — problem framing, trust design, bottom-up sizing, a one-line revenue formula, and proof of execution.