HomeThere is a number that quietly unsettles every founder who hears it. One of the most valuable companies the world has ever produced — a business whose chips sit underneath nearly every serious AI system on the planet — runs on a workforce that, set against its scale, is astonishingly small. Nvidia crossed valuation milestones that read like typographical errors, and it did so without the sprawling armies of staff that those numbers seem to demand. The widely reported figure that travels with the company is its revenue and valuation per employee — a ratio so high it makes most large technology peers look bloated by comparison.
Read that too fast and you draw exactly the wrong conclusion. Lean team, enormous value — so the lesson must be: hire fewer people. Founders all over Singapore are reaching for that conclusion right now, usually with a spreadsheet open and a hiring plan they are about to slash. They have heard that AI lets you "do more with less," they have seen the Nvidia numbers, and they are quietly congratulating themselves on the headcount they are not going to add.
That is the misread. And it is expensive.
Because Nvidia is not a story about hiring fewer people. It is a story about redesigning the work so that a small number of people can carry an extraordinary load — about leverage, focus, and structure, not subtraction. The lean team is a result, not a strategy. It is what you get when you design the work brilliantly, not when you simply refuse to hire.
This is that lesson, decoded for a Singapore founder — what the world-class move actually is, why the obvious reading of it is wrong, and how a ten-person firm in Tanjong Pagar can copy the design rather than the headline.
The world-class move: leverage, not headcount
Start with the thing the Nvidia story is actually about, because it is not the thing most people think.
The naïve version goes like this: Nvidia is worth an absurd amount and employs comparatively few people, therefore the magic is fewness. Keep the team tiny and the value-per-head stays enormous. It is a tidy story, and it is almost entirely backwards. The fewness is downstream of the design. Nvidia did not get there by capping headcount and hoping the value would follow. It got there by building a business in which a relatively small group of exceptionally leveraged people could move an enormous amount of value — and the lean ratio is simply the arithmetic that falls out the other end.
So what is the leverage actually made of? Strip away the chip-industry specifics and four structural moves remain, and every one of them is portable to a business of any size.
The first is brutal clarity of mission. Nvidia is famous internally for a flat, fast structure where information moves directly rather than climbing and descending a tall management ladder. The point of that flatness is not to look modern in a culture deck. It is leverage. When everyone is pointed at the same clearly understood goal, you do not need layers of managers translating intent downward and status upward. The org chart itself is a tax on output, and a flat one collects less tax. A small team that all understands exactly what it is building can out-execute a large team that spends half its energy keeping itself aligned.
The second is that the routine layer is carried by machines and tooling, not by people. This is the part most observers skip because it is invisible from the outside. A leveraged organisation does not have its scarce, expensive humans doing retrieval, reconciliation, status-chasing, formatting and the thousand small repetitive tasks that fill an ordinary company's week. That work is automated, tooled, or systematised so that human attention is reserved for the things only humans can do well. The lean ratio is not "fewer people doing the same jobs." It is the same people freed from the jobs a machine should be doing.
The third is that the humans who remain are kept relentlessly on high-value work. Judgment. Design. The hard call. The relationship. The problem that has never been seen before. In a leveraged company, you almost never find a senior person buried in administrative drudgery, because the entire system is engineered to keep its best minds pointed at its hardest problems. That is not a perk. It is the whole source of the per-employee number.
The fourth is a culture that treats the team as a multiplier, not a cost. Lean organisations that work are not stingy; they are concentrated. They would rather have ten people operating at the absolute top of their capability than thirty operating at half. The investment per person is high — in tools, in trust, in autonomy — precisely because the return per person is high. This is the opposite of the cost-cutting instinct the Nvidia story is usually mistaken for. It is an investment thesis dressed as a headcount number.
A lean team is not a small team that does less. It is a team where the routine work has been engineered away, so every person operates at the top of their judgment — and the headcount number is just the receipt.
Put those four together and the Nvidia "lesson" stops being about size at all. It becomes a question every founder can ask of their own business, today, regardless of headcount or capital: Is each of my people operating at the top of their capability, with the routine cleared beneath them and the mission perfectly clear above them — or are my best people drowning in work a machine should be doing? Answer that honestly and you usually find the real opportunity is not to hire fewer people. It is to redesign what the people you already have actually spend their days on.
This is the move that has become newly, dramatically available to ordinary businesses. For most of corporate history, leverage of this kind required either a uniquely defensible product or a decade of internal systems-building that only giants could afford. AI collapses that. The routine layer that used to demand a dozen contract staff to absorb can now, increasingly, be carried by tooling that a small firm can stand up in weeks. The structural advantage that made Nvidia's per-employee number possible is, for the first time, within reach of a Singapore SME. That is the genuinely new thing — and it is also exactly where most founders are about to get it wrong.
A small, intensely focused team working in a calm modern Singapore studio surrounded by ambient data and tooling
The misread: "lean" does not mean "understaffed"
Here is the mistake, made in earnest by good founders with sound instincts.
They see the lean-team story, they feel the pull of AI's "do more with less" promise, and they reason straight to headcount: freeze hiring, keep the team small, let AI absorb the gap. The logic feels airtight. The savings are immediate and visible. The hiring plan shrinks, the burn rate drops, and on the spreadsheet it looks like discipline.
It is not discipline. It is a category error — and it fails for a reason that is almost mechanical.
A lean team and an understaffed team look identical on a headcount line and are opposites in reality. A lean team is one where the work has been redesigned so a small group can carry a large load comfortably, because the routine has been engineered away and each person is operating at the top of their value. An understaffed team is one where the load was never redesigned at all — the same sprawling pile of routine and judgment work still exists, but now fewer people are buried under it. One is leverage. The other is just overload wearing leverage's clothes.
The founder who freezes hiring without redesigning the work gets the second one every time. The routine tasks do not vanish because you declined to hire; they simply land on the people who stayed. Your best engineer is now also doing the reconciliations. Your one operations hire is now also chasing every status update. Morale erodes, the senior people quietly start looking, and the very leverage you were chasing collapses — because your highest-value people are now spending their days on your lowest-value work. You have not built Nvidia's machine. You have built its photographic negative.
There is a second, subtler failure hiding inside the first. AI does not replace jobs; it replaces tasks. This is the load-bearing idea, and the lean-team misread runs straight past it. A role is a bundle of tasks — some routine, some judgment-heavy, some relational, some that genuinely require a person to own and answer for the outcome. When you point AI at a role, it does not vaporise the role. It dissolves the automatable tasks inside the role and leaves the rest standing, often more exposed and more important than before. So the move was never "this role can go." It was always "which tasks inside this role can move to the machine, and what does the human do with the time that frees up?" The founder who skips straight to "we don't need the hire" has answered a question nobody asked, and missed the one that actually matters.
This is exactly the trap we watch SMEs walk into when they adopt AI tooling without redesigning the work around it — the pattern we explored in how Microsoft is putting Copilot into every job. The tool arrives, the routine is technically automatable, and yet nothing improves, because the work was never re-shaped to take advantage of it. The capability was bought; the redesign was skipped. The tool is the easy part. The redesign is the work.
So the real lesson Nvidia teaches is not "stay small." It is "redesign the work so that staying small becomes a strength rather than a liability." Those are different universes. One ends in a burnt-out skeleton crew apologising to customers. The other ends in a small team that genuinely punches far above its weight. The difference between them is not headcount. It is design.
Redesign, not reduction: the three-bucket model
If "hire fewer people" is the wrong frame, what is the right one? It begins with a better question — arguably the single most useful question a founder can ask of AI:
"If the machine clears the routine work, what could my people finally do with that time?"
That reframing produces a concrete, repeatable operating model, and the good news is that it does not require a chip giant's resources to run. You do not start with the org chart. You start with tasks. Take any function in your business — sales, customer service, operations, finance, delivery, even parts of your own founder role — and decompose it into the discrete tasks people actually perform week to week. Then sort every task into one of three buckets.
Bucket one — what machines do better
These are the tasks where capable tooling genuinely outperforms a person on speed, consistency, availability and cost. Drafting the first version of a proposal or a reply. Retrieving and summarising information. Logging, formatting, reconciling, status-chasing. Pre-filling forms. Triaging an inbox. Generating the routine report nobody enjoys writing. In a small business this bucket is almost always larger than the founder expects, because routine work hides inside senior roles — the experienced person who spends a third of their week on tasks a tool could do in seconds. Route this work to the machine without sentiment. It is genuinely better at it, and every hour it reclaims is an hour your scarce people get back.
Bucket two — what humans do better
These are the tasks where the person is not merely preferable but load-bearing. The judgment call that no rule quite covers. The difficult client conversation. The creative leap. The relationship that took years to build. The decision someone has to own and answer for. The novel problem the business has never faced. In a small Singapore firm, where trust is dense and word travels fast, this bucket is disproportionately valuable — getting it wrong does not cost you one task, it costs you a relationship and a reputation. This bucket is small in volume and enormous in value. Automate it carelessly and you do not save money; you bleed it.
Bucket three — what they do better together
This is the bucket most founders forget exists, and it is precisely where Nvidia-style leverage lives. It is the salesperson who now runs three times the meaningful pipeline because the tooling prepped every account, drafted every follow-up, and surfaced every signal before the human walked into the room. It is the consultant whose AI assembled the analysis and whose judgment decided what it meant. It is the founder who reclaimed ten hours a week from admin and spent them on the two decisions that actually move the company. Together they are not a smaller team doing the same job. They are the same team doing a far higher-value job.
The reason bucket three is so easy to miss is that it never shows up in a cost spreadsheet. A spreadsheet that asks "how many hires can I avoid?" will find buckets one and two and stop. It has no column for "value created when a freed person is pointed at higher-value work," because that value is diffuse, arrives later, and lands on the revenue line rather than the cost line. So the headcount-first instinct structurally undercounts the upside and overcounts the savings. It sees the hire you can skip and is blind to the growth you could unlock. The redesign-first instinct inverts this: it treats the freed capacity as fuel for growth, not as a line item to delete.
Run this exercise honestly across even a small company and the lean-team number stops looking like a constraint you imposed and starts looking like a result you earned. Bucket one shrinks the routine load. Bucket three multiplies the value of the people you have. Bucket two is fiercely, deliberately protected. That is redesign. The small headcount is a by-product of the leverage, not the goal of it — and it is exactly the discipline a founder can run this quarter without a single new dollar of capital. Mapping that automation surface and building the tooling safely on top of it is the work our strategy-and-implementation sister practice, Freemansland, exists to do alongside a founder rather than leaving them to guess at it.
The house rule we keep returning to is simple enough to put on a wall: redesign before you reduce. Reduce first and you overload your best people, automate the wrong tasks, and spend the following year firefighting and rehiring. Redesign first and the leanness takes care of itself — cleanly, defensibly, and without burning the team you built.
What this means for Singapore founders
Now bring it home, because Singapore is a particular kind of place to run a lean, high-leverage business, and the local physics matter.
Singapore is, in many ways, the ideal environment for the Nvidia lesson — and a few of its constraints are precisely what make redesign the only sensible path. Consider the realities a founder here actually operates under.
Talent is scarce and expensive, and that is an argument for leverage, not against it. Singapore's labour market is tight; good people are hard to find and costly to keep, and hiring is further shaped by manpower policy and quota considerations that mean you cannot simply throw bodies at a problem. For most founders this feels like a ceiling. Reframed through the Nvidia lesson, it is an incentive. If you cannot easily add the tenth, fifteenth or twentieth hire, the entire value of redesigning the work so your existing nine carry the load of fifteen goes up. The constraint that frustrates a headcount-first founder is the exact condition under which a redesign-first founder thrives. Scarcity of talent is not the enemy of a lean team; it is its reason for existing.
The market is small, which makes every relationship dense and every reputation fragile. Singapore is a high-trust, word-travels-fast economy. A customer will happily let an automated system handle a routine request and will walk over one badly handled exception — and tell their entire network why. This is precisely why the lean-team move here must be a redesign, not a reduction. If "going lean" means understaffing the human judgment that protects your most important relationships, you will save a little on payroll and lose the thing the whole business runs on. The bucket-two work — the difficult conversation, the goodwill exception, the relationship — is disproportionately valuable in a market this tight, and it is exactly the work a careless lean strategy destroys first.
The founder is usually the most over-loaded person in the building. This is the Singapore-specific point that matters most, and it is the one founders feel in their bones. In a small local firm, the founder is the bottleneck — selling, delivering, hiring, doing the accounts, answering the emails, making every call. The Nvidia lesson, applied honestly, often starts not with the team but with the founder's own week. Decompose your own role into the three buckets and you will almost always find a third of it sitting in bucket one — routine work you are doing because there was never anyone else to do it. The single highest-leverage redesign in most Singapore SMEs is the founder reclaiming the hours they currently spend on tasks a machine should own, and pointing those hours at the two or three decisions only they can make. That is not a luxury. In a one-founder business, it is often the difference between a company that scales and one that stays permanently stuck at the size of its founder's calendar.
There is a competitive dimension to all of this that founders here should sit with. The companies redesigning around AI are not only the giants; they are increasingly the disciplined small firms, because a small firm has less legacy structure to unwind and can re-shape its work faster than a large one. This is the same structural truth playing out in Google's AI-first mandate and in Meta flattening its org chart around AI — except a Singapore SME can move faster than any of them, precisely because there is no sprawling hierarchy to convince. The lean-team advantage is, paradoxically, easier for a small founder to capture than for a giant. The giants have to dismantle decades of accumulated process. You just have to design the work right the first time.
What this adds up to is a genuinely optimistic position for a Singapore founder willing to do the design work. You operate in a market that punishes bloat, rewards trust, and is studded with national infrastructure built to help you move people up the value chain rather than out the door. The lean, high-leverage company is not a Silicon Valley import that does not translate here. It is arguably better suited to Singapore than almost anywhere — provided you copy the design and not the headline.
A Singapore founder at a calm desk, freed from administrative clutter, focused on a single high-value decision
The Singapore enablers: a system built for redesign
Here is the part of the story that founders elsewhere genuinely envy, and that Singapore founders too often overlook because it is so familiar it has become invisible. When a company in most economies redesigns work and frees up capacity, the displaced worker is largely the firm's problem or the individual's. In Singapore, there is a dense, funded, deliberately constructed system whose entire purpose is to move a person from a role that is shrinking into a role that is growing. The infrastructure for redesign already exists. Most founders simply do not use it.
Job Redesign and the reskilling ecosystem. Workforce Singapore, e2i and SkillsFuture run Job Redesign initiatives and Career Conversion Programmes built precisely for this moment — to fund and support the journey from routine tasks toward higher-value work. The framing matters enormously. A founder who approaches AI as "redesign and upskill my people" is not just choosing the kinder path; they are choosing the path the national system is actively built to subsidise and support. A founder who approaches it as "cut headcount" forfeits that support and absorbs the reputational cost. The incentives are pointed, on purpose, at redesign rather than reduction. For a small firm with thin margins, the difference between tapping that support and ignoring it can be the difference between a redesign you can afford and one you cannot.
Tripartism — the social contract behind the schemes. Singapore navigates economic disruption through tripartism: government, employers and unions moving together. This is not decoration; it is how the economy has metabolised every major shift — manufacturing, globalisation, the financial crisis — without the social fractures other economies suffered. AI is simply the next wave. For a founder, tripartism means the national stance is not "automate and let the chips fall." It is "redesign, reskill, and bring people along." A founder who runs their AI transition in that spirit — gradual, transparent, upskilling-led — moves with the grain of the entire system rather than against it, and earns the goodwill that comes with it.
Trust and governance as the enabling layer. A lean, AI-leveraged business only works if the AI is trustworthy — if it does not quietly mishandle customer data, make decisions it cannot explain, or expose the founder to a compliance problem they did not see coming. For regulated or data-sensitive work, this is not optional. Building the governance posture in from the start — so that the leverage is defensible, not just efficient, and so the firm can actually qualify for the grants and schemes that reward responsible redesign — is precisely the work our governance, risk and grants sister practice, FMC Collective, exists to handle. The point is not bureaucracy. It is that trust is the thing that lets a small team operate with the autonomy a lean model requires. A team you cannot trust to handle AI responsibly is a team you have to supervise — and supervision is the opposite of leverage.
Taken together, these enablers mean the Nvidia lesson is not a lonely Silicon Valley fantasy for the Singapore founder. It is a move the entire national system is, in effect, set up to underwrite — if you frame it as redesign. The schemes fund the human transition. The tripartite contract protects the trust. The governance layer makes the leverage defensible. A founder who reaches for "cut headcount" walks past all of it. A founder who reaches for "redesign and upskill" finds a surprising amount of the country already standing behind them.
The operator's playbook: five moves to run now
Strategy is only as good as the next action it produces. If you are a founder in Singapore — running an agency, a SaaS, a services firm, a small manufacturer, anything where people are your main cost and your main constraint — the Nvidia lesson compresses into five concrete moves. Run them in order.
1. Map your own week first
Before you touch the team, decompose your own role into the three buckets. Take a representative two weeks of your work and tag every task: routine, judgment, relational, or founder-only. You will almost certainly find a third of your week sitting in bucket one — admin, drafting, chasing, formatting — work you do only because there was no one else. This is the highest-leverage redesign in the company, and it costs nothing but honesty. A founder who reclaims those hours and points them at the two decisions that actually move the business has already captured more value than most AI projects ever deliver.
2. Map tasks across the team, not roles
Now do the same for the business. Pull a representative month of work — tickets, deals, deliveries, internal processes — and tag every task by type. Do not start from job titles; start from what people actually do. You will reliably find that a large share of the volume is genuinely routine, and that it is scattered inside roles that look senior. That map is the single most important artefact in the whole exercise. Skip it and every later decision is a guess.
3. Automate the routine — visibly to the team, invisibly to the customer
Deploy tooling against bucket one. But how you frame it to your own people determines whether it works at all. Tell them plainly: this clears your queue so you can own the work that matters. Adoption collapses the instant staff believe AI is in the building to replace them — they route around it, withhold the tacit knowledge that makes it useful, and wait for it to fail. Frame it as the thing that finally takes the drudgery off their desk and they become its best trainers. To the customer, the automation should be invisible: faster answers, fewer errors, no sense of being handed to a bot.
4. Redesign the human role upward
This is the move almost everyone skips, and it is the one that separates a lean team from an understaffed one. Once the routine is gone, rewrite each role around judgment, complex problem-solving and relationship ownership. The role did not get smaller; it got harder and more valuable, and pay, title and expectations should reflect that. Automate a third of someone's tasks and leave their job definition untouched and you have created a confused, under-rewarded employee. Redesign the role around its new high-value core and you have created your most productive one.
5. Reskill, don't release — and use the schemes
Move freed capacity into the redesigned roles, supported by Singapore's reskilling infrastructure — Job Redesign initiatives, Career Conversion Programmes, Workforce Singapore and e2i support, and SkillsFuture. The reclaimed hours should become growth, service quality and retention, not a one-time cost cut booked in a single quarter. A firm that releases people banks a small saving once. A firm that reskills them compounds a capability advantage for years — and keeps the hard-won institutional knowledge that walks out the door with every departure. End-to-end, this is the redesign we help Singapore founders run: finding the real automation surface and building the tooling safely with Freemansland, and locking down the governance, grants and risk posture with FMC Collective so the move is as defensible as it is efficient.
Run these five and the Nvidia lesson stops being a number you admire from a distance and becomes something you execute deliberately, on your own terms, with the national system and your own team moving alongside you rather than against you.
The investor's close: revenue per employee is the real scoreboard
Now for anyone allocating capital — and for the founder who is the capital allocator in their own business — because this is where the whole argument cashes out.
The naïve reading of Nvidia is "a small team can be worth a lot." True, but useless as guidance, because it points you at the wrong lever. The number that should actually move is revenue per employee — and beneath it, the operating leverage of the whole business.
Here is the mechanism. A service or product business has historically scaled the way a galley scaled: more output meant more oars, more rowers. Cost-to-serve and headcount marched together; growth and labour were chained. The lean-team move breaks that chain. When the routine work migrates to tooling that costs a fraction of a salary and scales without hiring, the relationship between growth and headcount finally decouples. The firm can take on more customers, ship more, deliver more — without the labour curve rising in lockstep. That is operating leverage of a kind small businesses have rarely had access to: closer to software economics than to traditional services. It is, stripped of the chip-industry glamour, exactly what the Nvidia per-employee number represents.
But here is the crux, and it is the same crux for a unicorn and for a five-person Singapore agency: the leverage only shows up on the income statement if the business is redesigned to capture it. Two founders can adopt the identical AI tooling and end up in opposite financial places.
The founder who merely buys AI and bolts it onto an unredesigned business will show, a year later, a slightly lighter admin load, a meaningfully larger software bill, an overloaded senior team, and — if they cut headcount without the redesign — a quiet drift in quality and a couple of departures. Cost-to-serve barely moves, because the savings were eaten by tooling spend and churn. On paper they "did AI." In reality they spent money to stand still.
The founder who redesigns around AI will show something categorically different: rising output, flat-to-falling cost-to-serve, and revenue per employee climbing as the same people — now freed from the routine — carry materially higher-value work, deepen relationships, and grow the book. Same tooling. Same starting headcount. Completely different result. One bought a tool. The other rebuilt the machine around it.
The question is no longer "is this business using AI?" Everyone is. The question is "is this business redesigning around AI, or just buying it?" Only one of those shows up as durable revenue per employee — and that ratio, not the logo on the slide deck, is the real scoreboard.
Nvidia's lean team is the visible result of years of that second kind of work — leverage engineered into the structure of the company until the per-employee number became extraordinary almost as an afterthought. A Singapore founder cannot copy the chips. But they can copy the design: clarity of mission, the routine carried by machines, the best people kept on the hardest problems, and a small concentrated team trusted to operate at the top of its capability. That design is not capital-intensive. It is discipline-intensive — and discipline is the one input a founder fully controls.
Nvidia did not get lean by hiring fewer people. It got lean by designing work so brilliantly that fewer people were all it needed. The founders who copy the headline will overload their teams and spend next year apologising and rehiring. The founders who copy the design will quietly build companies that punch far above their weight — more leveraged, more resilient, and more valuable per person than anyone expected of a firm their size. The lesson is not in the smallness. It is in the redesign — and it is sitting in plain sight, waiting for Singapore's founders to read it the right way. For more on how the world's most valuable companies are decoding the AI workforce shift for Singapore, explore the rest of our Insights.