Category leader Apple required more than four decades to reach a $1 trillion valuation, powered by arguably the most successful consumer hardware products ever created.
Google took more than twenty years, to do the same, becoming the de-facto gateway to the internet in the process.
Anthropic, founded in 2021, is reportedly approaching a valuation of roughly $965 billion after raising an additional $65 billion and filing confidentially for an IPO. If current trajectories hold, it could cross the trillion-dollar threshold around its fifth birthday.
OpenAI isn’t far behind, reportedly valued at approximately $852 billion. Both are likely to be dragged through by the success of the SpaceX IPO.
But both appear to be heading toward a milestone that previous generations of technology giants took decades to achieve.
It’s not acceleration. It’s step change enabled by a completely different operating environment. Yet much as I’d like to focus on very juicy valuations that would be missing a much more interesting story.
It’s how few people they’re using to grow at an unprecedented pace.
The revenue-per-employee revolution
Anthropic is reportedly operating at roughly $47 billion in annualized revenue with approximately 5,000 employees. That works out to around $9.4 million in revenue per employee.
Whether you like them or loathe them, Apple and Alphabet until recently ranked among the most efficiently managed companies ever built. Yet Anthropic is operating at nearly four times their revenue-per-employee levels.
What’s even more remarkable is how early this efficiency is appearing: when Google reached roughly $30 billion in revenue, it employed around 32,000 people; when Salesforce reached similar scale, headcount approached 79,000. Anthropic arrived in comparable revenue territory with approximately 5,000 employees.
Historically, revenue and headcount moved together within fairly predictable ratios. More customers meant more salespeople. More support staff. More managers. Eventually, more managers managing managers.
Today, that relationship is beginning to break – the arrival of AI has driven three revolutionary structural shifts.
Product is labour
When companies sell intelligence through an API, the software itself performs work that previously required human labour. Every additional customer does not necessarily require additional employees as product creates capacity to satisfy demand.
Historically, scaling revenue meant scaling people. AI-native businesses increasingly scale through software-generated output instead.
Compute power is the primary input
Traditional businesses scale through growing technology-enabled human capital. AI businesses scale primarily through growing computing infrastructure.
Compute power remains expensive – ask any firm that burning through tokens. But unlike labour, it becomes more efficient over time as hardware improves, inference costs decline, and utilization increases.
Humans become more expensive and servers do not require sick leave, annual reviews, holiday schedules, or HR departments.
As a result, revenue – and potentially margin – can grow dramatically faster than employee count.
Low headcount is a strategy
Anthropic isn’t operating with 5,000 employees because it couldn’t hire more people fast enough. It’s operating with 5,000 employees because it can.
For much of modern business history, larger organizations were an inevitability. Bigger teams meant greater capability.
Today, many AI-native companies seem to view lean, simply structured organisations as a competitive advantage.
M&A compression
The phenomenon isn’t limited to public market valuations. We’re seeing similar compression in acquisition timelines.
Consider Cursor. Founded in 2022, the company has become involved in a transaction with SpaceX valued at approximately $60 billion within four years of launch.
Regardless of how one evaluates the structure or ultimate outcome, the scale is extraordinary: a company measured in the low hundreds of employees reaching a multibillion-dollar revenue run rate in just a few years.
Then there’s Wiz.
Founded in 2020, it ultimately agreed to a $32 billion all-cash acquisition by Google, creating one of the largest venture-backed start-up exits ever recorded. Historically, that level of value creation would have taken far, far longer.
What founders should learn from this
The temptation from an M&A point of view is to focus on valuations. That’s the wrong lesson. The deeper insight is that the traditional relationship between growth and margin and organizational size and complexity is changing.
The old operating model was straightforward: Hire more people to create more capacity to generate more revenue.
The emerging model looks different: Build leverage to automate work to scale revenue faster than organizational complexity.
That doesn’t mean people no longer matter. It doesn’t mean every company can remain tiny in organisational terms forever.
And it certainly doesn’t mean founders can ignore customers beyond innovators and early adopters. Crossing the “vibe chasm” into mainstream adoption still requires operational excellence, customer understanding, and execution discipline.
But it does mean that adding headcount is no longer the default answer to every growth challenge.
Increasingly, the best companies are asking a different question: How much output can we create without increasing complexity and minimising overhead?
The ratio that matters most
If you’re trying to identify the next generation of category leaders – or build one yourself – the most revealing metric will not be total revenue. It will be revenue per employee.
Because the companies defining the next decade won’t necessarily be the ones with the biggest org charts. The next trillion-dollar companies won’t be an HR bonanza. They’ll be the ones whose productivity scales way faster than their payroll.