$4 trillion, with only 36,000 people. NVIDIA reveals a harsh truth: Labor is decoupling from wealth.

[Introduction] When Walmart's revenue increased by $200 billion in a decade with zero growth in the number of employees, the "decoupling of labor" has penetrated non - technology sectors. When NVIDIA set a new record with a market value of $4 trillion, the major decoupling of labor and capital had already begun. At this time, we need to re - understand AGI: it is not a sudden event but a gradual revolution.

After its market value exceeded $4 trillion, NVIDIA's "market value per employee" also reached an unprecedented height.

This company is like an economic machine driven by scientific creativity, capital power, and a workforce the size of a small town, and is finally valued by the market at $4 trillion in assets.

According to the latest statistics, NVIDIA achieved the "number one in history" with only 36,000 employees.

Think about it carefully - a market value of $4 trillion is equivalent to the annual GDP of 1.5 billion Indians.

This ratio is crazy. Comparing the GDP of a country with the market value of a company is not fair.

What matters is the ratio and the astonishing historical trend behind it:

The major decoupling of labor and capital.

When NVIDIA, the world's largest company by market value, achieved $100 billion in revenue in 2024, it had only 30,000 employees, and it is estimated that only 6,000 - 8,000 more employees will be needed for the next $100 billion in revenue!

The major decoupling of labor and capital

In 2007, HP became the first technology company to exceed the $100 billion annual revenue threshold. At that time, HP had 172,000 employees.

The next year, IBM also joined this rank, but its employee size was close to 400,000.

Today, all the trillion - dollar technology giants in the United States show a common characteristic: their growth is almost completely decoupled from the number of employees.

Actually, this is not news.

For example, Daniel Fernández, the founder of UFM Market Trends and currently a professor of economics at Francisco Marroquín University, published relevant research in 2021:

He compared the changes in the wages of private - sector workers in the United States with the overall economic productivity and reached the conclusion of the major decoupling of labor and capital:

• Net productivity (the green line in the figure below) showed significant growth, reaching 252.9% in 2018.
• Hourly compensation (the gray line): the growth rate was slower, reaching 115.6% in 2018.

As early as 10 years ago, economist Erik Brynjolfsson warned about the paradox brought by machine learning: Productivity continues to rise, but employment may not grow synchronously.

But the devil is in the details. When examining these numbers closely, the degree of this decoupling is still surprising.

Meta's equity analyst, using the pen - name "MBI Deep DIVES", shared his observations and understanding of the major decoupling of capital and labor.

For convenience, let's call "MBI Deep DIVES" "Mr. Know - It - All". The following relevant data table is from the original article of "Mr. Know - It - All".

Apple: For every $100 billion in new revenue, the number of new employees is halved

In 2011, Apple achieved its first $100 billion in revenue with only 60,000 employees.

To achieve the next $100 billion in revenue growth, the number of new employees needed was only half of the previous time!

For the most recent $100 billion in revenue growth, Apple only added about 17,000 employees.

Of course, after adjusting for inflation, $100 in December 2011 is equivalent to $144 today. During the stage of revenue growth from $200 billion to $300 billion, Apple broke the overall rule.

Let's see how other companies gradually grew into business giants.

Alphabet (Google's parent company): 11,000 employees = $100 billion

Google has not yet exceeded the $400 billion revenue mark, but it is likely to achieve it in the next quarter!

Back then, Google used 76,000 employees to reach its first $100 billion in revenue.

Assuming that 3,000 more employees are added in the fourth quarter of 2025, Google only needs 11,000 employees for its most recent $100 billion in revenue growth!

Google follows the overall rule: the number of employees required for each additional $100 billion in revenue decreases significantly!

Microsoft: 7,000 employees = $100 billion in new revenue

Microsoft has not yet reached the $300 billion revenue milestone, but it will achieve it in the next quarter.

Historically, Microsoft was a typical labor - intensive enterprise - it hired 124,000 and 97,000 more employees respectively to reach its first $100 billion and $200 billion in revenue.

But for the most recent $100 billion in revenue growth? Only 7,000 employees!

Meta: The ratio of employees to $100 billion in revenue decreases by two - thirds

Meta is the youngest member among these companies.

It may exceed $200 billion in revenue next quarter. It took 63,000 employees to achieve its first $100 billion in revenue, while only one - third of the original number of employees is needed for the latest $100 billion in growth!

As people often say: The first mountain of gold is always the hardest to climb, and the first $10 billion in revenue is always the hardest to achieve!

Amazon: Adding 36,000 employees for $200 billion in revenue growth

Different from other technology companies, Amazon did not show an obvious pattern in the process of reaching $500 billion in revenue.

Its recruitment model is a typical example of over - recruitment after the pandemic -

The company was in the midst of a demand frenzy caused by the pandemic but misjudged the market callback in the post - pandemic era.

Although historically, it needed 200,000 - 400,000 employees for each additional $100 billion in revenue, only 36,000 more employees were added for the most recent $200 billion in revenue growth!

It should be noted that in an internal memo in June 2025, CEO Andy Jassy clearly stated:

With the promotion of GenAI and AI agents, the way of working will be completely revolutionized. The manpower required for some existing positions will decrease, and the demand for new positions will increase.

Although it is difficult to accurately predict the final balance point, the total number of employees is expected to be reduced in the next few years through the company - wide application of AI to improve efficiency.

If Amazon achieves $1 trillion in revenue with only 100,000 - 200,000 new employees in 3 - 4 years, it will not be surprising.

This means that in the early stage, 1.5 million employees created $500 billion in revenue, and in the later stage, only 10% - 15% more employees are needed for an additional $500 billion in revenue!

When $1 trillion in revenue only needs 100,000 employees, re - understand AGI

This trend may not be limited to technology companies.

In the past decade, Walmart's full - time employee count has remained relatively stable, while its revenue has increased by $200 billion during this period.

Actually, Walmart recently said that its employee size will also remain unchanged in the next three years. This means that since 2015, Walmart may have achieved $300 billion in revenue growth with almost zero employee growth!

If we calculate the cumulative incremental revenue of $100 billion recently achieved by Apple, Microsoft, Google, Meta, and NVIDIA, plus Amazon's $200 billion in incremental revenue and Walmart's approximately $300 billion in incremental revenue, we are facing approximately $1 trillion in revenue growth with only about 100,000 more employees!

It should be noted that all this happened before the rise of GenAI. The next $1 trillion in incremental revenue may require even less manpower!

This is the "major decoupling of labor and capital" - this trend has been quietly continuing for many years.

This is also the reason why there may be misunderstandings in the discussion of "Artificial General Intelligence".

Many people talk about AGI as if they will suddenly receive a company email on a Monday saying:

"We have achieved AGI and no longer need your services."

The blogger "Mr. Know - It - All" believes that it is more likely that AGI is not a moment that suddenly arrives but an asymptote that we are gradually approaching.

AI will accelerate our progress towards this "AGI".

Transcend binary thinking and re - understand the AI economy

Even if strictly from a theoretical level, we never reach AGI, in two or three decades, the actual difference between "reaching" and "not reaching" AGI may be negligible.

Such a world naturally involves deeper social, political, and philosophical impacts.

The blogger "Mr. Know - It - All" especially emphasized several considerations related to investment inspiration:

He increased the probability of OpenAI achieving $200 billion