In just 1 quarter! That is ~$5.68 Million of pure profit per worker per year!
NVIDIA + Broadcom make 44.1% of the group’s profit with only 2.5% of the workers!
80% of all workers are at just two companies: Amazon (1.58M) & Accenture (814k).
Total bottom-line profit made by all 9 giant companies in one single quarter.
3 Grown-Up Words Turned Into Lemonade Stand Words
Before we read the story, here is the 15-second cheat sheet so none of the big words can trick you:
Operating Leverage (Non-Linear Scaling)
In Plain English:Being able to double or triple your sales and profits WITHOUT having to hire double or triple the workers.
Fabless Semiconductor Company
In Plain English:A company (like NVIDIA or Broadcom) that designs the brain blueprint on computers and pays a partner factory (TSMC) to physically bake the silicon chips.
Linear People-Scale Business
In Plain English:A business where the only way to make twice as much money is to hire twice as many human workers.
Two Very Different Ways to Build a Giant Business
Why counting total profit without counting workers only tells half the story!
Suppose two lemonade companies both brag that they made $2 billion in profit this summer. At first, they sound equally awesome. But then you peek behind the curtain: Company A made that $2 billion with a small team of 1,500 inventors, while Company B needed an army of 814,000 people working day and night to make the same $2 billion!
When we line up 9 of the world’s most famous technology titans—NVIDIA, Broadcom, Apple, Meta, Google, Marvell, Intel, Amazon, and Accenture—they made a combined $165.20 billion of profit in a single quarter across 3.00 million workers.
But when you divide each company’s profit by its number of workers, the tech world splits into two totally different planets: "Secret Recipe Magicians" vs. "Giant People Armies."
The Secret Recipe Magicians: NVIDIA ($1.42M) & Broadcom ($397k)
How 2.5% of the workers captured 44.1% of the entire group’s profit!
Look at #1 NVIDIA: with just 42,000 employees, it made $59.69 billion in net profit in a single 3-month quarter. Divide $59.69 billion by 42,000 people and you get $1,421,190 of pure profit per employee in 90 days—or ~$5.68 million per worker per year!
#2 is Broadcom, making $396,667 in profit per worker in one quarter (~$1.59 million/year) across 33,000 people. Add NVIDIA and Broadcom together: they employ only 75,000 people (just 2.5% of the 3 million workers in our study), yet they pocketed $72.78 billion—44.1% of all the profit!
How is that physically possible? Because both are "fabless" designers! Their engineers invent the world’s smartest AI chip blueprints once, and TSMC’s robot factories stamp out millions of chips. When demand doubles, NVIDIA doesn’t need to double its workforce—profits shoot straight up!
Apple (#3 at $179k profit/worker per quarter) grew its workforce by only 43% over an entire 10 years, while its quarterly profit floor more than tripled from $8.7B to $30B+!
The People Army: Accenture ($2,445) & Intel’s Factory Pain
What happens when your business needs more people for every dollar—or when factories sit half-empty!
Now look at the mirror image of NVIDIA: #9 Accenture (ticker: ACN). Accenture employs 814,000 people—more workers than Google, Apple, Intel, Meta, NVIDIA, Broadcom, and Marvell combined! Yet Accenture made $1.99 billion in profit, which works out to just $2,445 of profit per worker per quarter (~$9.8k/year).
Why? Because Accenture is an IT consulting company that charges customers by the hour for human helpers. To triple its profit over the last decade, Accenture had to more than double its workforce from 384,000 to 814,000 people.
Meanwhile, Intel (#7 at $39,013/worker) shows how tough it is to own your own chip factories when you fall behind in technology. After losing the chip race to TSMC and NVIDIA, Intel had to shrink its workforce by 35.5% (from 131,900 down to 85,100 workers) just to climb out of a -$16.6B loss back to +$3.32B in profit!
Amazon’s Robot Turnaround & The AI Cash Question
How Amazon kept its workforce flat for 4 years while profits jumped by +$24 Billion!
What about Amazon? With 1,580,000 employees (more than half of all workers in the whole 9-company group!), Amazon’s profit per worker is $13,089 per quarter. During 2020–2022, Amazon quadrupled its warehouse staff from 341,000 to 1.6 million people.
Then something amazing happened: for the last four years, Amazon stopped adding total headcount—keeping its team flat near 1.58 million while filling its warehouses with smart robots and growing its AWS cloud and ad business. Over that flat-headcount stretch, Amazon’s quarterly profit swung from a -$3.9B loss to a +$20.68B profit!
One last detective tip: remember that while Google ($129k/worker) and Meta ($148k/worker) make huge accounting profits per employee, they are currently spending mountains of cash on AI supercomputers—whereas fabless winners like NVIDIA and Broadcom turn their worker profits directly into cash in the bank!
⚡ The 9-Titan Profit-Per-Worker Scoreboard (Latest Quarter)
Ranked from highest profit per employee to lowest in a single 90-day quarter
| Rank & Company | 90-Day Profit | Total Workers | 90-Day Profit / Worker | Yearly Profit / Worker | Business Style |
|---|---|---|---|---|---|
| 🥇 #1 NVIDIA (NVDA) | $59.69B | 42,000 | $1,421,190 | ~$5.68M / yr | 🧠 Pure AI Chip Blueprints (Fabless) |
| 🥈 #2 Broadcom (AVGO) | $13.09B | 33,000 | $396,667 | ~$1.59M / yr | 🔌 Custom AI Chips + VMware Software |
| 🥉 #3 Apple (AAPL) | $29.79B | 166,000 | $179,458 | ~$718K / yr | 🍏 Outsourced iPhone Factories + App Store |
| #4 Meta (META) | $9.86B | 66,775 | $147,660 | ~$591K / yr | 📱 Global Social Apps & AI Ads |
| #5 Alphabet (GOOG) | $26.47B | 204,757 | $129,275 | ~$517K / yr | 🔍 Google Search, YouTube & Cloud |
| #6 Marvell (MRVL) | $308M | 7,480 | $41,176 | ~$165K / yr | ⚡ Data Center Optical & Custom Chips |
| #7 Intel (INTC) | $3.32B | 85,100 | $39,013 | ~$156K / yr | 🏭 Owns Heavy Chip Factories (Cut 35% Staff) |
| #8 Amazon (AMZN) | $20.68B | 1,580,000 | $13,089 | ~$52K / yr | 📦 1.58M Delivery/Warehouse Army + AWS |
| #9 Accenture (ACN) | $1.99B | 814,000 | $2,445 | ~$9.8K / yr | 💼 814k Consultants Billing by the Hour |
The Lemonade Stand Rule of Operating Leverage
The strongest businesses on Earth don’t need to hire 10,000 new workers every time they want to sell more products. When a small team invents a digital recipe or chip blueprint once and sells it millions of times, every extra sale turns almost entirely into pure profit!
3 Fun Facts You Can Tell Your Family Tonight
1 NVIDIA Worker = 581 Accenture Workers in Profit
In a single quarter, one average employee at NVIDIA generates $1,421,190 in net profit—the same amount of profit generated by 581 employees at Accenture ($2,445 each)!
Big Tech Stopped Growing Headcount
Out of all 9 tech titans in this study, Google (+2.9%) was the only one adding workers—7 companies held their workforce flat at 0.0% while pouring their money into AI computers instead!
Amazon’s $24.6 Billion Robot Swing
Without growing its 1.58 million headcount over the past four years, Amazon swung its quarterly net profit from -$3.9B to +$20.68B thanks to warehouse robotics and cloud software!
Cross-referencing Net Earnings against Employee Headcount across this 9-company cohort reveals a stark divergence in operating leverage, unit economics, and economic moats. While the cohort generates $165.20B in combined net earnings across 3.00M total employees, value creation and labor intensity are inversely concentrated.
Metric Lab Grid View: Net Earnings (GAAP)
The small-multiples gallery below captures the bottom-line GAAP net income trajectory across the study cohort. Notice the staggering earnings breakout of compute and networking leaders relative to legacy capital operators:
Exhibit 1: Net Earnings Small-Multiples Gallery ($B)
Quarterly GAAP net income trajectories extracted from SEC Form 10-K & unaudited Form 10-Q filings
1. The Ultimate Moat Metric: Net Profit per Employee
Combining both charts exposes how efficiently each company converts human capital into bottom-line GAAP profit:
| Company | Net Earnings (Latest) | Share of Cohort Earnings | Employees (Latest) | Share of Cohort Headcount | Net Profit / Employee (Period) | Annualized Profit / Employee (Est.) |
|---|---|---|---|---|---|---|
| $NVDA NVIDIA | $59.69B (#1) | 36.1% | 42,000 (#7) | 1.4% | $1,421,190 | ~$5.68M |
| $AVGO Broadcom | $13.09B (#5) | 7.9% | 33,000 (#8) | 1.1% | $396,667 | ~$1.59M |
| $AAPL Apple | $29.79B (#2) | 18.0% | 166,000 (#4) | 5.5% | $179,458 | ~$718K |
| $META Meta | $9.86B (#6) | 6.0% | 66,775 (#6) | 2.2% | $147,660 | ~$591K |
| $GOOG Alphabet | $26.47B (#3) | 16.0% | 204,757 (#3) | 6.8% | $129,275 | ~$517K |
| $MRVL Marvell | $308.00M (#9) | 0.2% | 7,480 (#9) | 0.2% | $41,176 | ~$165K |
| $INTC Intel | $3.32B (#7) | 2.0% | 85,100 (#5) | 2.8% | $39,013 | ~$156K |
| $AMZN Amazon | $20.68B (#4) | 12.5% | 1,580,000 (#1) | 52.7% | $13,089 | ~$52K |
| $ACN Accenture | $1.99B (#8) | 1.2% | 814,000 (#2) | 27.1% | $2,445 | ~$9.8K |
Metric Lab Grid View: Employee Headcount Footprint
To see why the profit-per-employee metric diverges so dramatically, we inspect the workforce scale across the peer group:
Exhibit 2A: Total Employee Headcount Gallery (Thousands)
Disclosed full-time workforce footprint across all peers, showing Amazon and Accenture's massive labor scale
Unmasking Specialized Tech & Semiconductor Headcount (Excluding AMZN & ACN)
Because Amazon (~1.58M) and Accenture (~814k) absorb nearly 80% of the entire group’s headcount, they compress the Y-axis. When we filter out these two labor-heavy giants, the subtle talent inflections of pure-play cloud, software, and semiconductor peers become vivid:
Exhibit 2B: Specialized Tech & Semiconductor Headcount (Ex-AMZN & Ex-ACN)
Focused gallery of cloud, software, and fabless silicon peers re-anchored on a 0–250k scale
2. Structural Business Model Archetypes
Non-Linear Fabless Compounders (NVDA, AVGO)
$NVDA NVIDIA and $AVGO Broadcom employ just 2.5% of the cohort’s combined workforce (75,000 people total) yet capture 44.1% of the cohort’s total net earnings ($72.78B). By outsourcing capital- and labor-intensive wafer fabrication to TSMC, their R&D headcount scales linearly while earnings scale exponentially during the AI infrastructure cycle.
Linear People-Scale Services (ACN)
$ACN Accenture is the exact mirror image of Nvidia. It ranks #2 in headcount (814,000 employees—more than GOOG, AAPL, INTC, META, NVDA, AVGO, and MRVL combined) yet ranks #8 in Net Earnings ($1.99B). Over the last decade, ACN had to more than double its workforce (from 384K to 814K) to roughly triple its quarterly earnings (from ~$670M to ~$2.0B). Because revenue is tied to billable hours, it lacks non-linear operating leverage, though its tight sawtooth channel reflects low-volatility predictability.
Outsourced Manufacturing Leverage (AAPL)
$AAPL Apple ’s 10-year headcount chart is a masterclass in operational discipline. Apple grew total employees by only 43% over an entire decade (from 116K in Q4 ‘16 to 166K in Q2 ‘26, currently flat at 0.0% growth), while scaling its quarterly net earnings floor from $8.7B to ~$30B–$42B (depending on Q1 holiday seasonality).
3. Capital Allocation & Turnaround Signatures
Broadcom’s M&A Integration Playbook (AVGO)
$AVGO Broadcom ’s two charts capture its classic acquisition-and-streamline model. On the Employees chart, headcount stayed flat near ~20K for years, spiked to 37K following the VMware acquisition, and was promptly rationalized down to 33,000. Simultaneously, Net Earnings broke out to a record $13.09B (+40.6%), validating massive post-merger margin expansion alongside custom AI ASIC and networking demand.
Amazon’s Robotics & Fulfillment Inflection (AMZN)
$AMZN Amazon accounts for 52.7% of all employees in the peer group (1.58M). After quadrupling its workforce from 341.4K to 1.6M through the 2020–2022 pandemic buildout, Amazon kept headcount completely flat for nearly four years. Over that exact same flat-headcount period, quarterly net income swung from a -$3.9B trough to +$20.68B, driven by AWS mix shift, high-margin advertising, and warehouse automation.
The Cost of Lost Foundry Leadership (INTC)
A decade ago, $INTC Intel generated significantly higher net income than Nvidia, AMD, or Broadcom. Today, Intel illustrates the reverse operating leverage of an Integrated Device Manufacturer (IDM): after bloating headcount to 131.9K while losing process-node leadership, Intel suffered a -$16.6B net loss trough and slashed its workforce by 35.5% down to 85,100 (the cohort median) just to claw back to $3.32B in net income.
Post-2022 Big Tech Labor Discipline (META, GOOG)
Out of all 9 companies, $GOOG Alphabet (+2.9%) is the only company expanding headcount sequentially; 7 peers are at 0.0% and $META Meta is down 11.5% (to 66,775, well below its 87.3K peak). Big Tech is funding record AI capex via compute and depreciation rather than headcount expansion.
Comparative Context: Micron (MU) & The Memory Torque
For equity analysts comparing semiconductor archetypes, $MU Micron provides the quintessential cyclical foil to fabless category leaders. Operating with 61,000 employees and capital-intensive fabrication cleanrooms, Micron generated $37.70B in quarterly net earnings during the peak of the High-Bandwidth Memory (HBM) supply bottleneck.
While Micron’s net profit per employee surges to ~$618k/quarter during supply deficits, its capital expenditures and manufacturing overhead prevent it from sustaining the pure, asset-light margins enjoyed by NVIDIA or Broadcom across downcycles.
Capex Disparities & The Free Cash Flow per Headcount Frontier
While GAAP net income isolates operating efficiency and gross-margin leverage per employee, it amortizes physical infrastructure through depreciation schedules over 3 to 15 years. However, in the current AI transition, cash capital expenditures reveal a deeper structural bifurcation across the cohort:
- LLM Training & Datacenter Infrastructure ( $GOOG Alphabet , $META Meta , $AMZN Amazon ): These hyperscalers are deploying tens of billions in cash every quarter toward high-density GPU/TPU clusters, liquid-cooled data centers, and power purchase agreements. While their internal engineering talent becomes dramatically more productive—driving surging accounting net income per worker—their actual quarterly cash conversion is heavily reinvested back into compute hardware.
- Manufacturing Cleanroom Retooling ( $INTC Intel ): For an Integrated Device Manufacturer constructing leading-edge fabs (Intel 18A/14A), extreme ultraviolet (EUV) lithography scanners and cleanroom shells require enormous multi-billion-dollar cash outlays that must be disbursed years ahead of volume wafer delivery.
- Pure Fabless Compounding ( $NVDA NVIDIA , $AVGO Broadcom ): By passing fabrication capital expenditures to dedicated silicon foundries, fabless designers preserve an exceptionally clean bridge between Net Income and Free Cash Flow. Their human talent productivity translates directly into unencumbered sovereign free cash flow.
5 Analytical Observations on Operating Leverage & Cash Conversion
- The Free Cash Flow per Headcount Filter: Triangulating Net Income per employee against Free Cash Flow per employee highlights how capital intensity affects cash conversion. Companies exhibiting expanding GAAP profit per head while deploying substantial cash toward LLM training clusters or foundry tooling face different capital-reinvestment dynamics than asset-light peers.
- Terminal Margin Structure: Long-term DCF frameworks often distinguish fabless architectures from physical operators. Fabless models have historically sustained higher operating margins because incremental software and silicon architecture units require minimal marginal labor.
- Wage Inflation Immunity: Firms like NVIDIA ($1.42M net profit / head / quarter) are mathematically immune to tech salary inflation; payroll is a de minimis fraction of operating cash flow. Conversely, for linear consultancies like Accenture ($2,445 profit / head), wage inflation directly compresses operating margin.
- Turnaround Operating Leverage: Intel’s 35.5% headcount reduction demonstrates that IDMs face steep fixed overhead. Until fab utilization rates rise, incremental revenue will be absorbed by unabsorbed depreciation rather than falling to equity holders.
- Automation Dividend: Amazon’s flat 1.58M headcount paired with a +$24B quarterly profit swing proves that warehouse robotics and AWS automation are decoupling volume throughput from physical labor constraints.
Data and charts compiled from publicly filed SEC EDGAR Form 10-K & unaudited Form 10-Q inline XBRL filings via the FolioCue Metrics Lab Terminal.
🍋 Kid Note on Grown-Up Rules: Even when we explain big companies using lemonade stands and Pokémon cards, this website is for learning and fun—not a tip to go buy or sell stocks! Always talk with a parent or financial grown-up.
FolioCue, its principals, contributors, or related accounts hold or may hold long or short equity or derivative positions (directly or via pooled/index vehicles) in one or more of the issuers discussed herein, and may buy or sell any security at any time without notice.