Deep dive
Tech's Changing Relationship Between Employment and Growth
For twenty years, software companies converted revenue into jobs at a steady, almost mechanical rate. Nine charts trace how that conversion weakened after 2022 — across Big Tech and traditional software alike — from twenty-year hiring paths to the billions of new revenue that now arrive with no new people at all.
Every company has an implicit exchange rate between growth and jobs: how many people a new billion of revenue requires. For the software industry that rate was remarkably stable — for fifteen years you could predict a company's headcount from its revenue alone, to within a few percent. This story follows that exchange rate through SEC filings from 2005 to fiscal 2026, and watches it break. The break is not one company's layoffs and it is not a return to pre-pandemic normal: it is a synchronised, ongoing shift in how much labor growth requires — visible at Microsoft, Alphabet, Meta, Amazon, Apple, Adobe, and Oracle, and conspicuously absent at the airline included as a control. Every chart updates as new fiscal years are filed.
01
The path each company walked: revenue vs. headcount
Fifteen years on a straight line
On log-log axes, a company that hires in proportion to its growth walks a straight line. Seven software companies did exactly that for fifteen years — through the smartphone boom, the cloud transition, and Covid itself. After 2022, every path bends below its own line. Delta, the airline included as a contrast, keeps walking its line.
What it adds
On a log-log plot of revenue against headcount, every software company tracked its own straight line for fifteen years — and every one bends below that line after 2022. Delta, the non-tech contrast, does not.
Open chart, data, and methodology ↗A bend on a chart like this means the relationship itself moved. The cleanest way to see the old relationship is year by year, in growth rates.
02
Big Tech: hiring vs. revenue
The old exchange rate
Company-year observations of headcount growth against revenue growth, 2013–2025. In the years before the pandemic the points cluster round a clear upward line: grow revenue ten percent, grow headcount six to nine percent. This is the exchange rate the rest of the story watches break — visible here as the recent years peel away from the cloud of points that came before.
What it adds
Hiring and revenue growth moved together across Big Tech before the pandemic, weakened during 2020–2021, and show essentially no pooled relationship from 2022 onward.
Open chart, data, and methodology ↗One line through company-years shows the average relationship. To see it change, fit it repeatedly through time and watch the slope itself.
03
The weakening link between revenue and hiring
The slope has never been lower
The elasticity of hiring to revenue, refitted every year over the trailing eight: 1.0 means headcount keeps pace with revenue, zero means growth arrives without people. All seven software companies are at their historical low simultaneously — Oracle at 0.25, Apple 0.42, Microsoft 0.48 — and most of the descent came after 2022.
What it adds
The slope of headcount on revenue over the trailing eight years is at its historical low for all seven software companies at once — Oracle 0.25, Apple 0.42, Adobe 0.45, Microsoft 0.48, Alphabet 0.61, Meta 0.62, Amazon 0.83.
Open chart, data, and methodology ↗A rolling window smooths the turn. Splitting time cleanly at the pandemic shows how far each company actually moved.
04
Hiring's response to growth, before and after
Before and after
Fitted on 2013–2019 and again on 2021 onward, the change is not subtle: Meta's elasticity fell from 0.89 to exactly zero, Amazon's from 1.48 to slightly negative. The hollow markers say something stronger than decline — for most of these companies, revenue no longer explains headcount at all. The relationship did not weaken; it disconnected.
What it adds
Fitted before and after Covid, the elasticity of hiring to revenue collapsed at every software company: Meta from 0.89 to 0.00, Amazon from 1.48 to −0.02 — and for most of them revenue no longer explains headcount at all.
Open chart, data, and methodology ↗Elasticities are slopes on logs. The same fact in ordinary units is blunter: how many people does a new billion of revenue come with?
05
What a new billion of revenue used to buy
What a billion used to buy
In 2013–2019, a billion dollars of new revenue arrived with six hundred to three thousand new employees, depending on the company. In 2021–2025 it arrives with eighty to seven hundred — and at Amazon, $247 billion of new revenue arrived with 32,000 fewer employees. Plain arithmetic on filed numbers, no model involved.
What it adds
In 2013–2019 a new billion of revenue came with 600–3,300 new employees at these companies. In 2021–2025 it comes with 80–670 — and at Amazon, revenue grew $247B while headcount fell by 32,000.
Open chart, data, and methodology ↗If growth no longer requires proportional hiring, actual headcounts should be falling below what revenue would once have predicted. They are — measurably.
06
Software headcount below its pre-Covid revenue scaling
The missing workers, company by company
Fit each company's own pre-Covid revenue-headcount relationship, hold it fixed, and ask how many people today's revenue would have predicted. Microsoft's freshly filed fiscal 2026 sits 28% below its own benchmark — five thousand fewer people than a year earlier on eighteen percent more revenue — with Alphabet and Meta at 26% and 24% below.
What it adds
At its most recent annual filing each company employed fewer people than its own pre-Covid revenue-headcount relationship predicts: Microsoft 28% (2026), Alphabet 26% (2025), Meta 24% (2025).
Open chart, data, and methodology ↗That benchmark fits each company separately against its own longest history. A stricter pooled version asks the same question of all five giants at once.
07
How many workers would Big Tech’s revenue predict?
The pooled version agrees
A single panel model across Apple, Microsoft, Alphabet, Amazon, and Meta, fitted on 2016–2021: Big Tech reported 27% fewer workers in 2025 than its own pre-2022 relationship would predict. The gap appears at every company, and it is the deeper baseline — one that includes the pandemic hiring surge — which is why it reads larger than the per-company chart before it.
What it adds
Big Tech reported 27% fewer workers in 2025 than its pre-2022 revenue–headcount relationship would have predicted. The gap appears at every company, suggesting revenue has recently grown with much less labor.
Open chart, data, and methodology ↗Fewer people per unit of revenue has an arithmetic mirror: revenue per employee, which is now setting records.
08
Big Tech revenue and profit per employee
Output per person breaks away
Revenue and profit per employee across the five, indexed over a decade. The series drifted upward for years; after 2022 it accelerates sharply as revenue compounds against flat or falling headcount. Part of the level is inflation and pricing — but the inflection in the slope is not, and it lands in the same two years as every other chart in this story.
What it adds
Revenue and profit per employee vary dramatically across Big Tech companies and have generally risen after the pandemic hiring boom.
Open chart, data, and methodology ↗Put the pieces together and the aggregate picture is a pair of scissors: revenue one blade, headcount the other.
09
Big Tech’s Growth-without-Hiring Years
Growth without hiring
The combined picture across Big Tech: revenue keeps compounding while total headcount has gone flat since 2022 — the scissors that every preceding chart measured one way or another. Whether the cause is AI, discipline after the 2021 hiring surge, or maturity, the pattern has now persisted through three years of record revenue growth.
What it adds
Revenue growth with flat or falling headcount appeared in 6 of 20 Big Tech company-years from 2022 onward, after appearing in none of the 25 observations from 2017–2021.
Open chart, data, and methodology ↗Three years is a trend but not a law. What follows is what would confirm, extend, or break it.
The quick read
What changed, and what we still cannot say
The evidence in these nine charts is descriptive but consistent: the exchange rate between software revenue and software jobs weakened abruptly after 2022, at every company measured, under every specification tried, and at no point in the preceding fifteen years. What the filings cannot say is why. AI is the obvious candidate and the timing fits, but the same years brought the unwinding of a pandemic over-hire, higher interest rates, and margin pressure from capital spending — and annual headcount disclosures cannot separate those forces. What the data does rule out is the comfortable reading: this is not one company, not one bad year, and not a return to the pre-pandemic trend. The trend itself moved. A companion deep dive, “The Great Hiring Freeze,” follows the same shift into the labor market — hiring rates, job openings, and the graduates who arrive at doors that no longer open.
Each chart is linked to its own methodology, source list, latest observations, and downloadable chart image. Associations shown here are descriptive unless a chart explicitly states otherwise.