Amazon is not cutting 30,000 jobs

because they have too many managers.

Whatever they may say, they're cutting

30,000 jobs because they need the money

to buy GPUs. That's the story nobody

wants to tell you. Not Andy Jesse on the

earnings call. Not the business press

dutifully repeating the culture

narrative. Not the analysts who'd rather

talk about AWS growth rates than the

brutal arithmetic reshaping big tech

from the inside out. So here's what

really happened. Amazon's quarterly free

cash flow went negative. Negative4.8

billion to be exact. At the exact moment

their capital expenditure hit $125

billion, the highest of any company on

Earth with 75% of it going directly to

AI infrastructure. They raised 12

billion in debt this year to fund data

centers. And then they eliminated 10% of

their white collar workforce. That's the

30,000 jobs. This is not a layoff. This

is a capital reallocation. Human

headcount is being converted to compute

capacity. Salaries are being transformed

into silicon. And if you understand why

Amazon is making this trade, you

understand something very important

about what's coming for the entire tech

industry because the numbers just don't

add up. Let's start with the

contradiction that should make you

suspicious of every official explanation

you've heard. Amazon just posted one of

its strongest quarters in years. Revenue

hit $180 billion, up 13% year-over-year.

AWS is growing at 20%, the fastest clip

since 2022. Net income surged 38%. The

stock jumped 10% after earnings. And by

every conventional measure, this company

is firing on all cylinders. And then

they fired the 30,000 people. Not

warehouse workers, not seasonal

fulfillment staff, corporate employees,

engineers, product managers, the people

who build things. and what will become

the largest layoff in Amazon's 30-year

history. The company is eliminating

roughly 10% of the white collar

workforce. 14,000 in October, another

16,000 this week. And that was planned

together. CEO Andy Jasse wants you to

believe this is about culture. The

announcement was not really driven

financially, he said. And it's not even

AIdriven, he said. Instead, he said it

was about culture. Too many layers, too

much bureaucracy. Amazon, he continues

to beat this drum. Needs to operate like

the world's largest startup. That is a

compelling narrative. It's also

carefully constructed to obscure what

the balance sheet makes obvious.

Amazon's free cash flow tells a story

that Andy Jasse, I think, would rather

not discuss on earnings calls. In 2024,

Amazon generated $ 38 billion in free

cash flow. That sounds healthy until you

understand what happened next. As of Q3

2025, quarterly free cash flow went

negative. The trailing 12-month figure

dropped 61% year-over-year. Free cash

flow margin collapsed from 8.73% of

sales to just 2.7%.

What changed? Capex. Capital expenditure

exploded. Amazon spent $83 billion on

capex in 2024. In 2025, that number

jumped to $125 billion, 61% increase.

CFO Brian Olsavski has already told

investors to expect even higher spending

in 2026. Amazon isn't stopping. Roughly

75% of all of this money is going

directly to AI infrastructure, GPUs,

custom tranium chips, data centers, and

the power systems that run them. To put

these numbers in perspective, Amazon is

now spending more on infrastructure in a

single year, one year, than the entire

gross domestic product of the country of

Morocco. They're adding 3.8 gawatt of

data center capacity annually, enough to

power 3 million homes every single year.

They're on track to double their entire

computing power by 2027. This is not a

company with excess cash lying around.

This is a company that went to the debt

markets and raised $12 billion in bonds

specifically to fund infrastructure

investments. And oh, by the way, it's

not a zero interest rate environment

anymore. So even if it's Amazon and it's

a blue chip bond, it's not free. When

Bank of America's credit strategists

analyzed hyperscaler spending, they

found that aggregate capex among the big

five, that is Amazon, Microsoft, Google,

Meta, and Oracle, now consumes, wait for

it, 94%

of operating cash flows after dividends

and buybacks. That's approaching the

theoretical maximum a company can spend

without taking on really significant

leverage. So, what does this really

mean? Amazon isn't cutting people

because the culture is broken. Amazon is

cutting people because they need the

money. Look, let's do the math that

Amazon's PR team hopes that you and I

wouldn't do. 30,000 corporate employees

eliminated. So, what does that actually

save us? Amazon's median total comp for

corporate roles runs around $217,000

per year. That includes base salaries,

stock grants, and benefits. Assume the

layoff pool skews a little bit below the

median. it. Make it make the math. Call

it $200,000 per head all in. 30,000

employees times $200,000 equals roughly

$6 billion in annual savings. Now $6

billion sounds like a rounding error

against $125 billion in capital

expenditure. But that is the wrong

comparison. The relevant comparison is

free cash flow. The actual money

available to fund investments after

you've paid for everything else. When

your quarterly FCF is $4.8 billion,

well, an extra $6 billion a year

actually matters a lot. That $6 billion

is roughly what Amazon spent building

project Reneer, its massive AI computing

platform packed with half a million

custom tranium 2 chips. It's enough to

fund half of their annual investment in

new data center campuses. It's the

difference between issuing more debt and

funding expansion internally. The

pragmatic engineer argued against this

thesis when the first wave of layoffs

hit in October. And his analysis pointed

to Amazon's $93 billion in cash reserves

and $32 billion in trailing free cash

flow, concluding that the layoffs won't

make a big difference to how much it can

invest in data centers. I would argue

that analysis used backward-looking

metrics. The cash flow picture had

already deteriorated even more.

Quarterly FCF was negative. Capex

guidance had increased by $7 billion.

And Amazon had just raised another $12

billion in debt, something they wouldn't

have had to do if they were swimming in

excess cash. The numbers don't lie.

Amazon is under financial pressure. Not

because the business is weak, but

because the AI infrastructure race

demands capital at a scale that even

Amazon's profit engine struggles to

generate. So what do we make of the

culture narrative that Josie gave?

Understanding why Jasi frames this as a

culture problem rather than a financial

reallocation reveals something really

important about how CEOs communicate

during times of technological

transformation. There are three

audiences that Jasse has to satisfy and

each one needs to hear frankly a

different version of reality. For

employees where cutting you to buy more

GPUs is a really brutal message. It

implies your work doesn't matter that

you're simply a line item to be

optimized away. The culture framing is a

lot more gentle. It suggests that the

cuts are about organizational

effectiveness and not your individual

value. You weren't eliminated.

You weren't eliminated because you cost

too much. You were eliminated because

there were too many layers between you

and the customer. And to be fair to Andy

Jesse, having worked at Amazon into the

2020s, he's not wrong. There are a lot

of layers. It is possible that the

culture narrative is true at the same

time as the hard numbers forced his

hand. For investors, admitting financial

pressure is dangerous. Amazon trades at

a premium because the market believes

AWS and AI will generate massive

returns. If Jasse said we're cash

constrained and need to cut headcount to

fund infrastructure,

the stock would be in trouble. The

culture narrative lets him position the

cuts as proactive optimization rather

than reactive cost cutting. For

regulators in the public, the AI

replacement story creates problems. If

Amazon admits these layoffs are about

freeing capital to build AI systems,

they invite a lot of scrutiny, perhaps

congressional hearings, perhaps labor

activism, bad headlines, it's better to

frame it as routine corporate

restructuring. And I think that's part

of why Jasse chose to call out

explicitly that these layoffs were not

about AI, as a productivity tool. The

tell in all of this is in Jassis's own

contradictions. In June of 2025, just

four months before that first round of

the 30,000 were laid off, he sent a memo

to employees explicitly warning that AI

would mean Amazon needs quote fewer

people doing some of the jobs that are

being done today. That memo is still on

Amazon's public blog. But by the October

earnings call, the message had shifted.

Not AIdriven. Not right now, at least.

The hedge in that sentence, not right

now, at least reveals the tension. Jasse

knows AI is part of this story. He just

can't say it out loud when the analysts

are listening. And again, to be fair to

Jasse, the culture story isn't

fabricated. It's just not the whole

truth. Did overhire during the pandemic.

Between 2017 and 2022, the company's

corporate headcount tripled. I was there

during a lot of that time. When you grow

that fast, you inevitably add layers.

Middle managers accumulate. Decisionm

slows down. I saw that happen. The

ownership culture that defined Amazon's

early years gets diluted across too many

people, too many meetings, and too many

documents. Josie has been talking about

this problem for over a year. And in

September of 2024, he mandated a return

to 5 days in the office as part of his

goal to reintroduce startup discipline

to Amazon. Every major organization, he

said, had to increase its ratio of

individual contributors versus managers

by at least 15%. and he created a no

bureaucracy email alias where employees

could flag unnecessary processes. And

that initiative, as much as it sounds

silly, generated 1500 responses and led

to apparently 450 changes. I appreciate

that focus. As someone who saw

decision-m slow down, who saw too many

meetings, too many docs, I love that

he's doing this. And so when Jasse says

these layoffs are about culture, he's

drawing on a real diagnosis. Amazon

really did become bloated. The company

does move slower than it used to. This

is not the day one thinking Jeff Bezos

like to talk about. It's very day two.

But here's what the culture and

narrative cannot explain. The timing. If

the problem was pandemic overhiring, why

didn't Amazon fix it in 2023 when

everyone else did? They laid off 27,000

people that year. Is that enough?

Apparently not. If the bureaucracy

problem persisted, why wait until late

2025 to address it with the largest

workforce reduction in company history?

And why announce those cuts the same

week you're telling investors the

capital expenditure will hit $125

billion and be even higher in 2026. The

culture problems are real, but they

didn't suddenly become urgent in October

of 2025. What became urgent was the cash

because free cash flow turned negative.

To fully understand what's happening at

Amazon, you have to zoom out and see the

competitive landscape driving this

spending. The AI infrastructure buildout

is the largest capital deployment in the

history of technology, bar none. Bigger

than the moon race, bigger than the atom

bomb, anything you can think of, bigger

than the pyramids. Goldman Socks

projects the top hyperscalers will spend

$1.15 trillion on infrastructure between

2025 and 2027, more than double what

they spent in the previous 3 years

combined. For 2026 alone, aggregate

capex among the big five is going to

exceed a half a trillion dollars. It's

estimated to be $600 billion. These are

not discretionary investments. For the

big five, these are existential. The

companies that build the most advanced

AI infrastructure first are going to

capture the majority of enterprise AI

spending for the next decade at least.

The companies that fall behind are going

to find themselves locked out of the

most lucrative technology market ever

created. Microsoft has OpenAI kind of.

Google has Gemini. Amazon has Enthropic

kind of and its own Nova models. Each of

them is racing to build the compute

capacity that will run the AI

applications every business needs. And

that is why nobody is pulling back

despite the enormous expense. Meta just

raised its capex guidance to hundred

billion for 2026. Microsoft's capital

intensity has hit 45% of revenue.

historically unthinkable for a software

company. Oracle, which barely registered

in cloud at all 5 years ago, is now

deploying tens of billions in AI

infrastructure. And Amazon's position in

this race is particularly vulnerable. It

remains the dominant cloud platform, but

it has remains the dominant has lost

ground to Microsoft, Azure, and Google

Cloud and AI specific workloads. The

perception that Amazon was slow to AI,

late to partner with leading model

providers, late to ship competitive AI

services is very widespread and has

weighed heavily on the stock. The $ 38

billion OpenAI infrastructure deal

announced in January sort of helped to

address that perception, but it also

represents yet another massive capital

commitment. When you are in the

infrastructure arms race and you need

every dollar you can find, cutting $6

billion in annual headcount to flip

yourself toward free cash flow positive,

that's not just attractive, it becomes a

necessity. So what does this mean for

the rest of us? The Amazon layoffs

matter beyond Amazon for two big

reasons. First, they reveal the true

cost structure of the AI transition.

Every hyperscaler is facing the same

pressure. How do you fund a h 100red

billion plus in annual infrastructure

spending while maintaining

profitability? The answer increasingly

is that you fund it by cutting

everywhere else. Research budgets,

experimental projects, moonshot bets,

and yes, people. This isn't about AI

replacing workers in the sense that most

people imagine. Robots are not doing the

jobs that Jasse cut. robots doing your

job is something that people

increasingly envision as the future. But

that's not where we're actually seeing

these cuts. The more immediate reality

is that AI creates capital demands so

enormous that companies have to shrink

their human workforces simply to afford

the infrastructure to play the game.

You're not being replaced by an AI

really. You're being replaced by the

need to buy GPUs.

Second, Amazon's cuts signal something

broader about tech sector employment.

The industry added workers at a very

unsustainable rate during the pandemic.

And the correction that started in 2023

is not over yet. Microsoft cut 15,000

people. Meta continues its year of

efficiency. Intel eliminated 24,000

positions. Uh I found data from

Challenger that shows that over 1.1

million job cuts hit across the economy

in 2025. And that's the highest number

of job cuts since the pandemic itself.

So we've lived through these cycles up

and down in tech. You know, I remember

2001. I remember 2008. What makes this

cycle different is that the companies

doing the cutting are not struggling.

They're not losing money. They're not

fighting for survival. These are some of

the most profitable enterprises in human

history. If they eased up on the AI

capex, Amazon would be minting money,

but their capital allocation priorities

have shifted decisively from human labor

to physical infrastructure. And that

shift is, I think, structural rather

than cyclical. The workers who remain

are going to be expected to do more with

less on a long-term basis. Amazon

employees report that managers now track

AI tool usage via dashboards.

Performance reviews increasingly f

performance reviews increasingly factor

in how effectively you leverage

automation in your role. The implicit

bargain is very clear. justify your

existence by being more productive than

the machines that are otherwise going to

come for the role. Use the machines

instead. Leverage yourself. And this

comes back to one of the big things I

like to talk about in these videos. How

are you going to use AI as a mech suit

to expand your spam to be able to do

more because you have so many more

tools? I think it's the critical job

skill for the next decade. I don't want

to be too pessimistic here. There is a

version of this story that is more

optimistic. Amazon is investing $125

billion in infrastructure that will

power the next generation of AI

services. Those services will generate

new businesses, new jobs, and new value.

And the short-term pain and layoffs is

going to unlock long-term gains across

the economy as AI unlocks productivity

improvements. History tends to suggest

that industrial revolutions, however

painful in the moment, ultimately create

much, much more prosperity than they

destroy. The problem is we're not

through the pain yet. All we know for

certain right now is that Amazon is

making a very deliberate choice. They're

choosing the GPUs over the people, data

centers over headcount, infrastructure

over organization. And they're making

that choice not because their business

is weak, but because the competitive

dynamics of AI demand capital at a scale

that forces hard trade-offs even among

the most prosperous companies on Earth.

Andy Jasse is right to keep talking

about culture. He's going to keep

invoking the startup mentality and the

ownership mindset and the need to move

faster. He's right on all of those

counts. Those things matter. They're

overdue at Amazon and they are some of

the changes that Amazon needs to make to

get back to day one thinking and

reinvent itself. I'm glad he's doing

that. But I don't want to ignore the

numbers and I think they're not being

talked about enough. Free cash flow

going negative is unacceptable. That is

a company that would have difficulty

servicing its debt before long. Capex

hitting $125 billion forces hard choices

at the margins. And the hard choices

were 30,000 people out the door. And a

12 billion bond offering to fund what

cash flows can't cover. This is the

equation that's starting to reshape big

tech. And I think we we should just be

honest about it. Human capital is at

risk when it competes with compute

capital. And if you think Amazon is

alone in making this calculation, then

you haven't been paying attention. Every

major technology company is running the

same arithmetic. The only question is

who announces their layoffs next. And

for those of you who are impacted by the

Amazon layoffs, you are going to make

it. You are going to get through. I am

an ex Amazonian. There's a whole bunch

of ex Amazonians out there. Reach out to

us. We're all alums. We'd love to be

helpful. Best of luck out there.