跪拜 Guibai
← Back to the summary

Big Tech's 2025 Layoffs Are Not About Survival — They're a Strategic Bet on AI

Recently, people keep asking whether the internet industry is finished.

I usually don't answer directly. Because "finished or not" is a subjective judgment, and what everyone really wants to ask is another question: Why did those big-tech jobs, once seen as "iron rice bowls," suddenly lose their appeal?

This question is much more practical than "Is the internet industry finished?" Today, I'll try to talk it through thoroughly.

1. First, Look at Some Unsettling Numbers

I don't like using vague terms like "winter" or "reshuffling." Numbers are more direct.

Overseas, data tracking site layoffs.fyi shows that US tech industry layoffs in 2025 have exceeded 110,000 people. This isn't just "optimization" at one or two companies; it's a line-up:

That's Silicon Valley. Looking back at the domestic scene, it hasn't been much gentler.

Alibaba's financial report was the most startling—by the end of fiscal year 2025, its total headcount had dropped from 205,000 a year earlier to 124,000, a loss of 80,000 people in one year. The first time I saw this number, I was stunned.

Of course, there's an impact from changes in reporting scope, like the spin-off of Sun Art Retail and the sale of Intime department stores. But even stripping those out, the "slimming down" of the core business was real.

Baidu launched its "largest layoff in recent years" at the end of 2025, with some departments in the MEG business group optimizing 20% to 30% of staff, covering fresh graduates, un-regularized employees, and even veterans with over ten years of service. Some entire teams disappeared.

As for ByteDance, although it didn't explicitly announce layoffs, its net profit fell by over 70% year-on-year—the bulk of its earnings were eaten up by AI.

Tencent is one of the few still bucking the trend and adding staff, growing from 82,000 to 87,000 employees, but in the whole industry, this kind of example is a minority.

Laying these numbers out together, you'll find this isn't a localized problem at one company, but an industry-level trend.

2. How This Wave is Different from Before

This isn't the first time internet companies have laid people off.

The O2O bust in 2015, the "capital winter" from late 2018 to 2019—each time, companies fell and people left.

But this wave is genuinely different.

The difference is this: In the past, unprofitable ventures were cut; now, profitable ones are being cut too.

Before, the logic for internet company layoffs was clear—a business burned cash for two years without finding a model, capital dried up, and the company naturally contracted. Those laid off felt bad but could understand: my business lost, on to the next one.

But this time, many companies are laying off while profits are still growing.

Meta, Alphabet, Amazon, and Microsoft are projected to spend $350 to $400 billion in capital expenditure in 2025, most of it poured into AI. On one hand, unprecedented investment; on the other, unprecedented streamlining. This isn't a "can't hold on" layoff; it's an "active choice" layoff.

A second difference: This time, there's a very convenient reason—AI.

Challenger's data shows that for full-year 2025 layoffs, about 54,000 to 55,000 people were explicitly labeled by companies as "due to AI and automation."

Anthropic's own report is even more blunt: 75% of programmer tasks are already covered by AI. Global programmer jobs fell 8% in 2025, and 1 in 4 junior programmers faces unemployment.

This gives companies a particularly smooth narrative: It's not that I don't want you, it's that the times don't want you.

But I think simply attributing this wave of layoffs to "AI replacing people" is lazy thinking.

3. AI is a Good Excuse, But Not the Whole Truth

Let's be clear about one thing first: I acknowledge that AI's impact on employment is real, especially for junior programmers, entry-level coders, simple script developers, and some testing roles. These jobs are characterized by high repeatability, clear rules, and high standardization—exactly the kind of work AI is best at eating up.

But to say that today's internet company layoffs are all because of AI is to miss the forest for the trees.

The first overlooked fact is: The internet industry's own growth has peaked.

China's mobile internet user base hit its ceiling long ago. The growth rate of the overall internet advertising market has fallen from over 30% to just over 10%. E-commerce penetration is so high it's hard to tell new stories. An industry that is no longer growing at high speed should naturally enter a phase of "intensive cultivation"—meaning it no longer needs so many people to expand territory, because there's no territory left to expand.

The second overlooked fact is: Over the past decade, the human efficiency of internet companies has not been high.

This industry lived too long under the logic of "get big first, get strong later." Hoarding people during expansion was a rational choice; but once the growth curve flattens, these "hoarded people" become a burden on the financial report. The emergence of AI just gave CEOs a more respectable reason to do what they've long wanted to do.

Ultimately, layoffs are never determined by a single variable. They are the result of a pile-up of factors: peak growth, capital retreat, AI substitution, management inefficiency, strategic mistakes... Putting all the blame on AI is neither accurate nor fair.

4. It's Not Just the Laid-Off Who Are Scared Off

The deepest impact of this wave of layoffs isn't actually on those who've already taken their severance package and left.

Those laid off certainly suffer, but at least they have closure. The ones suffering more are those still at the table but who can no longer see the future.

Caixin had a report with a jarring headline—"Big Tech Lays Off the Middle-Aged, Scares Off the Young."

This isn't an exaggeration. I have a few examples around me: programmers born in the 90s and 95s, who used to see "getting into a big tech firm" as the standard career path, are now hesitating. What they fear isn't the grind inside, but being kicked out in their thirties with nothing left on their resume but a company name.

And those already inside, in their thirties? Their anxiety is visibly rising. Before, job-hopping was a shortcut to a raise; now, it's a prelude to unemployment. The market has shifted from a "talent grab" to a "job grab." Stories of senior engineers sending out a hundred resumes with no response are no longer isolated cases.

The spread of this sentiment is more alarming than the layoffs themselves. If an industry makes young people unwilling to go all-in and middle-aged people afraid to take a leap, its vitality is fundamentally broken.

5. So, Who's Doing Well in This Wave?

I don't want to write a whole article of doom and gloom. Because in this wave, some people are genuinely doing well.

One type is the high-end roles deeply committed to technology.

System architects, AI algorithm engineers, and compound technical talents who understand business—demand for these hasn't fallen but risen, with salary increases still between 120% and 210%. The market is eliminating "repetitive mental labor" but is even hungrier for genuine "creative mental labor."

Another type is companies that bet on the right new battlefield.

JD.com aggressively entered food delivery in 2025, burning cash to create a loss (Q3 operating loss of 1.05 billion yuan) but grabbing future market share; Pinduoduo locks in talent with high salaries and incentives, with 7.9 billion yuan in stock-based compensation in 2025, averaging 310,000 per person; ByteDance poured profits into AI, making its books look ugly, but its determination to secure a position is clear.

There's another type: those who were laid off but didn't see themselves as victims.

I know a few programmers who were "optimized," took a sum of money, and turned around to do independent development, become tech bloggers, or build small, beautiful SaaS products. What they say is very consistent: They should have left long ago, they just lacked the courage.

This sounds a bit like a motivational cliché, but it's not entirely. Behind it is a sober judgment—in an industry that is no longer growing at high speed, tying your fate to a single company is itself a high-risk act.

6. Some Personal Opinions

Having talked this far, I want to share a few of my own judgments. They might not be right, but they are honest.

First, the "golden age" of the internet industry is indeed over.

But "over" doesn't mean "finished." An industry returning to a normal growth rate is still one of the most lucrative in society in terms of compensation. It's no longer the 2015 casino where anyone who entered could turn their fortunes around, but it remains a track worth taking seriously.

Second, "AI replacing people" will happen slower than imagined, but more thoroughly than imagined. Slow, because corporate organizational inertia, business complexity, and compliance requirements can't be switched overnight; thorough, because once a position is proven replaceable by AI, it's never coming back.

So instead of being anxious about "Will I be replaced?", it's better to think clearly: "How much of what I do is truly something AI cannot do?"

Third, for practitioners, "climbing the corporate ladder" is no longer the only path.

I'm not saying big tech is bad, but rather that the risk of single-point dependency has never been this high. Side hustles, independent products, personal branding, cross-disciplinary skills—these things once seen as "not serious work" have today become tools to hedge career risk.

Fourth, regarding the industry itself, I'm actually not that pessimistic.

The internet is not a dying industry; it's just transforming from an "expansion-type industry" into an "operation-type industry." Expansion-type industries reward boldness and speed; operation-type industries reward patience and depth. This isn't regression; it's maturation.

It's just that the price of maturation always has to be paid by someone.

And this wave of layoffs is paying that price.

Final Words

If you are currently going through a period of career confusion, I want to leave you with one thought:

Don't mistake an industry cycle for your own personal failure.

Amidst the great storms of the era, individual effort certainly matters, but what matters more is recognizing which boat you're on. When the boat is sinking, paddling desperately is useless; but if the boat is just slowing down, what you need to do isn't jump into the sea, but learn to live well even on a slow boat.

This internet boat hasn't sunk; it's just slowed down.

A slowed-down boat can still reach the shore. It's just that this journey will test patience, and real skill, more than ever before.

Data Source Notes: The data involved in this article is synthesized from public sources including layoffs.fyi, various company financial reports, Caixin, 36Kr, Huxiu, ChinaVenture, Challenger Report, Anthropic Report, and the World Economic Forum's "Future of Jobs Report 2025." Data is current as of late 2025 to early 2026 and is for reference only.