Jensen Huang has also become Liu Qiangdong's "brother".

On May 27th, Liu Qiangdong, the founder of JD.com, who hadn't made high - profile statements for a long time, made an unusually bold commitment: "In the face of the impact of new technologies, whether it's about innovating business models or developing more new businesses, we must do everything possible to secure the jobs of hundreds of thousands of employees, including our blue - collar workers (our brothers' jobs)."

He emphasized that "JD.com will not lay off any front - line employees who are replaced by machines."

If this statement had been made two or three years ago, it might have been regarded as just "boss talk." However, in 2026, when Meta has just laid off 8,000 employees, Microsoft and Google are continuously downsizing, and Mark Zuckerberg even asks employees to "compete with AI," Liu Qiangdong's claim of "not laying off anyone" has created a significant gap in the business ethics of the tech circle when facing the impact of AI.

What's even more interesting is that when the idea of large - scale white - collar layoffs caused by AI is sweeping Silicon Valley, Jensen Huang, the CEO of NVIDIA, has recently made a series of intensive statements with a core logic surprisingly consistent with Liu Qiangdong's - AI is not a knife for cutting jobs, but a ladder for upgrading people.

These two giants representing the tech industries of China and the United States have rarely united on the "AI and employment" issue, which is like a difficult test question.

01

AI is here. What about our brothers?

Liu Qiangdong and Jensen Huang have both told the same story of not laying off employees.

In November 2025, at NVIDIA's all - staff meeting, an employee was worried about whether AI would make them lose their jobs and asked the boss a question. Jensen Huang then promised on the spot: "I guarantee to you that you will still have jobs."

Jensen Huang doesn't just talk the talk. The facts prove it: NVIDIA not only didn't lay off employees because of AI, but also added thousands of new employees last quarter. There aren't even enough parking spaces. He also added: "To be honest, I think we may still be short of about 10,000 people."

This sounds counter - intuitive. Isn't AI supposed to improve efficiency? If so, why are there more shortages of workers?

On May 26th, 2026, when Jensen Huang was interviewed by Singapore's AsiaNewsNetwork, he directly criticized those CEOs who linked layoffs to AI, saying it was a "lazy" narrative. He asked a question that no one can avoid: AI has only become efficient and useful in the past six months. How could people have laid off employees because of AI two years ago?

This statement is quite sharp because it exposes a convenient excuse. When a company wants to cut costs, adjust its organization, or clean up inefficient businesses, AI can easily be packaged as a decent reason. "It's not that I want to lay off employees; it's that the times have changed." "It's not the company's operating pressure; it's that AI has improved efficiency." "It's not a choice made by management; it's a natural replacement of technology."

Jensen Huang said that such statements are "too perfunctory." Some executives do this just "to seem smart," and "I really hate this kind of behavior."

His core view is that AI improves productivity. After productivity is improved, enterprises have two choices. One is to do the same work with fewer people, thus laying off employees; the other is to do more work with the same number of people, thus expanding.

Technology itself doesn't automatically determine which result will occur. It's the management that truly decides the outcome.

For companies with imagination and ambition, the efficiency improvement brought by AI is not a reason for layoffs, but fuel for expansion. Since efficiency has increased, more businesses can be carried out, more markets can be entered, and more users can be covered, which requires more people, not fewer.

NVIDIA's own practice is the best proof. Within a year, the total number of employees has increased from 29,600 to 36,000, a net increase of more than 6,000. And Jensen Huang also said that they are "still short of about 10,000 people." This is not a contraction in the face of AI, but an expansion.

So Jensen Huang said that laying off employees in the name of AI only shows that these CEOs "lack imagination." When they see efficiency improvement, their first reaction is "we can hire fewer people"; while Jensen Huang's first reaction is "how much more can we do and how many more people can we hire."

The difference between these two types of management is essentially not a difference in technical judgment, but a difference in ambition and imagination.

Liu Qiangdong's main logic has three points: "JD.com will still be one of the companies with the largest number of employees in China 20 years from now," "not laying off any front - line employees replaced by machines," and "minimizing the impact of new technologies on more than 500,000 blue - collar workers and their families."

This statement may sound like brotherly affection, but in fact, it is completely in line with Jensen Huang's logic.

JD.com's "Nirvana Project" has built more than 80 robobases across the country, teaching blue - collar workers to repair robots, do maintenance, and manage intelligent warehouses. This is not charity; it's a judgment: When AI and robots come, JD.com won't choose to "do the same work with fewer people," but to "do more and more advanced work with the same number of people."

Liu Qiangdong said that "the work our brothers do is not what humans should be doing" - he admits that working in all weather is hard work, but the solution is not to replace people with machines and then dismiss them. Instead, it's to use machines to free people from hard work and then train them to operate the machines. Blue - collar workers become robot repairmen and intelligent warehouse managers, with higher incomes, stronger skills, and higher irreplaceability.

This is using AI for expansion, not contraction.

One is in the chip industry, and the other is in e - commerce logistics. Although they have different starting points, they reach the same conclusion: When AI comes, companies with imagination will keep people, upgrade them, do more things, and hire more people.

They are calculating the same account. Jensen Huang thinks that with the support of AI, the business scope of NVIDIA can be much larger than it is now, so they are short of 10,000 people. Liu Qiangdong believes that with the support of AI and robots, JD.com's fulfillment ability and service scope can be expanded, so 900,000 employees are not the end.

Jensen Huang's evaluation of those CEOs who lay off employees in the name of AI is "lacking imagination." In other words, they only see the cost, not the opportunity. They are calculating "how to spend less money," not "how to earn more money and then distribute more."

This is the fork in the road for two types of bosses.

02

AI is here. How to calculate the number of employees?

Of course, reality is not all warm - hearted.

If Liu Qiangdong and Jensen Huang represent the idea that "AI should amplify human potential," then Mark Zuckerberg and Meta show another colder path: After AI enters the company, the labor cost will be recalculated.

Mark Zuckerberg publicly made a very impactful judgment in January last year that AI can soon undertake code work similar to that of a mid - level engineer. This statement shocked the tech industry.

In the past, programmers were considered one of the high - skill positions that were least likely to be replaced. Especially in Internet companies, engineers are often the most core, most expensive, and most protected group. However, when large - scale models can start writing code, fixing bugs, doing tests, generating documents, and even completing some relatively complete development tasks, the value chain of software engineer positions begins to be disassembled.

Which work is creative architectural design? Which work is just repetitive code implementation? Which work requires the judgment of senior engineers? Which work can be handed over to an AI agent to run first?

Once a company starts to disassemble positions like this, "people" are no longer seen as a whole but are broken down into tasks, processes, and costs.

Meta's recent rounds of adjustments have occurred in this context.

On the one hand, there is a huge investment in AI. Computing power, data centers, model training, and top - notch AI talents all require money, and the cost is getting higher and higher. On the other hand, there is the labor cost. In the past, the most expensive thing in Internet companies was people, but in the AI era, the most expensive things may be becoming GPUs, data centers, and model teams.

When Meta invests more resources in AI infrastructure, it will inevitably re - examine the labor budget. So, organizational adjustments, job mergers, low - performance eliminations, and transfers to AI workflows will happen simultaneously. This is what really deserves attention in the Meta case.

It's not simply "AI replacing humans," but "AI changing the company's cost structure."

In the past, the expansion logic of an Internet company was to hire more people, develop more products, and cover more users. Now, a new logic is emerging: Use stronger models, greater computing power, and fewer but more capable people to do what a large team used to do.

For the company, this is an efficiency revolution. But for employees, it means the collapse of their sense of security.

The rumors about NetEase in China also hit the same anxiety.

In March, the rumor of layoffs in NetEase's game outsourcing positions detonated social media. A response of "normal business adjustment and personnel replacement" combined with the industry's hot discussion of "the localization of AI horror stories" made the topic of AI replacing human labor in the game industry continue to ferment.

Although NetEase has denied the statement of "using AI to lay off all outsourcing employees," it also admits that there have been normal business adjustments and personnel replacements in some projects recently and plans to gradually let some outsourcing employees in basic - skill positions leave.

This statement needs to be viewed from two aspects. It doesn't mean that NetEase really "uses AI to lay off all outsourcing employees," as this statement has been denied by the official. But it also shows that basic - skill positions, outsourcing positions, and project - based positions are indeed the first groups to be re - evaluated.

Why are outsourcing employees the ones being laid off?

Because outsourcing positions are usually farther from core business decisions, have weaker bargaining power, and the job tasks are easier to be standardized, streamlined, and project - based. Once AI tools can improve the efficiency of processes such as original painting, modeling, animation, audio, levels, and testing, companies often consider the external production capacity pool first, rather than full - time employees.

This is the real order of AI replacement.

It doesn't necessarily replace the CEO right away, nor does it necessarily replace the core R & D leader first. It usually impacts those positions that can be described by processes, assisted by tools, and outsourced as projects first. In other words, AI doesn't impact everyone equally. It will first impact those on the edge of the organization, those with low bargaining power, and those who haven't received long - term protection commitments from the company.

This is why the NetEase rumor has resonated in the industry.

What people really fear is not whether a certain company will lay off outsourcing employees, but a clear trend: When AI tools enter a mature industrial chain, companies will start by making cuts from the people who are easiest to target. Outsourcing, art execution, basic testing, operation review, customer service, junior code implementation, and low - complexity content production - these positions may all be re - priced.

This is not a problem of a single company, but a "dividend" brought by the development of AI as a whole.

When we look at Zuckerberg's approach and the NetEase rumor together, we can find a common logic. In the AI narrative of these companies, the question is not "how to upgrade people," but "do we still need so many people" and "which people can be replaced in a cheaper way."

This is completely different from the path of Liu Qiangdong and Jensen Huang. One path is: Upgrade people and then expand together. The other path is: Disassemble positions and then re - price labor costs.

The technology is the same, but the choices of management determine whether people are regarded as assets to invest in or costs to optimize.

This article is from the WeChat official account "Super Focus Foci", written by Sean, and is published by 36Kr with authorization.