Silicon Valley AI Transformation Record #1: Organizational Changes Triggered by Layoffs at Silicon Valley Tech Giants

This is not an era where there are increases of 10% or 20%.

"Silicon Valley's AI Transformation Chronicles" is a brand - new observation series initiated by the Tencent Research Institute. AI has begun to reconstruct every aspect of our work at the fundamental level. This "AI revolution" is not merely a simple upgrade of production tools but a profound transformation regarding production relations, collaboration methods, and value creation.

Yuan Xiaohui and Yu Yi from the Tencent Research Institute will join hands with Silicon Valley consultant Chen Ran and more industry pioneers to delve into the front - line of innovation and change in Silicon Valley, focusing on two core points: First, how AI, as a fundamental capability, penetrates and reconstructs our work, creation, and competition. Second, how different groups, especially pioneer enterprises and individuals in Silicon Valley, create new paradigms of human - machine collaboration and how they adapt to or even lead this transformation. We not only care about what is happening but also why it is happening and where it is heading.

[Highlight Quotes]

Layoffs and restructuring have been very hot topics in Silicon Valley in the past two years. When we look at the core plans of these companies, we will find that this is not a short - term situation but a long - term, systematic, and drastic change that is taking place.
The era of major restructuring has arrived. This is not an era of a 10% or 20% improvement but an era where you need to conduct extensive restructuring, reconstruction, and re - thinking around AI.
AI enables an enterprise to return to a more fundamental thing, which is manufacturing and sales. A large number of other functions or organizations are being outsourced and tooled. Enterprises are returning to spending more time on manufacturing and selling products.
What is the most important thing for a person? In the past, people might think it was ability and technology, but now it has become agency, that is, subjective initiative. The bigger the things you want to do, the greater the positive impact of AI on you.
We are moving away from the employee system and starting to enter a larger - scale partnership system. If a company can implement a so - called partnership design or a clearer system of rights, responsibilities, and incentives within the company, its growth will also be very rapid. Because it's like training a large - scale model. Once you set the benchmark, it will progress rapidly.
This kind of unicorn small - team is becoming more and more common. People will consider making money a more important thing than financing. The era of saying 'we are short of a programmer' is over. We won't hear such a statement anymore. Now what people lack is how to make money and how to find customers.
A good AI - native organization should send out a very safe and even encouraging signal. Everyone can confidently say that 70% of this work is done by AI, and I am exploring a workflow where AI can do 50% to 70% of the work, and I am willing to promote it.
If you dare to think about and build a company worth 100 million or one billion today, something you only dared to think about in the past, it is really possible to achieve it today.

Introduction to Guests of This Episode:

Chen Ran is based in San Francisco, Silicon Valley. He is the CTO of Pure Global (AI helps medical technology go global). He was the former director of machine learning at Tubi TV and has decades of experience in AI implementation and development. He is an excellent answerer in the fields of machine learning and artificial intelligence on Zhihu.

Yuan Xiaohui is the director and senior expert of the Innovation Research Center at the Tencent Research Institute. He has long been concerned about the impact of artificial intelligence on the economy and society, as well as the innovation models of organizations, industries, and cities. He is a member of the AI Governance Alliance of the World Economic Forum.

Yu Yi is a senior researcher at the Tencent Research Institute, mainly researching AI - native product innovation and corporate transformation. He has many years of experience in venture capital and ecosystem incubation. He is an annual expert on LinkedIn China, an annual outstanding AI expert and excellent sharer at Tencent, and a tutor in the AI learning circle on Get.

Silicon Valley's Major Restructuring: Behind the Layoff Wave is an AI - Driven Systematic Transformation

Yu Yi: Xiaohui, Chen Ran, and I have talked a lot about work reshaping and the transformation of large companies before, which is also a topic we have been very concerned about recently. I'd like to invite Chen Ran to introduce some strange situations in Silicon Valley. For example, Microsoft and Salesforce have good financial reports, but at the same time, they are announcing large - scale layoffs.

Chen Ran: This is a very objectively existing phenomenon. When people talk about Silicon Valley, their first reaction is how AI is changing and innovating. But in fact, on the other hand, we have seen that large companies have been undergoing systematic and large - scale restructuring and layoffs in the past two or three years, especially after the pandemic. On the one hand, this is due to the previous economic impact, and on the other hand, it is also because AI has brought about some systematic changes.

So, layoffs and restructuring have been very hot topics in Silicon Valley in the past two years. How to lay off employees, how to restructure, and these issues are basically discussed together with 'how to introduce AI'. I have many friends and classmates who are in management positions in large companies. Many of the core topics they discuss are how to implement AI, how to measure the effectiveness of AI, and how to really make some changes. At the same time, they also need to discuss how to make organizational changes.

This is actually very common in large companies. Some people even think it is a short - term phenomenon. But in fact, when we look at the core plans of these companies, we will find that this is not a short - term situation but a long - term one where restructuring and layoffs are carried out with specific targets every year for many years to come. So, this is a drastic change that is taking place.

Yuan Xiaohui: Around 2022 was a period of concentrated layoffs in Silicon Valley. Before the pandemic, many junior and middle - level engineers were recruited, resulting in many positions not actually requiring that many people. So, that round of adjustment was actually a kind of 'leaning down'. But from the second half of last year to this year, the number of layoffs in the entire Silicon Valley has been increasing. There is information that more than 90,000 people were laid off in 2024, and 80,000 people were laid off by August 2025. Do you think this is caused by AI or the aftermath of over - recruitment before?

Chen Ran: This is a very good question. In terms of attribution, it is really hard to say whether it is completely due to economic reasons or brought about by AI. But we have indeed seen many changes at the organizational structure level of companies, including the expectations for employees, the selection of talents, and the changes in compensation.

Let me give a few obvious examples. A company like Facebook (Meta), which has always been famous for its 'Move Fast and Break Things' culture, proposed to cancel or reduce middle - level management early in its reform. In the past, there might have been many situations where directors were nested within directors and VPs were nested within VPs, but now it has been drastically reduced. Front - line managers have to make a choice: either leave or go back to be programmers and write code again. This is equivalent to a large - scale 'leaning down' of the middle layer. They believe that in today's technological situation, there is no need for so many middle - level managers. If a person manages a team, they should manage a larger team.

Another example is also from Meta. Recently, Meta hired people at extremely high salaries to form a super artificial intelligence organization for a very ambitious plan. Each person hired is very expensive. The internal discussion is that to 'support' such a person, they may have to lay off 100 or even 1000 people to free up the salary for this person. The logic behind this is that the value that many core talents can generate in the AI model may be far higher than the so - called 'ten - times engineer', but rather 'hundred - times engineer' or 'thousand - times engineer'. So, in this situation, it will also bring about a large number of organizational structure adjustments and layoffs.

Yu Yi: In the case of companies with good performance and simultaneous layoffs, is it because of talent replacement or for future preparation, using good performance as a foundation for adjustment?

Yuan Xiaohui: Yes, in fact, there is not a strong inevitable connection between performance and the logic of layoffs. I think management mainly sees the pressure from this wave of AI. For example, Microsoft has good performance, but its recent layoffs are quite significant. Why are many middle - level employees laid off? It may be thought that having such a position in the organization does not seem to bring additional increments. When the business is expanding, people are needed to charge forward, but now that the business has stabilized, there is no need for so many managers. Of course, what we are discussing this time is to see what kind of impact it is from the perspective of AI.

The New Work Paradigm in the AI Era: Towards Independence, Standardization, and 'De - Middle - Leveling'

Yuan Xiaohui: Is it because of the emergence of AI that middle - level management is no longer needed? Why do people now think that management should be flatter?

Chen Ran: Yes, I think flattening is an objectively existing trend. Let's take the example of food delivery. A regional manager can manage many food delivery workers. Why? Because in fact, they don't need much management. It is more the algorithm platform that assigns orders. In the past, an engineer team or a more complex team of knowledge workers had many non - standardized work models and needed to discuss together. So, a team would have a certain scale. For example, in the United States, a manager might lead three to five people, no more than ten.

But now, on the one hand, AI tools have improved communication efficiency, and on the other hand, many people's work has been more severely standardized. After the delivery and handover with others become more standardized, there is no need for as much communication as before. People can do things relatively independently, and the dependence on each other is not as strong. Then one person can lead a larger team. This is a relatively obvious change.

Yuan Xiaohui: This is equivalent to the fact that in the past, so many middle - level managers were needed in an organization because of the 'transaction cost' - they needed to report and convey information, and this transaction cost needed to be borne by people. Now, with the emergence of many communication tools, digital tools, and even AI tools, the transaction cost of communication is decreasing. One person can master more information within the organization in a shorter time and can manage more people.

Chen Ran: Yes, and what each person does is more independent than before.

Yuan Xiaohui: Then let's go back to the bottom of the pyramid. Just now we said that the middle - level is shrinking, and top - tier talents are being scrambled for, forming a 'barbell structure'. But for the lower level, that is, junior engineers or entry - level positions, it seems to be very difficult this year. Many graduates have difficulty finding jobs in the technology industry. What do you think?

Chen Ran: Enterprises are profit - seeking by nature and definitely hope that a person can start working and generate commercial value as soon as they come. So, from this perspective, not only junior engineers but also people in all positions are having a hard time. This hardship comes from two aspects. On the one hand, our education is a bit backward. We train a lot of skills, but these skills are actually useless in the face of the AI era. You work hard to learn a lot of things in school, but when you start working, you find that most of the work requirements are not these. So, why should the enterprise hire you?

On the other hand, for enterprises, they actually desire people to come in and contribute new ideas and generate greater commercial value. But I have seen a big change. Many smart students start entrepreneurship earlier and generate commercial value earlier. They don't need to rely on a corporate position to prove themselves. This process, whether it is entrepreneurship or running a business, is becoming more and more youthful and independent. I think this is a kind of impact on our existing 'employment' system, and this impact may be quite large.

From 'Short of a Programmer' to 'How to Make Money': AI Coding Brings Business Value Back to the Center

Yu Yi: When we shared our research on AI Coding before, we also saw some data, such as the rising unemployment rate of computer science graduates. AI Coding has played a very important role in this wave of work restructuring. Chen Ran, from your perspective, what new consensuses or controversial 'non - consensuses' has AI Coding brought?

Chen Ran: I can share some interesting observations. The Hackathon in Silicon Valley has always been a very popular culture. Many people gather on weekends to work on interesting projects. About a few years ago, the mainstream model of Hackathon was still programmers and non - programmers getting together to form teams. One side lacked ideas, and the other side lacked people to implement them.

But if you go to a Hackathon this year, it's completely different. The mainstream no longer involves pairing. People come and use AI Coding to implement their ideas by themselves. People gather together more to exchange ideas about the pitfalls in the AI Coding framework, how to solve technical, business, or tool - related problems, and most importantly: marketing, how to acquire customers, and how to make money. No one says 'I'm short of a programmer' anymore. The era of 'just short of a programmer' has passed.

So, what do people lack? It comes back to the most fundamental question: How can I make money? How can I find customers? Now the mainstream discussion has returned to a more business - oriented, fundamental, entrepreneurial, and core question. This is a big change. We have moved away from the discussion about technology and tools and returned to business value itself. This will be a very painful thing for all employees because the technology, tools, and even the advanced skills of 'climbing the ladder' in large companies that people usually focus on at work are not that important in the real business environment.

Yu Yi: A similar situation has also appeared in China. I have a friend who is doing low - code entrepreneurship. The trend change in their company is very similar to what you just described. First, the team has indeed made some shrinkage and adjustments. They were originally doing low - code development, and sales and development were originally separate. Now they use AI to call a large number of pre - developed components, and the speed of creation and development is much faster than before. So, they require developers, and the compensation system has also changed: a basic salary plus a direct commission after sales. This requires developers to also stand at the front line to complete sales work.

A European company called Rekki. Its CTO shared another reform idea, which has several interesting points. First, the role of developers has changed. In the past, developers quickly implemented requirements once they received them. Now they are required to take a step back and think about how to empower business personnel to develop on their own by developing components. Second, the boundaries between many roles such as sales, product, and development are becoming blurred, and there are more and more generalist - type roles. This seems to be a trend that some leading people in Chinese and American startups are moving towards.

The Maze and Path of Enterprise AI Transformation: Is the Partnership System the Optimal Solution in the Future?

Yu Yi: Startups are flexible and can change quickly. But will there be any changes in large companies? Dr. Xiaohui, you and Professor Yang Guoan have recently been working on a project on AI transformation and interviewed some leading Chinese enterprises. How exactly are they doing AI transformation?

Yuan Xiaohui: From the perspective of AI transformation, the progress is still in a very early stage. Even some industry leaders are mostly considering things at the productivity level, that is, how to apply AI to their business. In fact, very few are really considering organizational change. It may involve some department mergers or how to cooperate closely with the AI department, but no disruptive changes in the entire organization have been seen.

Currently, there is an experience that if you want to start a new business, a relatively easy way is to establish a new department or even a new company, gather people who are very interested in both AI and the business, and let them collide and rethink how to reshape the organizational process. Because making adjustments in an existing institution is bound to involve many interests, so it is very difficult.

Another way is for the founder to strongly promote it from the top - down, requiring all employees to embrace AI, conduct a lot of training, and let everyone reconsider the workflow from the perspective of 'AI First'. Of course, there are also bottom - up attempts. Some 'preachers', people who are passionate about this direction or think this may become a key node in the future, promote some organizations to make transformations and adjustments from the bottom -