A profit of $18 billion was exchanged for 14000 termination letters.
This is the Asura arena that Amazon is currently staging. The stock price of this tech giant is approaching a historical high, but the big blow of layoffs has quietly fallen.
Moreover, this is just the beginning – the final number of layoffs could reach up to 30000, accounting for nearly 10% of the total number of employees.
99% of Amazon Game Studios San Diego employees laid off
Someone is still sunbathing on the beach in Mauritius when a notification pops up on their phone: You have been fired.
The party who was laid off was surprised, and the spectators were even more surprised.
Why did Amazon suddenly resort to such brutal layoffs after performing well in the past quarter?
The official reason given is the same as the companies that have laid off employees in the past period, because of AI.
Beth Galetti, head of human resources at Amazon, wrote in a memo to all employees, “AI is the most revolutionary technology we have seen since the Internet.”.
She explained that precisely because AI can enable companies to innovate faster, Amazon needs a “more streamlined organization with fewer levels”.
Listening to it is a familiar language, using AI can reduce costs and increase efficiency, with fewer levels. The reason behind Amazon’s layoffs this time is not just because AI can replace these employees.
Operate like the world’s largest startup company
Over the past decade or so, the hallmark of a large company has probably been having a large workforce and a high output, with each department having one or even multiple of their own products.
But the logic of the AI era has directly become more computing power and fewer people. Like OpenAI, the valuation is approaching trillions of dollars, and the total number of people in the company has only increased from a few hundred in 2023 to about four thousand now.
At Amazon, this transformation is even more from top to bottom, with clear pricing. CEO Andy Jassy is systematically building a narrative of ‘AI changing everything’.
Jassy’s core philosophy is that he wants Amazon to operate like the world’s largest startup. He will spare no effort to simplify, reduce bureaucracy, and eliminate hierarchy; The official layoff announcement perfectly echoes this point.
This’ AI streamlined narrative ‘is not groundless, it even accurately targets a specific group of people: middle-level managers.
According to internal data obtained by Business Insider, over 78% of the laid-off employees in the first batch of US layoffs were managers at the L5 to L7 levels.
The departments involved include approximately 15% of positions in HR (People Experience&Technology); Equipment and Services Department (Echo, Fire, Kindle teams); Audible Audio Department; And some AWS enterprise support teams.
According to Levels.fyi data, the total salary for an L5 in 2025 is approximately $120000, $210000 for an L6, and $280000 for an L7
These middle-level managers’ daily work, in summary, mostly involves coordinating cross departmental communication, summarizing information, writing status reports, and drafting memos.
It sounds like this is currently the task that generative AI excels at the most. Amazon seems to be using actions to prove that AI is hollowing out the middle layer of the enterprise.
One of the hardest hit areas of layoffs is the equipment department (responsible for products such as Kindle and Echo). In an internal letter after the layoffs, Tapas Roy, the vice president, bluntly requested that the remaining employees embrace AI (lean in on AI) to improve efficiency.
On the surface, this is a standard story of ‘AI improving productivity leading to human white-collar unemployment’. But even Amazon’s own people may not fully believe this script.
Is AI the real culprit or the scapegoat
The biggest doubt about this layoff is: Has AI really become powerful enough to immediately replace 30000 white-collar workers?
The answer is likely to be negative.
The most ironic thing is that, according to CNN’s report, an Amazon spokesperson privately admitted that AI was not the reason for the “vast majority” of the 14000 layoffs announced on Tuesday.
This statement has also received support from several management experts. They mentioned that there is currently a trend of “AI washing” in the entire technology industry, which roughly means packaging anything as AI.
From July 2024 to June 2025, the top 15 most commonly used keywords in job postings related to artificial intelligence are AI, rather than specific AI skills such as computer vision and natural language processing
Very few companies are really replacing people with AI now, “Jessica Kriegel, Chief Strategy Officer of executive consulting firm Culture Partners, told CNN. She believes that what we are seeing now is more like a form of ‘preemptive layoffs’.
In other words, Amazon is laying off employees not because AI is ready to take over human work, but because executives want to free up the “financial runway” to explore the possibilities of AI.
After all, for now, in addition to efficiency, AI also brings a lot of “workrolls”, which refer to the “work garbage” generated by chatbots that seem like something but are actually worthless and even full of errors.
The work waste generated by AI, such as slides, summaries, code, etc., requires more time and effort from humans to process. A survey has found that 40% of AI participation is work waste, with each piece of waste wasting employees two hours and costing $186.
Recently, MIT also released a study stating that 95% of enterprise AI investments currently have zero returns.
From pilot to practical application, the introduction of AI (light blue) for specific tasks has significantly decreased, with only 5% successful implementation
Although AI has the ability to handle simple recording tasks, it still has no way to handle some details and high accuracy requirements in the work.
Amazon itself acknowledges that the company is still recruiting and will provide 90 days for most affected employees to search for new positions within the company.
If AI can really systematically replace these positions, why bother with internal job transfers? Obviously, the story of AI “replacing” humans is not convincing enough. So, what is the real motivation?
The AI arms race is too costly
If AI is not the direct culprit, then it is likely to be the motive. The first real reason for this layoff may be related to AI, but not because it is cheaper to use AI, but because AI is too expensive.
Amazon is currently in a trillion dollar, winner takes all AI war. In this war, it must desperately chase after its opponents.
By 2025, Amazon Web Services (AWS) will hold a leading market share in the market
Amazon’s core profit source, AWS (cloud services), has fallen behind Microsoft’s Azure and Google Cloud in terms of growth rate. Amazon needs to quickly release a large amount of cash to recruit expensive AI experts and build data centers.
This expense is rising at an unprecedented rate. In 2024, their expenditure on server and data center construction has already exceeded 85 billion US dollars. It is expected to increase by another 40% to reach 120 billion by 2025.
Cutting 1% of the white-collar salary budget means the company can buy several thousand more H100 GPUs.
Analysts pointed out sharply that “Amazon is facing pressure in the short term and needs to offset long-term investments in building AI infrastructure.
The ChatGPT and Claude we use strongly rely on these cloud services; Claude mainly relies on AWS, Gemini is its own Google Claude, and OpenAI is mainly supported by Microsoft Azure. However, Anthropic also announced a partnership with Google a few days ago to use their TPU; OpenAI also uses AWS cloud services.
Now, the logic of layoffs is clear. Amazon is cutting off bloated management in mature businesses such as retail, as well as employees in non core profitable departments such as HR and equipment, converting their billions of dollars in wages into “ammunition” to support the AI arms race.
These laid-off L5-L7 managers were not replaced by AI, they were ‘squeezed’ by the cost of AI.
This is a cruel but inevitable business decision: trade today’s people for tomorrow’s computing power.
It’s only because we recruited too many people during the epidemic that we were able to be run on
If raising funds for AI is a strategic consideration for the future, then there is another more awkward and past oriented real reason for this layoff, which is to correct Amazon’s excessive recruitment during the pandemic.
Yale University economist Ernie Tedeschi directly said, ‘I think this is mainly a correction of the dynamics during the pandemic, rather than new things like AI impacting these companies.’. ”
Data does not lie.
During the COVID-19 e-commerce boom in 2020 and 2021, Amazon excessively recruited and hired hundreds of thousands of employees. The total number of its global employees (including warehouse and corporate staff) has almost tripled from 2017 to 2024, reaching over 1.6 million at one point.
Now, as e-commerce demand slows down and the economy enters an uncertain period, Amazon has found itself employing too many ‘surplus’ employees. This company is starting to pay for the crazy expenses of the past and optimize costs.
After all, while cutting off the fastest-growing white-collar and managerial positions during the pandemic, it is also recruiting 250000 seasonal warehouse workers on a large scale. Just invested $1 billion to increase the average hourly wage of warehouse workers to over $30.
On one hand, we are increasing salaries and hiring blue collar workers, while on the other hand, we are laying off white-collar managers. The wave of productivity replacement triggered by AI has not come as fast as imagined, but its potential impact has begun to make these large enterprises walk on thin ice and cannot take another wrong step.
Blaming layoffs on AI is a win-win situation for the company. It has packaged these past wrong paths (a traditional cost cutting behavior caused by poor management and excessive recruitment) into a forward-looking, future oriented technological transformation.
Image source: https://x.com/KobeissiLetter/status/1983209896721236139
Although it is different from the traditional narrative that layoffs are necessary due to the high efficiency of AI, they are actually discussing the same thing.
The tech giants led by Amazon, Microsoft, Meta, and Google are in a life and death AI arms race. The ticket to the competition is to build a data center, purchase sky high priced GPUs, and develop basic models.
And Amazon’s layoffs are precisely because it needs to free up the salary budget of these 30000 people to pay for the huge bill of building AI infrastructure.
Another group of companies that are also laying off employees, UPS (United Parcel Service), Target (American retail giant), and even Lufthansa, are doing so precisely because AI is already working.
These companies do not need to bear the sky high research and development costs themselves. They only need to rent mature AI solutions from platforms such as Amazon Web Services (AWS) and Microsoft Azure, such as automated customer service, supply chain optimization, and streamlined management levels.
These two motivations for layoffs, one is to buy AI for the future, and the other is to use AI now, both feeding the same economic giant – computing power.
It is evident that AI is creating wealth, which is flowing at an unprecedented rate towards computing power rather than labor.
Semiconductor companies, such as NVIDIA, sit at the top of this value chain, collecting rent from everyone.
The world’s first company with a market value exceeding $5 trillion, Nvidia, only surpassed $4 trillion in July
For laid-off employees, regardless of the reason, the outcome is the same.
暂无评论内容