The First Wave of Tech Giants' AI Layoffs Are Already Returning to Work
- Core Viewpoint: Recent layoffs by some tech companies under the guise of AI-driven efficiency gains, followed by rehiring, reveal the practical challenges AI still faces in fully replacing human labor—such as high costs and organizational integration. The so-called "AI layoffs" may often serve as an excuse for cost-cutting or strategic realignment.
- Key Elements:
- After Block laid off over 4,000 employees, some staff (e.g., engineers, HR) were rehired due to "documentation errors" or the critical nature of their roles, indicating AI cannot seamlessly take over all tasks.
- Enterprise-level AI applications are costly. For instance, individual use of the Claude model can cost up to $6,000 per month, while training an AI customer service agent capable of handling complex tasks far exceeds the cost of low-wage human labor.
- The "Jevons Paradox" exists: after AI improves individual efficiency, companies may increase employee workloads rather than reduce headcount, leading to heavier work burdens.
- AI cannot comprehend or integrate into a company's informal organization and "internal dynamics." Layoffs may weaken organizational collaboration and risk-taking capacity rather than simply optimizing efficiency.
- NVIDIA CEO Jensen Huang criticized management for lacking foresight in using AI as a pretext for layoffs, arguing that AI application should focus on business expansion, not workforce reduction.
- Cases similar to Twitter's post-acquisition layoffs and rehiring suggest that "AI layoffs" can sometimes be a smokescreen for cost reduction or business adjustment, not an inevitable outcome of technological maturity.
Original | Odaily (@OdailyChina)
Author｜Golem (@web3_golem)
The first batch of employees laid off by AI are already returning to work.
On February 27th, Jack Dorsey's fintech company Block laid off over 4,000 employees at once, reducing its total workforce from 10,000 to less than 6,000. Jack's reason for the layoffs was that "AI tools have changed everything." While it's a societal consensus that AI will eventually eliminate certain professions, the fact that it's first replacing white-collar workers in mid-to-high-level roles has intensified workplace anxiety. (Related reading: At Jack Dorsey's Company, 4,000 White-Collar Workers Are Being Replaced by AI)
However, less than a month later, some of the laid-off employees have already received offers to return...
According to Business Insider, these recalled employees come from various departments, including engineering and recruitment. A design engineer at Block posted on LinkedIn, stating that leadership told him he was laid off by mistake, a "clerical error." An HR employee, in a now-deleted post, said they were only recalled after their manager persistently advocated for them. Others mentioned receiving a call from Block out of the blue a week after being laid off, asking them to come back.
Jack has not publicly responded to the recalls. Proportionally, these recalled employees represent only a tiny fraction of the initial layoffs, but it perhaps illustrates a point: for some roles and tasks, AI isn't as effective as humans.
First, from a cost perspective, an enterprise-grade AI "employee" is certainly more expensive than ordinary human labor.
Hiring people costs money; using AI costs tokens. Claude Opus 4.6's standard base price is $5 per 1 million input tokens and $25 per 1 million output tokens. Domestic large language models are cheaper; Qwen3.5 plus's standard base price is 0.8 RMB per 1 million input tokens and 4.8 RMB per 1 million output tokens.
Taking the recently popular OpenClaw as an example, a senior "shrimp farmer" within Odaily stated that using OpenClaw merely as a life and research assistant for just over a month burned through about $6,000 worth of tokens (using Claude 4.5/4.6 models). $6,000 a month—what kind of highly educated professional couldn't you hire with that (outside of Europe and America)?
If personal use is this costly, integrating AI into enterprise workflows is even more expensive. Take the simplest example of replacing customer service. In regions with degree inflation, you can hire a good-looking college graduate as a customer service representative for 3,000 RMB. But training an AI customer service agent that can truly replace a human, handle complex tickets, access multiple knowledge bases, conduct multi-turn conversations, and remain stable online—that cost is definitely not covered by 3,000 RMB per month.
In 2024, the Swedish payments company Klarna high-profile laid off over 1,000 people, claiming its AI customer service could handle the workload of 700 human agents. But in May 2025, Bloomberg and other media outlets reported that Klarna had started rehiring human customer service staff, with its CEO admitting they had indeed "moved too fast" with AI.
Furthermore, AI replacing human labor also faces the "Jevons Paradox."
The Jevons Paradox is an economic concept stating that efficiency improvements don't necessarily lead to reduced use of a resource. Instead, because the cost of use decreases and demand expands, total consumption may rise. Applying this theory to the AI-era workplace means that when AI technological advancements improve employee efficiency, companies won't allow employees to rest; they will instead demand they complete more tasks within the same timeframe.
So-called efficiency gains become another, more hidden form of increased burden. The idea of AI liberating human labor is a complete illusion.
Capitalists also believe that in the AI era, companies don't need as many employees, as Jack said, "smaller teams with more intelligent tools." But in reality? The current situation is that after layoffs, the original work isn't entirely inherited by AI; rather, the remaining employees take on increased workloads with the help of AI.
If it were just simple work tasks, that might be one thing. But it's crucial to remember that, ultimately, a company is a human organization. Where there is organization, there is "politics." AI can integrate into a company's formal structure, but it can never understand or integrate into its informal/invisible structure.
Therefore, when AI-driven layoffs occur, they cut not just labor but organizational muscle. The remaining employees not only shoulder heavier workloads but also absorb the anxiety, risk, and responsibility that came with the eliminated positions. There are fewer people to collaborate with, fewer people to execute, and most importantly, fewer people to take the blame.
During NVIDIA's GTC 2026, Jensen Huang criticized companies that use AI efficiency as a reason for layoffs in an interview. "Leaders who resort to layoffs in response to AI simply can't think of better ideas. They have nothing new in their minds. Even with the most powerful tools, they won't use them for expansion," were Huang's exact words.
Jensen Huang's point is that AI isn't here to eliminate employees but to help companies expand and develop new businesses. Instead of layoffs, they should increase hiring. If management doesn't realize this, they are fools. But jokes aside, managers in companies are often the shrewdest of the shrewd. They are certainly aware of AI's current high costs and the continued necessity of human labor.
For tech company layoffs, perhaps AI is just a pretext; cost-cutting is the real goal.
AI has become a universal excuse for tech company layoffs. In reality, what AI is truly eliminating is not individuals, but those enterprises and business models still stuck in the old era. When companies fail to keep pace with AI advancements, leading to stagnant business growth and shrinking profits, the AI revolution instead becomes a new tool for companies to pressure employees—reducing headcount, cutting costs, piling more work onto those who remain, and then making each person reflect on why they haven't become someone more adaptable to the AI era.
If they unfortunately cut a vital artery, they can quietly invite it back. This layoff method is also common in Silicon Valley. After Elon Musk completed his acquisition of Twitter in October 2022, he laid off about half the employees (over 3,000 people) in early November. Subsequently, due to mistaken layoffs or realizing key positions couldn't function without people, he rehired dozens of the laid-off employees.
Returning to the present, in the end, AI will change many things, but it's not so magical that it can compensate for a company's strategic sluggishness, business aging, or managerial laziness. Whether the reason behind being laid off by AI and then recalled is that the company realized some work doesn't just disappear with a statement like "AI changed everything," or it was merely a cost-cutting excuse, the situation isn't inspiring, nor is it a dramatic reversal.
It just shows us that before the future has truly arrived, some people have already been hurt by it in advance.
