Dan Schulman's statement sounds like he watched the scene from "The Matrix: Revolutions" again, where Neo fights against a swarm of Agent Smith. The CEO of Verizon said to analysts during the third-quarter earnings call, "I think that in the coming years, the world's population will be around 8.5 billion, but I believe that the number of agents could be at least 100 billion."

Of course, these are not the kind of sinister "agents" you see in Kennu Reeves' movies, who smile coldly and are determined to eliminate anyone who is not connected to the machines. Schulman continued, "We will ensure that unless we are 100% certain that it is indeed a customer and not an out-of-control agent, they will not be allowed to access our platform." He described agentic AI as an opportunity to Verizon increase customer spending and expand sales.

He believes that the number of so-called autonomous software programs will far exceed that of humans – the U.S. President considers these programs to be “super intelligent” – which could be unsettling for anyone who has tracked the predictions made by Schulman since he took over as the head of this American telecommunications company about a year ago. In April of this year, in an interview with The Wall Street Journal, he suggested that AI could push the unemployment rate to 30% within the next two to five years. If he is correct, the unemployment rate should have started to rise already, and may even have spiked by now.

Relevant: As the rebound sentiment around AI heats up, AT and Verizon lay off thousands more employees

However, so far, the accuracy of Schulman's predictions about the "end of the world" as predicted by AI seems to be similar to McKinsey's predictions regarding the internet API. Nearly three years ago, this consulting firm predicted that in the following five to seven years, telecom operators would see an additional $300 billion in sales due to these API factors stimulating usage. If there is no unprecedented surge in the next two years and four months, this will become one of the worst predictions since Winston Churchill's forecast that the Gallipoli campaign would be a piece of cake.

Despite what Schulman thinks, the unemployment rate in the United States did not change between April and September, remaining at 4.2%, according to data from the U.S. Bureau of Labor Statistics. Moreover, there is hardly any evidence to suggest that AI is improving productivity or is responsible for large-scale layoffs. Nevertheless, Verizon continues to lay off thousands of employees each year.

The Storming of the Bastille

By the end of 2025, Verizon had 9,700 fewer employees than a year earlier, due to Schulman announcing plans to lay off 13,000 staff after taking office. Subsequently, the number of employees was restored after the acquisition of broadband operator Frontier Communications at the beginning of 2026. Based on Frontier's previous employee base of around 13,000 people, Verizon's total workforce should have been approximately 112,600 by the end of 2024. However, as of June this year, the number of employees at Verizon was only 97,600. Three months later, according to the financial report released this week, that figure has dropped to 93,300.

Relevant: AI makes the wave of layoffs in the telecommunications industry seem irreversible

One day before the update release of Verizon, competitor AT also disclosed that they laid off another 3,700 people during the same three-month period. This means that these two large American telecommunications companies have collectively cut 8,000 jobs in one quarter, which is nearly 4% of their total workforce. As of September, the two companies employed approximately 220,500 people, which is even less than the number of employees AT had on its own in 2020. Since then, the combined employee count of the two companies has decreased by 142,460 people.

Layoffs over several consecutive quarters have become so commonplace that hardly anyone mentions them anymore. No analyst would think of asking how Verizon managed to increase its annual sales by nearly $10 billion between 2020 and 2025, while laying off about a third of its employees during that time. In 2020, Verizon generated approximately $970,500 in revenue per employee; by last year, that figure had soared to over $1.53 million.

If it weren't for the employees, investors might be delighted to hear that such a large-scale downsizing is entirely the result of the efficiency improvements promised by AI. In the past, layoffs were often seen as a sign that a company was in trouble. However, if AI can enable the company to operate with a much smaller number of employees than before while still generating the same or even higher revenues, then profits should increase, and shareholders will also benefit.

Relevant: Verizon Confirms Plans to Lay Off Over 13,000 Employees

Regrettably, there are many issues with this judgment. Firstly, AT &T and Verizon began to cut staff before anyone even heard of Sam Altman and AI becoming part of daily news. In fact, the largest round of layoffs occurred before November 30, 2022, which was also the day ChatGPT made its public announcement to the public – the "Storming of the Bastille" moment of AI.

In addition, compared to a few years ago, these telecommunications companies have not seen a significant improvement in profitability despite a substantial reduction in costs. Verizon recorded total operating expenses of approximately $25.7 billion in the most recent quarter, which is slightly higher than the $24 billion in the same period five years ago. AT &T's corresponding figures for these two quarters were both $28.2 billion. During this time, Verizon laid off 17% of its employees, while AT &T's workforce was reduced by 28%.

Of course, if such large-scale layoffs had not taken place, the financial situation of the operators might have been even worse. The problem is that any savings are eroded by wage inflation, severance payments, and the costs of outsourcing work. People also often forget that labor costs account for a relatively small proportion of the total costs of any large operator—in telecommunications companies in the United States, it may be only about one-fifth.

Difficult and dangerous

The real reasons for layoffs may still be those well-worn factors: routine efficiency measures, reducing redundancy in middle management, increasing reliance on external contractors, exiting non-core businesses such as television, eliminating duplicate positions after mergers and acquisitions, closing stores to accommodate customers shifting to online shopping, and more of the same mundane automation.

None of these can provide much comfort to the affected employees, or even to those who remain. If operators can lay off so many people without even needing AI, then might this rapidly developing technology help them to further significantly downsize in the coming years? It seems that the software part of telecommunications services is particularly vulnerable to the impact of AI. The ability to write code and to solve algorithmic problems based on a few basic hints has shown concerning improvements this year.

However, it would still be very difficult and dangerous to have Agent Smith eliminate human programmers. The difficulty lies in the fact that the heads of telecommunications companies still regard tools like GitHub Copilot for writing code as virtual assistants, used to increase the output of a certain programmer, rather than as a one-to-one substitute. The danger is that entrusting professional knowledge and decision-making power to AI could lead organizations to become dependent on expensive, error-prone technologies that are rarely truly understood by anyone. If humans no longer receive training for these positions, such skills might be lost forever.

It would be very instructive to see whether AT &T and Verizon can continue to lay off employees at the current rate. In 2020, hardly anyone could have predicted that AT &T would reduce its workforce by 45% over the following six years. If this trend continues, by 2032 when the 6G deployment begins, it may only have around 70,000 employees left.

By then, if Schulman is right, agents will bring about an “explosive growth in data traffic,” making the Verizon network a “scarce and valuable asset,” he told the analysts. “And we will price it accordingly. You will see different tariff plans.” On the day Verizon announced its financial results, the stock price fell by 3%, and over the past six months, it has fallen by a total of 9% due to market concerns regarding Starlink. Telecom analysts are skeptical about AI’s claims of improving productivity. However, they seem even less convinced by its claims of sales growth.