AT&T and Verizon ended September with about 8,000 fewer employees between them than they had at the end of June, according to the two companies’ own quarterly filings. The figures, first highlighted in an analysis by Light Reading, show the largest US telecom operators still shrinking their workforces quarter after quarter.
AT&T reported 127,150 employees at the end of the third quarter, down from 130,870 at the end of the second, a fall of 3,720. Verizon, which reports full-time equivalents rounded to the nearest hundred, put its workforce at 93,300 on 30 September, down from 97,600 on 30 June, a fall of about 4,300. Together that is roughly 8,000 fewer people, or about 3.5% of the combined total of around 228,500 in June.
What AT&T and Verizon’s filings show
The headcounts come from the financial and operating schedules each company filed with the SEC alongside its results. AT&T’s third-quarter schedule, filed on 21 October, shows 127,150 employees against 135,670 a year earlier, a decline of 6.3%. Verizon’s third-quarter schedule, filed on 26 October, gives 93,300, compared with 89,900 at the end of 2025.
Verizon’s number is higher than at the start of the year because it completed its acquisition of the broadband operator Frontier on 20 January, its 10-Q says. The filing also shows the cost of shrinking: Verizon paid $272m in severance in the third quarter and $1.4bn over nine months, “primarily related to separations in connection with our workforce reduction initiatives”, and had $679m of severance still owed at 30 September.
Neither company’s filings attribute the reductions to AI. The two measures are also not identical: Verizon counts full-time equivalents, while AT&T reports total employees, so the combined figure is an approximation.
Light Reading argues AI is not the driver
Light Reading’s International Editor, Iain Morris, says the cuts “seem to have had little to do with AI”, even as Verizon’s chief executive, Dan Schulman, talks up an agent-driven future. Light Reading quotes Schulman telling analysts on the results call that “there could be at least 100 billion agents” in the next couple of years, compared with about 8.5 billion people.
Morris points out that both operators were cutting staff long before ChatGPT’s release in November 2022, and that the deepest cuts came before then. He argues the real drivers are more familiar: efficiency drives, fewer layers of management, more outside contractors, retreating from businesses such as TV, removing duplicate roles after mergers, store closures and older forms of automation. He also notes that, by his figures, the two companies together now employ fewer people than AT&T alone did in 2020.
Verizon itself is pitching the network as an AI asset, from a dark fibre deal with AWS for AI workloads to early work with Ericsson on a pre-standard 6G field call.
Why telecom job cuts are hard to pin on AI
The filings support the headline number but say nothing about why the jobs went, and that gap matters. A falling headcount at a large telecom operator can reflect asset sales, outsourcing, attrition, store closures or the end of merger integration as easily as automation. Verizon’s own 10-Q describes its severance as tied to “workforce reduction initiatives”, a phrase that covers all of these. Without a company breaking down which roles went and why, calling these AI layoffs would be a guess.
There is also a timing problem for the AI story. AT&T’s headcount was already 6.3% lower than a year earlier, and Light Reading’s longer series shows years of steady decline at both operators. A trend that predates generative AI cannot easily be credited to it, even if AI tools now play some part in customer service or software work. The more useful question for staff and regulators is whether AI accelerates a decline that was already happening, and that will only show up over several more quarters of filings.
This is the same pattern TSN has found when looking at what surveys and individual layoffs actually show about AI and jobs: claims of AI-driven job loss often run ahead of the evidence. Some governments are starting to ask for that evidence directly, such as a South Korean bill that would make firms report AI-driven mass layoffs. No such disclosure rule applies to US telecom operators, so their quarterly headcount line remains the main public signal.
The next data point should come with the operators’ fourth-quarter results, which they have usually published in late January, when both will report year-end headcounts.
Sources
- AT&T: Form 8-K, Exhibit 99.2, third-quarter 2026 financial and operational results, filed 21 October 2026
- AT&T: Form 8-K, Exhibit 99.2, second-quarter 2026 financial and operational results, filed 22 July 2026
- Verizon: Form 8-K, Exhibit 99, third-quarter 2026 results, filed 26 October 2026
- Verizon: Form 8-K, Exhibit 99, second-quarter 2026 results, filed 24 July 2026
- Verizon: Form 10-Q for the quarter ended 30 September 2026, filed 29 October 2026
- Light Reading: AT&T and Verizon normalize layoffs as 8,000 jobs go in a quarter, 30 October 2026
Related stories
- Verizon Signs $2bn-Plus Dark Fibre Deal With AWS for AI Workloads
- Verizon and Ericsson Say They Completed a Pre-Standard 6G Field Call in 7 GHz
- AI and Jobs: What a UK Survey, a French Study and One Layoff Actually Show
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X
