Verizon predicts AGI in a year and a half and "a century of progress in a decade": a CEO forecast, not a measured milestone

By Zendoric · September 30, 2026.

Dan Schulman, CEO of Verizon, said at the Ford Pro Accelerate event in Detroit that artificial general intelligence (AGI, a system capable of reasoning and acting with human-level competence in virtually any task, not just in specific domains) will arrive "in the next six to 18 months." In the same conversation with journalist Diane Brady, reported by Fortune, he added that the coming decade will compress "a century of progress": "the collapse of the 21st century," he called it.

Schulman is not an AI researcher, nor does he run a lab that trains models: he runs a telecommunications carrier with more than 250,000 employees and contractors. That does not invalidate his opinion, but it does require reading it for what it is -a forecast from an executive with his own stake in the disruption narrative-, not as a milestone verified by any lab or independent evaluation. There is no agreed definition today of what constitutes "AGI," and that vacuum is precisely what allows each executive to set the threshold wherever it suits them best. Worth remembering every time someone -from an AI lab or, as now, from a telco- puts a date on a term that not even the labs themselves can pin down precisely.

That said, Schulman's underlying diagnosis is not far-fetched, and it matches something we have been documenting for some time: the rate of improvement in frontier models remains high, with capability jumps every few months. What is different is the conclusion he draws for his own company. Verizon announced last week, as Schulman himself explained, a $70 million investment in training for American workers -not just the company's own employees-, with free AI training and professional development programs run with local organizations. Schulman was explicit about the reason: his workforce, organized today into "very defined roles and functions," will have to become "a much more fluid organization." And he was equally explicit about the human cost: "not everybody is magically going to land in a better place."

That honesty -that the adjustment will be neither automatic nor painless- is more valuable than the exact date he put on AGI. It is also the same thesis we have seen sector by sector in other coverage: disruption does not fall evenly, and judgment, human relationships and hands-on physical work hold up better than routine administrative tasks.

That was in fact illustrated by his own fellow panelists, in a contrast that probably says more than Schulman's forecast. Patti Poppe, CEO of PG&E, described an increasingly digitized electricity grid -with sensors, edge computing and predictive maintenance- that, far from doing away with field workers, will need people able to combine physical skill with technological competence: "somebody is going to have to still connect the wire to the pole," she said. Dave Regnery, of Trane Technologies, compared the current panic to articles about the personal computer 40 years ago -"they are identical," he said- and explained that his company uses AI to manage heating and cooling based on how buildings are actually used, not on their original design: 30% of all energy goes to buildings, and 30% of that is wasted, according to his figures, an inefficiency AI can trim while generating demand for more highly skilled technicians, not cutting headcount. Chris Nelson, of Stanley Black & Decker, described autonomous drilling robots that free electricians from repetitive tasks so they can focus on higher-value work. And Dan Peyovich, of Dycom, was the most direct: his problem is not a surplus of workers displaced by AI but a shortage of people willing to lay fiber and physical infrastructure; his answer has been to improve benefits and build in-house training, not to automate the workforce.

That mosaic -one CEO talking about AGI within 18 months alongside four others describing a shortage of skilled hands to lay cable and install sensors- captures the current moment well: software AI is advancing at great speed, but the physical world that sustains it -fiber, power grids, data centers, heating and cooling- still needs, and will need for years, human hands. It is the same paradox we documented when analyzing employment sector by sector: the administrative back office is the most exposed; field work, technical judgment and the customer relationship are what hold up best.

Our reading: the value of this story is not in the date Schulman put on AGI -which will probably shift, as has happened so often with this kind of executive forecast-, but in the concrete decision that goes with it: $70 million in training, aimed not only at his own workforce but at the community, in anticipation that the labor market may not adjust on its own. That is the right attitude, regardless of whether the horizon is 18 months or five years: treating disruption as something to be planned for and cushioned, not as something that will work out "magically." Over the long term we still think the trajectory points toward abundance -more efficiency, more productivity, resources wasted today (that 30% of energy in buildings is a small but telling example) freed up for other uses-. But that horizon does not arrive by itself, and the companies already investing in the transition, rather than merely announcing the big bang, deserve more attention than those who only put a date on the headline.