Every generation faces a new wave of technological disruption and responds with the same blend of overconfidence, short-termism, and reluctance to let go of what’s working. The internet did it. The mainframe era did it. AI is doing it now.
Onstage, in conversation with Great Place to Work CEO Michael C. Bush, Devgan sounded a similar tune about why he believes AI is a bit overhyped.
Backstage with Fortune, Devgan dismissed the idea that AI is unlike anything we’ve ever seen, even as he hailed its breakthroughs. He kept coming back to a constant refrain: Humans will be human, no matter what technological changes society undergoes.
That framing helps explain why Devgan is relatively unbothered by one of the loudest anxieties in tech right now: the idea that AI data centers will strain electric grids, spike utility bills, and ultimately prove energetically unsustainable.
He sees it as a classic first-derivative mistake—projecting a straight line from current conditions and ignoring the human ingenuity that always bends the curve. Calling it a “first-derivative projection,” he said people extrapolate from the data-center boom onto a spike in utility bills, “but human innovation always saturates.” He predicted software efficiencies alone—not quantum computing, not new energy sources, just better algorithms—will deliver the 10x improvements in AI computation that make today’s projections obsolete.
“It always happens in software,” the Silicon Valley veteran told Fortune. “One software change can give you 10x improvement.”
“The best time to do this is when you’re doing really well,” he said, “because the typical mistake is when you’re doing really well, you will just try to milk what you have.”
On the question of what comes next, Devgan gets expansive. He called Waymo “the biggest breakthrough in AI in the last five years”—a window into a $3 trillion to $4 trillion global transportation industry on the verge of total transformation. He estimated 25% of downtown Los Angeles currently consists of parking lots—real estate that should become available the moment self-driving goes mainstream. On defense, he said he sees the industry being “completely redesigned for autonomous”—noting the absurdity of a $1 million missile being fired in Iran to knock down a $30,000 drone. Robotics and drug discovery are the next frontiers, for him: “We can’t even imagine how different the world is going to look.”
And yet, in the same breath, he returns to his anchor: Human nature doesn’t change. Kids today have the same worries about careers and friendships that his generation did. The nostalgia for previous eras is always misplaced. The warnings about disruption are always slightly overblown, the timelines always slightly wrong—self-driving cars were supposed to arrive in 2012, he noted, and they’re only arriving now.
Onstage with Bush, Devgan framed this not as pessimism, but as a kind of operating principle. His biggest worry about AI adoption, he said, isn’t the technology—it’s the disconnect between executives who are enthusiastic and employees who are skeptical.
“The enthusiasm is very high at the leadership level,” he said, “but there’s more skepticism at the employee level—and that’s the real thing.” His advice to leaders: Stop positioning AI purely in terms of margins and efficiency.
“We need to bring everybody along and do it in a truthful manner, right, in a transparent manner,” he said. Not everything has to be positioned as a question of financial gains or increased margins, he added, but “also how it affects the whole organization.” (In other words, the human part.)
