Last month, Mark Zuckerberg announced that Meta would cut roughly one in 10 jobs and cancel thousands of open roles as the company races to spend billions on raw silicon and virtualized infrastructure in pursuit of what he described an “AI -forward” future.

The cuts added to a mounting list of 137 tech companies that have reduced headcount by 113,000 so far in 2026, a rate of 825 jobs per day.

For the Meta employees, who waited anxiously for an email Wednesday, this decision reads like a verdict on their future. As panic spreads across the industry, there is a growing sense that the bots have arrived, and white‑collar work is living on borrowed time.

Diane Bryant is not buying it.

Bryant, who experienced homelessness as a teenager, went on to become the head of Intel’s data center business and was once the COO of Google Cloud. She has lived through every major tech shock of the last 40 years. She now sits on several boards, backs and invests in new infrastructure startups, and has created the Diane Bryant Innovation Center in her local college district, a brick‑and‑mortar bet that the future of the tech industry still depends on human curiosity, talent and collaboration, not just more compute.

Speaking to me from the balcony of the New York Stock Exchange, she shared a macro reality check on what’s unfolding beneath the AI boom, and why she believes the industry is still in the early stages of a much longer transformation.

“I’ve seen the advent of the PC, then the internet, then mobile computing, then the cloud and now AI,” she says. “They’re all very, very similar in how they go from nascent to pervasive. This doesn’t happen overnight. I tell people all the time, true transformation takes about 10 years.”

The Trillion-Dollar Delusion: What the Birth of AWS Teaches Us About AI Hype

​Bryant’s core thesis is that AI, while it might be sold as sorcery, will not come with a magic wand for deployment. The way she sees it, breakthrough technology moves fast in labs and headlines but more slowly inside companies.

She shares an insider story on cloud computing, which hindsight has deemed inevitable, as a great lesson for why she believes the AI timeline is somewhat misunderstood.

In the early 2000’s, Andy Jassy, who would lead Amazon Web Services before eventually becoming Amazon’s CEO, approached her global infrastructure group with what seemed like a radical idea: He wanted to lease out Amazon’s dormant, post-holiday computing infrastructure to software startups and split the revenue.

"I took it to the board twice, and the notion of cloud computing had just been invented," she recalls. "The word 'cloud’ was first spoken by Eric Schmidt at Google in 2006. So nobody knew what a cloud was, and so the board response was, 'This sounds like IBM Professional Services and you know hosting, and we don’t do that.’ So, I went back to Andy and I said, 'Can’t get it over the finish line.’ And he said, 'Okay, we'll just do it.' That became AWS."

The 10-Year Reality Check: Why the Enterprise Factory Can’t Run Without the Builders

Yet, despite the immediate hype surrounding the launch, scaling corporate walls required an entirely different level of structural maturity.

“AWS launched in 2007, and everybody was talking about the cloud,” she says. “But it was around 2016, before most CIOs were willing to put their core enterprise systems there. It takes about 10 years for a technology to become broadly embedded in how companies actually run.”

She is clear not to equate AI with the cloud, acknowledging that AI is seeing much faster adoption than in the early days of the cloud era. What she is saying, however, is that the human and organizational bottlenecks remain strikingly familiar.

“Today, there are real blockers for adopting AI. Period. Full stop,” she says. “You need talent inside the company that really understands it. You have serious security concerns around your crown‑jewel data. You have accuracy issues, hallucinations, and you have people who simply don’t want to change how they work. Until those are addressed, you don’t get broad-scale adoption.”

I ask Bryant about Wall Street’s take on the SaaSpocolypse , a term used to describe the huge wipeout in market cap for a sleuth of software companies, as analysts hedge bets on a future where AI will eat all software and wipe out entire layers of knowledge work. She is clear she strongly disagrees.

“I don’t believe AI will eat all software and all jobs,” she says. “This idea that you won’t need software engineers anymore, that there won’t be any more software, that’s just wrong.”

Instead, Bryant views AI as an accelerator, not a replacement. “AI is the front end to every software solution, not the solution itself,” she says. “It sits in front of CAD. It sits in front of your CRM. It makes those systems more powerful and easier to use. It doesn’t make them disappear.”

She is clear; however, that doesn’t mean there won't be significant pain, driven predominantly by the fact that AI, like all tech evolutions, is, by definition, an efficiency play. “Every big technology shift is about efficiency,” she says. “Cloud lets you deploy applications faster. Mobile lets you do more from anywhere. AI lets you iterate and ship faster. They’re all efficiency plays.”

To explain why, Bryant leans on a 19th‑century economic idea that has quietly predicted each and every tech wave that’s come before: Jevons’ paradox. The belief that efficiency equals more need and more demand. “If you make something more efficient, you don’t drive demand down; you drive it up,” she says. “That’s what I’ve seen my whole career.”

​AI and the Real Fault Line: Revenue Vs Overhead

That paradox, she believes, splits labor risk into two camps. “If you’re an engineering company and you’re an engineer, AI is a gift,” she says. “You can offload the low‑value iterations on your design. You can do things twice as fast. That drives your top line, faster business growth, faster revenue growth.”

Those who are viewed solely as a cost center to the business face a bigger risk. “If you’re a staff function not directly tied to revenue,” she says, “that’s where companies will use AI to justify reducing OPEX.”

Bryant has also seen, up close, how people react when the foundation of their job changes.

Cloud Computing’s First Workforce Reckoning

When she led Intel’s IT organization in 2010, she decided the company needed to aggressively adopt cloud itself if it was going to credibly sell that vision to its customers. This created a huge change for long‑tenured infrastructure teams.

“There was a revolt,” she says. “If you’d spent twenty years managing servers and I suddenly told you the server didn’t matter anymore because everything was being virtualized, you heard one thing: you don’t matter anymore.”

Naturally, reactions were torn. Some employees chose to retrain into cloud roles. Others did not.

“You opt in, or you ultimately end up leaving,” she says. “The people who leaned into it became the experts everyone relied on.”

AI: The Biggest Mistake Is Thinking the AI Story Is Already Written

None of this makes this moment less brutal for those grappling with Layoffs. Bryant is clear‑eyed about the human impact. But she urges those impacted not to lose confidence in their role in this industry's future and to treat any single restructuring, even one as large as Meta’s, as proof that AI has already finished its sweep.

“Every time a big new technology shows up, people declare the world as we know it is over,” she says. “Ten years later, there’s actually more software, more systems, more complexity, and more need for people who understand it.”

In her view, AI in 2026 looks less like the final chapter and more like the start of another compressed decade of rewiring. Instead, this story is only beginning, and the most important decision ahead is whether to move with the work as it changes.

“This is the beginning of a long adoption curve, not the end,” she says. “The only real mistake is deciding the story is finished when it’s just getting started.”