Invisible Unemployment is going to really pick up in tech in 2026.
It’s already here. It just doesn’t show up in the numbers yet. And yes, it’s fueled by AI—but not in the way most people think.
The unemployment rate in tech looks fine on paper. Layoff announcements have slowed down from the bloodbath of 2023. VCs are deploying record capital again. But beneath the surface, something fundamental is shifting. And if you’re a founder, a leader, or someone trying to navigate their career in tech right now, you need to understand what’s actually happening.
At a recent gathering of CEOs, 66% of leaders surveyed said they planned to either fire workers or maintain the size of their teams in 2026. Only one-third indicated they planned to hire. The unemployment rate has already risen to 4.6%—the highest in four years. Economists at Indeed expect it to hover there throughout 2026.
This isn’t a blip. This is a regime change.
The root cause is multi-layered. Let me break it down.
#1. Almost Every B2B CEO I Know Wants to Run a Leaner Team in 2026
This isn’t new. CEOs wanted this in 2025 too. But here’s what’s different: now they’re actually executing the plan in full.
At a recent conference, Shopify CFO Jeff Hoffmeister was asked to describe the company’s hiring plans. His answer? “I don’t see us next year needing to increase head count in any way. It has been over two years we’ve been at this [same] head count. As I look to next year, I think we can continue to be disciplined on head count.”
That’s Shopify. One of the most successful e-commerce platforms on the planet. Flat headcount for over two years. And planning to stay that way.
I’ve had dozens of conversations with founders and CEOs over the past few months, and the theme is remarkably consistent. Everyone wants flat headcount—or close to it—paired with strong growth. Some are even more aggressive: they want to shrink headcount while growing 40%+.
Let’s do the math on what this actually means.
If you’re a $20M ARR company with 100 employees, you’re at $200K ARR per employee. That used to be pretty standard for a growth-stage B2B company. But leaders now want to get to $300K, $400K, even $500K ARR per employee. Even $1M+ for AI leaders like Replit, Elevenlabs, etc. Not eventually. Now.
At $500K ARR per employee, that same $20M company only needs 40 people. And when you scale to $50M ARR? You might only need 100 employees instead of 250.
This is a fundamental restructuring of what it means to build and scale a B2B company. Net net, it means far fewer jobs as startups scale. The hiring that would have happened at each stage of growth? A huge chunk of it simply won’t happen anymore.
The jobs aren’t being eliminated. They’re just never being created in the first place.
Even Microsoft is now beyond peak headcount:
#2. Backfilling Departures with AI is Now a Top 5 Priority
Here’s a critical distinction that most people miss: this is different than firing people to replace them with AI.
Almost no CEO I know is walking into the office and saying “Sarah, you’re fired, we’re replacing you with Claude.” That’s not how it’s playing out. That’s just media drama for the most part — outside of customer support.
Instead, it’s much gentler—and much harder to see in the data.
When someone leaves—whether they quit, get a promotion elsewhere, or just move on—the default question used to be: “Who do we hire to replace them?” Now the default question is: “Can we backfill this role with an AI agent instead?”
And increasingly, the answer is yes.
Wells Fargo CEO Charlie Scharf put it bluntly this month: the bank expects to have fewer people heading into next year. They’ll continue to retrain their workforce and use “attrition as our friend,” he said. “It’s not going to totally replace humans, but it does create an opportunity to do things significantly different.”
“Attrition as our friend.” That phrase should send a chill down the spine of anyone job hunting in 2026.
And here’s the IBM twist: when your voluntary attrition drops from 7% to under 2%, companies lose that “friend.” IBM’s RTO mandate—requiring three days a week in-office with badge swipes monitored, non-compliance leading to termination—isn’t just about productivity. One employee told The HR Digest it’s “a way to cut headcount without headlines.” The policy is expected to drive 10-15% voluntary attrition, particularly among workers who can’t or won’t relocate to Austin or New York.
Customer support? Obviously. But it’s spreading fast. Entry-level marketing roles. Junior data analysis. First-pass legal review. Basic sales development tasks. Content creation. QA testing.
Almost every CEO I talk to has this as an explicit goal now. Not “replace humans with AI” but “when humans leave, try AI first.” It’s a gentle reduction. Invisible. No layoff announcements. No bad press. The headcount just… doesn’t grow. Or slowly shrinks through attrition.
One founder told me last month: “We had 3 people leave in Q4. We replaced one of them. The other two roles? We’re running an experiment with AI agents for 90 days. If it works, those headcount slots just disappear from the plan.”
Multiply that across thousands of startups and you start to see the scale of what’s coming.
#3. There’s Zero Interest in Employees Who Can’t Truly Drive AI Change
This is the quiet truth that no one wants to say out loud. But I’m going to say it.
Most of the CEOs I know think half their team is still the wrong team.
Not wrong in the 2019 sense of “not a culture fit” or “B-player.” Wrong in the sense of: they don’t really understand AI. They don’t use AI tools daily. They can’t manage AI agents. They couldn’t architect an agentic workflow if their job depended on it.
And here’s the uncomfortable part: their job does depend on it. They just don’t know it yet.
I’ve talked to founders who have invested heavily in AI training for their teams. The results? Mixed at best. Some employees get it. They become 3x more productive. They’re automating parts of their own jobs and reinvesting that time into higher-value work.
But a significant chunk—maybe 40%, maybe 50%—just… don’t get it. They attend the training. They nod along. And then they go back to doing their job exactly the same way they did it in 2022.
These aren’t bad employees by traditional metrics. They’re often experienced, hard-working, loyal. But they’re operating in a paradigm that’s rapidly becoming obsolete.
And CEOs are noticing. They’re not mass-firing these people. But they’re also not fighting to retain them. They’re not promoting them. And when a “restructuring” happens, we all know who ends up on the list.
The invisible part? This often happens without anyone explicitly saying “we’re cutting people who don’t get AI.” It just becomes the reality.
#4. Everyone Needs More Engineers—But Not The Same Engineers
This one is creating massive disruption in dev organizations, and honestly, we’re still figuring it out in real-time.
Here’s the paradox: nearly every startup I know is trying to hire more engineers. Engineering headcount is the one area where most CEOs say “yes, I’d add more if I could find the right people.”
But—and this is crucial—they don’t want the same engineers they would have hired two years ago.
The profile has fundamentally changed:
Not wanted:
- Developers who resist AI coding tools
- “Good but not great” computer science grads
- Engineers who see GitHub Copilot as cheating
- Mid-pack graduates from non-elite CS programs
- Developers who can’t context-switch between AI-assisted and traditional coding
Highly wanted:
- Developers who are masters of AI-fueled coding
- Engineers who can architect and manage LLM-based systems
- People who understand prompt engineering at a deep level
- Developers who can build with, on top of, and around foundation models
- Engineers who view AI tools as multipliers, not threats
The mid-pack CS grad from a non-top-5 program who would have gotten 10 job offers in 2021? Many of them are now seeing zero offers. Not because they’re bad engineers. Because the definition of what makes a good engineer has shifted underneath them.
Meanwhile, the senior engineer who’s spent the last 18 months going deep on AI tooling, who can spin up an AI-native application architecture, who thinks natively in terms of human-AI collaboration? They’re getting multiple offers with comp packages that would have seemed absurd three years ago.
Same “engineer” job title. Completely different role. And a massive population of talented people caught in the gap.
#5. What You Can’t Say Outloud: Many People Just Don’t Want The New Jobs
This is perhaps the most underappreciated driver of Invisible Unemployment. And it’s the one that makes this problem so hard to solve. And one it’s tough to talk about without getting flamed. But it’s true.
The jobs are changing. In many cases, the new versions of the jobs are actually pretty interesting. But they’re not the jobs people signed up for. And many people—if they’re being honest—just don’t want them.
- Do CROs really want to manage 20 AI Agents themselves? To spend their days prompt-engineering SDR bots and tuning lead-scoring algorithms? Most CROs I know got into sales leadership because they love the human side of selling. Building teams. Coaching reps. The energy of a sales floor. Not debugging why the AI agent is hallucinating about pricing.
- Do CMOs want to go back to building campaigns themselves? To personally manage AI agents, generate content at scale, and be directly responsible for delivering leads to sales—not through a team, but through their own hands-on work with AI tools? Most CMOs evolved away from execution for a reason.
- Are AEs and SDRs willing to become mini-FDEs? To roll out agentic products for customers themselves? To become technical implementers on top of their sales responsibilities?
The honest truth is that many people in these roles don’t want these jobs. Not really. They want the 2021 jobs. The jobs where you managed a team of humans from home, where you had clear functional boundaries, where “scaling” meant hiring more people, not deploying more agents.
But many of those jobs are gone. Or going.
And here’s the even tougher part: this mismatch means that many open roles will actually remain unfilled. Not because there’s no budget. Not because the company doesn’t want to hire. But because the candidates they’re seeing don’t want the job as it actually exists today—and the candidates who might want the new version of the role are vanishingly rare.
So what happens? The role stays open. Eventually, the company tries an AI solution. Sometimes it works well enough. And that headcount slot quietly disappears from next year’s plan.
More invisible unemployment.
#6. Folks Are Staying Longer. Maybe You Should, Too.
IBM CEO Arvind Krishna just revealed that voluntary attrition at IBM in the U.S. has dropped to under 2%—down from a typical 7%. That’s the lowest rate in 30 years. Let that sink in. In an industry that historically runs 13-21% annual turnover, employees are now clinging to their jobs like never before.
“People aren’t looking to change jobs,” Krishna said. “That then leads to less hiring because people aren’t leaving.”
This creates a vicious cycle. Low attrition means companies don’t need to backfill. No backfilling means no job openings. No job openings means people afraid to quit. Which drives attrition even lower. And the wheel keeps spinning.
And maybe it should. If you’re not quite ready for the demands of the Age of AI … maybe stay. Seriously consider just … staying.
What Does This All Mean?
Let me be clear: I’m not celebrating this. This isn’t a “look how efficient AI makes everything” victory lap. The human cost of this transition is real and it’s significant.
Economists are calling it the “Great Freeze.” The corporate playbook for 2026? Don’t hire. Companies are looking to stay lean while relying on technology to take on more tasks. The weakest industries for new job openings include those in well-paid fields: data analytics, software development, marketing, and entertainment.
“We’re close to zero job growth. That’s not a healthy labor market,” Federal Reserve governor Christopher Waller said recently. And when I talk to CEOs around the country, they all say the same thing: “We’re not hiring because we’re waiting to figure out what happens next.”
But here’s the thing about the Great Freeze: it can’t last forever. At some point, workers retire. Companies need to replace natural attrition. Except now, with AI, they might not need to replace all of it. Or even most of it.
If you’re a founder: You need to think hard about the composition of your team. Not just for efficiency—though that matters—but for what your company actually needs to succeed in an AI-native world. The org chart from your 2023 board deck? It’s probably already obsolete.
If you’re an employee: This is not the time to coast. The ground is shifting fast. The employees who will thrive are the ones who are genuinely, deeply embracing AI tools and workflows—not as a novelty, but as a core part of how they work. Be honest with yourself about whether you’re one of them.
If you’re job hunting: Understand that the market is much harder than the headline numbers suggest. Yes, there are job openings. But there’s a massive mismatch between what companies are looking for and what many candidates are offering. Bridge that gap or face an extended, frustrating search.
2026 is going to be a year of significant increase in Invisible Unemployment. The layoffs won’t make headlines. The jobs will just… not materialize. People will be “between opportunities” for longer stretches. Hiring processes will drag on and then fizzle.
The numbers tell the story: when IBM’s voluntary attrition drops from 7% to under 2%—the lowest in three decades—in an industry that typically runs 13-21% annual turnover, something structural has broken. And when 66% of CEOs say they’re either cutting staff or holding headcount flat, while unemployment sits at a four-year high, we’re not looking at a normal cycle anymore.
It’s not a good thing. But it is coming. And the sooner we name it, the sooner we can start figuring out how to navigate it. It’s time for everyone to adjust. Fighting it and lashing out won’t help.
