The Duke CFO Survey dropped last week. 750 U.S. CFOs surveyed. 44% plan AI-related job cuts this year, projecting roughly 502,000 roles eliminated. That's 9x the 55,000 AI-attributed layoffs tracked by Challenger, Gray & Christmas in 2025.
Half of those cuts are white-collar. Developers. IT staff. Project managers. The people who actually build and maintain things.
Here's the part nobody's saying out loud: the work doesn't disappear when the people do.
The Regret Is Already Here
An Orgvue survey of 1,100+ C-suite leaders found that 39% already made employees redundant because of AI. Of those, 55% regret it. More than a third spent more on rehiring than they saved.
That tracks. RAND Corp puts the AI project failure rate above 80%, which is double the failure rate for non-AI IT projects. Companies are firing people for tools that don't work yet, then scrambling when they realize nobody's left to do the job.
A major fintech company is the poster child. They cut roughly 1,200 employees between 2022 and 2024, publicly claiming their AI chatbot replaced 700 customer service agents. Customer satisfaction tanked. Complaints stacked up. By February 2025, the CEO admitted they went too far and started rehiring with a hybrid model. AI plus humans. Which is what anyone paying attention would have suggested from the start.
Nine Out of Ten Don't Even Have the AI Built
This is the stat that should reframe how you think about 2026.
Forrester VP J.P. Gownder said it plainly: "Nine out of ten times, they haven't even started" building AI to replace those roles. Harvard Business Review found 60% of executives reduced headcount in anticipation of AI, but only 2% tied actual layoffs to real AI implementation.
Nearly 60% of hiring managers admit they emphasize AI's role because it's "investor-friendly" versus admitting financial distress. The Department of Labor called it "scapegoating AI."
So the layoffs are real. The AI replacing those workers is not. The work still needs doing. And now there are fewer people to do it.
Where That Work Goes
The managed services market hit $424 billion this year, up from $380 billion in 2025, growing at 12.8% annually. 72% of U.S. small and mid-sized businesses plan to increase managed IT spending in 2026. Cybersecurity services are the fastest-growing segment at 18% annually.
Forrester predicts half of AI-attributed layoffs will be quietly reversed, but rehired offshore or at lower salaries, "under the guise of AI." That's the outsourcing pipeline in action. Internal full-time roles become external contracts.
If you run an agency or MSP, this is the setup. Enterprise companies are gutting the teams that maintained their infrastructure, built their custom tools, and kept their compliance programs running. That work is about to show up as RFPs, emergency contracts, and frantic emails asking if you can start Monday.
The Catch
52% of MSPs say hiring is their number one growth constraint. 26% say they don't have enough staff to service more clients right now. 68% of IT leaders can't recruit cloud and security talent internally.
So the demand is about to spike, and the supply side is already stretched thin. If you can't absorb overflow work without a three-month hiring cycle, you're going to watch those contracts go to someone who can.
This is the actual problem. Not "how do we find clients" but "how do we deliver when the clients show up all at once."
The SMB Side Is Even More Interesting
The Duke CFO survey found something the headlines missed: AI adoption could lead to increased hiring among smaller firms. While enterprises cut, SMBs are bringing AI tools online and need help implementing them. They don't have internal teams to cut in the first place.
These aren't companies recovering from bad AI bets. They're companies trying to make good ones. They need builders, not consultants who talk about AI. They need people who can actually wire the systems together and make them work.
What This Means for You
If you're an agency or MSP, the next 12 months are a capacity question. Not a sales question. The demand is coming whether you market for it or not.
A few things to think about:
Your bench matters more than your pitch deck. Can you spin up a development team for a client engagement in two weeks, not two months? If the answer is no, figure out your white-label situation now, before the wave hits.
Cybersecurity is the sharpest edge. Companies cutting IT and security staff still need to pass audits, respond to incidents, and maintain compliance. That expertise doesn't get automated. It gets outsourced. If you can deliver security work, you'll have more inbound than you can handle.
Don't confuse "AI washing" with "AI doesn't matter." The companies making bad cuts are creating the demand. The companies making smart investments are also creating demand. Both sides need technical capacity they don't have internally.
The smaller players have an advantage here. A two-person studio with the right automation can absorb work that would take a traditional firm weeks of onboarding. We know this because we do it. Between the two of us at Kief Studio, we cover what would normally require a 10-14 person team, backed by 40+ custom tools we built ourselves. Our white-label clients bring us in when they need technical depth on tight timelines, and we've been doing it long enough to know the pattern.
The pattern right now: capacity is the bottleneck, not demand. Agencies that solve the capacity problem win the next two years. The ones that don't will spend 2026 turning down work they could have taken.
The Conversation Worth Having
If you're an agency or MSP staring at this wave and wondering how to absorb it without tripling headcount, that's the exact conversation we have with technical partners. No pitch, no pressure. Just figuring out if there's a fit.
First conversation is free. kief.studio/contact