This year alone, over 1.1 million U.S. jobs have been cut. October saw the highest monthly layoff total in 22 years. UPS eliminated 48,000 workers. Amazon targeted 30,000 roles. Target slashed 1,800 positions. Companies are making a specific bet: hold the line on headcount, pump billions into AI, squeeze more productivity from fewer people.

Brian Chesky of Airbnb put it bluntly:

“If people are getting more productive, you don’t need to hire more people”.

Goldman Sachs froze hiring and cut roles it believes AI will cover. Walmart projects sales growth without adding staff for three years. It’s the most aggressive corporate bet on automation since the dot-com boom.

But here’s the wrinkle no one is talking about publicly: 42% of companies have abandoned most of their AI projects this year, according to S&P Global Market Intelligence. That’s double the previous year’s rate. The result is a dangerous vacuum.

This isn’t a technology problem. It’s a potential strategic workforce planning catastrophe that will define winners and losers over the next five years. Every CHRO and C-level executive reading this is running into the same wall: aggressive headcount cuts based on AI’s promise, paired with AI’s failure to deliver on that promise.

The crash-landing is coming. The question is whether your organization will be ready to land it or if you’ll be scrambling to rebuild talent that took decades to develop.

The Bet That Looked Rational (Until October)

October 2025 broke something in the labor market. Over 153,000 job cuts in a single month, the worst October since 2003. These weren’t cyclical adjustments or necessary course corrections. Companies posting record profits cut staff anyway. UPS, with $21.4 billion in revenue and earnings that beat expectations, eliminated 48,000 positions. Goldman Sachs reported record third-quarter profits while implementing hiring freezes and modest workforce reductions. The script was identical across industries: technology will replace labor, so we cut first and optimize later.

Cost-cutting led the charge, accounting for 50,437 announced job reductions in October. But AI was right behind it, directly cited in 31,039 cuts. When you add federal workforce cuts (labeled “DOGE Impact” in the data), the dominoes keep falling—293,753 announced layoffs. Capital that once flowed to talent acquisition now flows to automation infrastructure. Microsoft and Amazon announced tens of billions in AI spending while announcing waves of layoffs.

Goldman’s projections were concrete: 4% workforce reduction over the next year, escalating to 11% within three years. Financial services faces the steepest cuts—14% over three years. Technology follows at 10%. Retailers are betting they can hold headcount steady while sales tick up.

The underlying bet is simple: if AI delivers on its productivity claims, you can do more with fewer people. If it doesn’t, you’ve already cut the people, so it doesn’t matter. That’s where the logic breaks down.

The Technology Promised Magic. It Delivered Pilots.

This is where the story gets uncomfortable for everyone who approved these strategies. S&P Global surveyed over 1,000 technology and business managers. The result: 42% of companies abandoned most of their AI initiatives by late 2024—a jump from just 17% the year before. The reasons are depressingly consistent. Unrealistic expectations. Lack of internal expertise. Bad data. Can’t scale past the pilot phase. Security risks and data privacy concerns emerged as top obstacles preventing production deployment.

Gartner predicts 40% of agentic AI projects will be canceled by 2027. NTT Data reports that 70-85% of GenAI enterprise deployments fail to hit ROI targets. Omdia found that nearly one-third of companies report complete failure of their AI proof-of-concept projects, with only 9% moving more than half their pilots into operational use. RAND Corporation did the math: 80% of AI projects fail—double the failure rate of traditional non-AI technology initiatives.

The real culprit? Leadership doesn’t understand what they’re building. Only 32% of companies have identified specific human tasks that AI should supplement or replace. Most don’t have the internal expertise to implement systems. MIT research is blunt: 95% of AI projects fail to deliver ROI, largely because leadership lacks practical understanding of what AI can and can’t do.

JPMorgan’s global CIO acknowledged the scale of the problem. The company shut down hundreds of AI projects. Some contributed code and concepts to other initiatives. Many just stopped. Salesforce cut 1,000 roles while hiring for AI positions, then reduced customer support from 9,000 to 5,000 employees—a 45% cut—by deploying its Agentforce platform. The platform works. Mostly. When it doesn’t, customers have fewer people to call because the people already left.

The Institutional Knowledge Vanishing Act

Here’s what no one talks about in earnings calls: when you cut 1.1 million jobs without a coordinated reskilling effort, you lose institutional knowledge that took decades to build. These aren’t just frontline workers. They’re analysts, recruiters, operations managers, subject matter experts—the people who understand why processes work the way they do.

The U.S. has no reskilling infrastructure at scale. Individual companies launched workforce academies. Walmart, Amazon, IBM invested in training programs. Federal dollars went toward regional initiatives under the Infrastructure Investment and Jobs Act. These are drops in an ocean. The World Economic Forum reports 44% of core job skills will change by 2027. Most organizations overstaffed during the pandemic and are now using AI as a convenient justification to correct what should have happened years ago.

The math on institutional knowledge loss is brutal. Harvard Business Review estimates the cost of losing an employee at up to 20 times average recruitment and training expenses—mostly because you lose what researchers call “deep smarts,” experience-based knowledge that’s almost impossible to replicate. Panopto calculated that new hires spend 200 hours chasing down information that departed employees took with them. For a small business under 1,000 employees, this costs $2.4 million annually in lost productivity. For a 30,000-person company? $72 million.

When tenured employees leave, relationships vanish. Problem-solving approaches disappear. Institutional memory walks out the door. AI can’t fill that gap. Research from the American Management Association shows companies lose both documented knowledge and tacit, experience-based knowledge during layoffs. New employees need months to understand roles that may not be properly documented. Meanwhile, laid-off employees join competitors, taking expertise that strengthens rivals while weakening the original organization’s market position.

Accenture saw this play out in real time. The consulting firm eliminated 11,000 roles through an $865 million restructuring. CEO Julie Sweet was explicit about the strategy: employees who couldn’t demonstrate tangible AI system training would be exited. “Where we don’t have a viable path for skilling—exiting people so we can get more of the skills we need,” she said. Companies are hiring talent with AI skills rather than reskilling experienced people who lack them. That’s faster. It’s also permanent displacement of the institutional knowledge that made your organization competitive in the first place.

Why Reskilling Isn’t Optional Anymore

Prior to my current role, I spent the last two years in the talent seat trying to figure out how I would reskill and redeploy my workforce. It’s the difference between thriving and getting stuck in cycles of layoffs, panic hiring, and persistent skills gaps. Deloitte research pointed out that 74% of organizations say reskilling is important for their success over the next 12-18 months, but only 10% report being very ready to pull it off. The World Economic Forum projects six in ten workers will need retraining by 2027 to keep pace with change—but only about half have access to adequate training. That’s the gap your organization has to fill.

Successful AI implementation requires workflow redesign, not just automation. McKinsey found that fundamentally redesigning workflows has the biggest effect on realizing impact from generative AI. But only 21% of organizations have redesigned workflows. Fewer than 20% track well-defined KPIs for their AI solutions, despite this being the strongest predictor of bottom-line results. The companies following BCG’s 10-20-70 rule—10% effort on algorithms, 20% on data and technology, 70% on people, processes, and cultural transformation—achieve 2.1 times greater ROI than tech-first organizations.

AT&T proved reskilling works at scale. The company committed $1 billion to retrain 100,000 employees—nearly half its workforce—for technology roles. Seagate saved $33 million by redeploying staff instead of laying them off. These organizations created clear pathways for people to use new skills in internal roles. The result: improved morale, reduced recruitment costs, retained institutional knowledge.

The obstacles workers face are organizational problems HR can fix. Emotional burnout (24%), cost (24%), lack of clarity on what to learn (21%), and time constraints (17%). PwC found that 77% of workers are ready to retrain, but fewer than one-third of companies reward people for developing new skills. Stanford’s research shows younger workers in AI-exposed fields like customer service, marketing, and software support saw a 16% employment decline between 2022 and 2025. Some of those roles won’t return in their previous form. But many could if companies invested in structured transition paths instead of treating displacement as termination.

The CHRO’s Three-Move Playbook

If you’re running HR, you have three moves. Make them now.

Move One: Reclaim the workforce narrative. Talent isn’t a cost center. It’s the thing competitors can’t easily replicate. McKinsey data shows CEO oversight of AI governance is the most correlated factor with bottom-line impact—yet only 28% of organizations have CEOs directly overseeing this. Position your talent strategy as competitive advantage, not a line item to cut when revenue tightens.

Move Two: Audit AI investments ruthlessly. Is this initiative built on solid data infrastructure, clearly defined use cases, measurable KPIs, and realistic ROI timelines? Or is it built on vendor hype? Companies tracking well-defined financial metrics from day one, redesigning workflows around AI capabilities rather than automating broken processes, and implementing human-in-the-loop safeguards significantly outperform peers. Do this audit now, before you cut more headcount based on promises that won’t materialize.

Move Three: Champion reskilling at scale. Protect learning and development budgets even during cost-cutting. Create structured pathways for employees to transition into AI-adjacent roles. Recognize that people who already understand your customers, processes, and culture bring advantages that no external hire can match. Guild’s Talent Resilience Index estimates that reskilling could generate $8.5 trillion in extra revenue by 2030. The World Economic Forum advocates for “reskilling at scale” combined with “redesigning roles for human-AI collaboration”.

Organizations implementing efficient, focused learning and development structures significantly outperform peers, according to Deloitte, because they recognize that HR is a strategic partner bringing integrated, future-focused perspectives. Current research shows 62% of laid-off workers believe AI or digital skills will increase long-term job security. Twenty-two percent of hiring managers acknowledge that upskilling in AI could have made employees more resilient to recent layoffs.

Conclusion: The Choice Is Still Yours

Layoffs are data. The absence of a workforce strategy when AI roadmaps hit walls is strategy failure at the executive level. Companies treating this as standard headcount reduction will find themselves in endless cycles: more layoffs, panic hiring, persistent skills gaps, repeat. Organizations treating this as workforce redesign will get stronger.

Your AI roadmap will hit a wall. The question is whether your organization will be ready when it does. Whether you’ll have the talent to pivot. Whether you’ll have the institutional knowledge to recover. Whether you’ll have anything left but processes and automation. The time to decide is now. The time to act is today.

Frequently Asked Questions

1. What’s really driving the 1.1 million layoffs in 2025?
It’s a three-part story. First, companies overstaffed during the pandemic and are correcting. Second, companies are aggressively investing in AI and freezing headcount to fund technology infrastructure. Third, cost-cutting to boost near-term profits. AI has been directly cited as the driver in 48,414 layoffs this year, while federal workforce reductions account for 293,753 announced cuts.

2. Why are 42% of AI projects actually failing?
Because organizations underestimated the complexity. Unrealistic expectations. Lack of internal expertise. Bad data quality. Inability to scale beyond pilots. Over 80% of AI projects fail—double the failure rate of traditional technology initiatives. Companies are solving the wrong problems with expensive tools.

3. How do you keep institutional knowledge from walking out the door?
Implement knowledge management systems before people leave. Conduct structured offboarding that captures expertise. Create mentorship programs pairing veterans with successors. Document tacit knowledge in accessible platforms. Harvard Business Review estimates losing an employee costs up to 20 times recruitment and training expenses due to institutional knowledge loss. Protect that asset before it’s gone.

4. What’s the actual difference between reskilling and upskilling?
Reskilling means learning skills for a completely different area or field—helping people embrace new technologies through degree programs or intensive training. Upskilling means enhancing existing capabilities to stay relevant—accomplished through workshops, online courses, or mentorship as a continuous process.