AI Workforce Strategy Why CEOs Must Choose Growth Over Cost Cutting

AI Workforce Strategy: Why CEOs Must Choose Growth Over Cost Cutting

A Wall Street Journal report published in May 2026, authored by Chip Cutter and Lindsay Ellis, captures a defining inflection point in corporate leadership. Coinbase is eliminating 14% of its workforce. PayPal is targeting a 20% reduction over the next two to three years. Meta is cutting 8,000 jobs—roughly 10% of its headcount. Meanwhile, Spotify is holding staffing levels flat while shipping more product, Axon Enterprise is actively reassuring its 5,000 employees that layoffs are not coming, and IBM’s chief human resources officer is predicting that her company will employ more people in three years, not fewer.

The article frames this divergence as a binary choice: shrink the workforce or stretch it. That framing is useful for a news story, but it is incomplete as a leadership framework. The more consequential question business leaders should be asking is not which of these two camps to join. It is whether their AI strategy is fundamentally oriented toward reducing yesterday’s cost base or building tomorrow’s competitive capacity. Evidence from organizational history, behavioral economics, and current AI deployment data suggests these are not equivalent paths—and leaders who conflate operational efficiency with strategic advantage are making a category error that will cost them dearly.

The Real Stakes of the AI Workforce Decision

The Cutter and Ellis piece correctly identifies that neither approach, at least as currently practiced, involves hiring. That detail deserves more scrutiny than it receives in the original reporting. The fact that virtually no major corporation is using AI as a catalyst for net new employment in the near term tells us something important: most executives are still processing AI primarily as a cost management tool rather than a growth engine.

This is understandable. Cost savings are immediate, quantifiable, and rewarded by equity markets. When Block and Snap announced AI-related job cuts, their share prices jumped. That feedback loop is powerful, and it creates a gravitational pull toward workforce reduction that operates independently of any genuine analysis of long-term organizational capability.

But history offers a cautionary parallel. During the business process reengineering wave of the early 1990s, companies that aggressively downsized in pursuit of efficiency gains frequently discovered that they had eliminated not just redundancy but also tacit knowledge, customer relationships, and the organizational resilience needed to adapt when conditions changed. A landmark study by Wayne Cascio, published in the Academy of Management Executive, found that companies that downsized did not consistently outperform those that avoided layoffs, and that frequent downsizers often suffered lasting damage to employee morale, innovation capacity, and long-term profitability (Cascio, W.F., “Downsizing: What Do We Know? What Have We Learned?”, Academy of Management Executive, 1993).

The parallel is not exact. AI genuinely does change the productivity calculus in ways that earlier automation waves did not. But the underlying leadership temptation—to treat a structural transformation as primarily a headcount optimization exercise—is identical.

Two Philosophies, Two Risk Profiles

The companies choosing to cut are not monolithic in their reasoning, and it is worth separating the legitimate from the opportunistic.

Coinbase’s Brian Armstrong presents what might be called the honest restructuring case. Cryptocurrency markets are down. The company built headcount during a bull cycle. AI now allows fewer people to accomplish more. Under these conditions, a reduction in force is arguably a rational response to both market reality and technological capability. Armstrong’s acknowledgment that engineers are now shipping in days what once took a team weeks is not spin; similar productivity claims are being validated across software development contexts by credible research, including a 2023 study from MIT’s Sloan School of Management that found GitHub Copilot users completed coding tasks 55% faster than those without AI assistance (Noy, S. and Zhang, W., “Experimental Evidence on the Productivity Effects of Generative AI”, MIT Working Paper, 2023).

The concern is not with companies like Coinbase that are rightsizing under genuine market pressure. The concern is with organizations using AI adoption as a narrative justification for workforce reductions that are really driven by investor relations and stock price management. When executives announce layoffs alongside AI adoption stories and watch their share prices climb, they are being taught a lesson by markets that may not hold over a five-to-ten year horizon. Short-term stock price appreciation driven by cost reduction is not the same as building durable competitive advantage.

Meta’s situation illustrates the complexity. The company is making massive capital investments in AI infrastructure while simultaneously reducing its human workforce. CFO Susan Li’s admission that the company does not know what its optimal size will be is refreshingly honest, but it also reveals a deeper strategic uncertainty. A company that is spending billions on AI capability while eliminating the institutional knowledge and human judgment needed to deploy that capability wisely is not executing a coherent strategy. It is placing a very large, very expensive bet.

The Productivity-to-Growth Reframe

IBM’s chief human resources officer, Nickle LaMoreaux, offers the most strategically coherent framing in the entire WSJ article, and it receives the least attention. Her question—whether leadership discussions are moving from “AI to productivity” to “AI to growth”—identifies the exact conceptual leap that most organizations are failing to make.

Productivity gains are a means, not an end. A company that uses AI to make its existing operations 40% more efficient but does not redeploy that efficiency into new products, new markets, or improved customer experiences has not created value. It has simply reduced cost while its competitive position relative to growth-oriented rivals deteriorates.

Spotify’s approach is instructive here. Co-CEO Gustav Söderström’s description of keeping headcount flat while doing “much more shipping, more value to consumers” reflects a clear-eyed understanding of what AI-driven productivity is actually for. The goal is not to do the same things with fewer people. It is to do more things with the same people, expanding the frontier of what the organization can offer.

This distinction matters enormously for talent strategy. A company that communicates to its workforce that AI is primarily a tool for identifying who can be eliminated will generate a specific kind of organizational behavior: self-preservation, information hoarding, resistance to AI adoption, and the departure of exactly the high-performing, autonomous employees who have other options. A company that communicates that AI is a tool for amplifying what each person can accomplish will generate a different set of behaviors: experimentation, collaboration, and a genuine cultural appetite for learning.

Axon Enterprise President Josh Isner’s email to staff, while somewhat blunt in its “block out the noise and keep kicking ass” register, gets this dynamic right at an intuitive level. His argument that even three-times productivity gains will simply surface new problems to solve reflects a view of organizational capacity as expansible rather than fixed. That is not naïveté. It is a strategically defensible position grounded in the observation that demand for capability tends to rise to meet available supply.

The Hollow Middle: Neither Cuts Nor Growth

The WSJ article identifies a third category that deserves more critical scrutiny than it receives: companies that are keeping headcount flat without a clear articulation of what they intend to do with the additional capacity. Synchrony Financial’s approach of preparing employees for “redeployments” rather than layoffs is humane and organizationally sensible, but HR chief DJ Casto’s acknowledgment that things will not be “black and white” also signals a strategic ambiguity that can become its own kind of risk.

Workforce agility is valuable, but agility without direction is just motion. Organizations that tell employees to prepare for change without specifying the nature or destination of that change create sustained anxiety that degrades performance. Research from organizational psychologist Adam Grant and colleagues has consistently found that uncertainty about job status and role identity is among the most corrosive forces for individual and team productivity (Grant, A., Give and Take: Why Helping Others Drives Our Success, Viking, 2013). Companies in the middle camp—neither cutting aggressively nor investing visibly in growth—risk accumulating the cultural costs of anxiety without capturing the financial benefits of efficiency or the competitive benefits of expansion.

What the Gartner Data Actually Tells Us

The Gartner survey cited in the WSJ piece, finding that approximately 80% of companies using AI agents or autonomous technologies are cutting staff, is striking—but it requires careful interpretation. Survey data from midlevel managers and above may reflect announced plans rather than completed actions, and it likely reflects the current moment’s orientation rather than a settled long-term equilibrium.

More importantly, the 80% figure tells us what companies are doing, not whether it is working. A more useful dataset would track three-to-five year revenue growth, market share, and innovation output for companies in the cutting camp versus those in the stretching camp. That data does not yet exist at scale, because the current wave of AI-driven workforce decisions is still in its early stages. But the historical evidence from previous automation transitions, including the introduction of ATMs in banking and robotic process automation in insurance, suggests that the relationship between workforce reduction and competitive advantage is far more complex than the share price bumps following layoff announcements would imply.

ATM introduction, for example, did not reduce bank teller employment over the long run. It reduced the cost per branch, which made it economical to open more branches, which increased overall teller employment even as each branch required fewer tellers. The mechanism here—that productivity improvements can expand the market they serve rather than simply reducing the inputs required—is likely to operate in many AI adoption contexts as well, particularly in knowledge work where the constraint on output is often imagination and execution capacity rather than raw labor hours.

What Strong Leadership Looks Like Right Now

The CEOs and executives navigating this moment most effectively share several observable characteristics that go beyond their position on the cut-or-stretch spectrum.

• They communicate with specificity rather than abstraction. Isner’s email to Axon staff works not because it promises job security in perpetuity, but because it grounds that assurance in concrete business reasoning. LaMoreaux’s framing works because she connects AI adoption to a specific outcome—more employees in three years—even while acknowledging she cannot be certain. Vague reassurances and vague threats both erode the psychological safety that organizations need to navigate transitions well.
• They invest in reskilling with the same seriousness they invest in technology. Amazon’s commitment to spend $1.2 billion on upskilling programs for its workforce, announced in 2019 and subsequently expanded, represents the kind of organizational investment that turns AI from a threat narrative into a capability narrative (Amazon, “Upskilling 2025”, Amazon.com, 2019). Companies cutting costs via AI adoption while simultaneously reducing investment in human development are optimizing for a very short time horizon.
• They resist letting AI strategy be driven by investor relations rather than competitive strategy. Companies whose share prices jumped after AI-related layoff announcements will face a reckoning if subsequent quarters do not show the revenue growth and innovation output that would justify the organizational disruption. Leaders making genuine strategic bets, even if those bets involve near-term cost, are better positioned for that reckoning than those surfing current market enthusiasm for AI efficiency stories.

The Question Leaders Must Answer

The WSJ article by Cutter and Ellis captures a real and consequential divide among corporate leaders, and it does so with admirable clarity and reportorial thoroughness. But the binary it presents—cut or stretch—is ultimately a tactical question nested inside a more important strategic one.

The organizations that will look back on 2026 as a period of genuine competitive repositioning are those whose leaders asked not “how many people can we afford to keep?” but “what would we attempt if our teams could accomplish three times as much?” The former question produces downsizing plans. The latter produces strategies.

Justin Briley, the laid-off Coinbase employee quoted in the original article, made a prediction worth taking seriously: AI, he said, will ultimately create “more work, not less.” He may be right. And the organizations positioned to capture that additional work will be those whose leaders had the strategic clarity—and the courage—to invest in human capacity even when markets were rewarding those who cut it.

The choice is not really between layoffs and retention. It is between leaders using AI to write the final chapter of their current business model and those using it to begin the next one.