A stark social experiment in corporate stratification is unfolding at Meta Platforms, Inc.. As artificial intelligence reshapes the organizational structure of major tech firms, a new fault line is deepening within the company—an AI elite enjoys astronomical compensation and limitless resources, while the remaining workforce faces the specter of layoffs, collapsing morale, and even a desire to be included in redundancy lists.
This phenomenon is termed the "K-shaped company," mirroring the logic of wealth polarization in a K-shaped economy. Internally, two distinct tracks are forming: salaries and status for the top-tier AI elite continue to climb, while ordinary employees at the bottom are increasingly viewed as replaceable components.
Some Meta employees have already publicly criticized management, initiated petitions, and in the UK, attempted to form a union.
Management has publicly acknowledged issues with these practices and is rolling out a series of appeasement measures. However, analysts point out that this structural imbalance cannot be resolved simply by improving employee benefits.
CEO Mark Zuckerberg recently conceded that the development progress of the company's AI agents has fallen short of executive expectations. Meta's newly launched Muse Spark AI system also failed to match the performance of cutting-edge models from OpenAI and Anthropic on core benchmarks.
The AI Elite's Feast: Nine-Figure Compensation and Special Privileges
Within Meta, the treatment enjoyed by top AI talent is worlds apart from that of the average employee.
The company's leading AI executives and researchers are receiving compensation packages worth nine figures (i.e., hundreds of millions of dollars), with some offered up to $500,000 in additional equity incentives, contingent merely on staying with the company. Meanwhile, the shadow of large-scale layoffs looms over other departments.
One Meta employee stated bluntly that internally, "the only people truly doing well are those with the highest pay and the closest ties to AI development." Another employee was more direct: "The only people not feeling dissatisfied are those executives."
This internal division is not unique to Meta but reflects a long-standing dual-track employment structure in Silicon Valley—the chasm in treatment between full-time employees and contractors is longstanding. However, the new variable introduced by the current AI wave is that even full-time employees face marginalization if they are not within the core AI circle.
The Plight of Ordinary Employees: Surveillance, Layoffs, and Morale Collapse
For employees on the lower track of this K-shaped structure, the situation is markedly different. Some have been forcibly assigned to teams described internally as "suffocating." Headcount continues to be cut, and the median salary has declined even as company profits grow.
More controversially, the company tracked employees' keystrokes and mouse clicks to train AI agents—a project halted only after encountering security vulnerabilities that employees had previously warned about.
Chief Technology Officer Andrew Bosworth offered a candid summary of management's overall attitude: "You can leave, or you can disagree and commit."
Low morale has evolved into a management crisis Meta can no longer ignore. In anticipation of the latest round of layoffs, some employees reportedly hoped they would be on the list—a deeply symbolic shift for a company once considered one of America's most desirable employers.
Harvard Business School professor Amy Edmondson noted in an interview: "They are conflating Silicon Valley's 'fail fast' mantra with reckless execution. This isn't the chaos necessarily accompanying AI innovation; it's a failure of basic management."
AI Progress Lags, Management Trust Depleted
The internal turmoil is beginning to negatively impact Meta's core strategic goals. Zuckerberg recently publicly admitted that the accelerated development of Meta's AI agents has not met executive expectations. The company's new Muse Spark AI system also lags behind the performance of leading models from OpenAI and Anthropic on key metrics and has faced setbacks in building open-source AI models.
Management internally is aware of the problems. Executives privately described the rollout of the company's new AI organizational structure as "terrible" and acknowledged the work environment has become "brutal."
When Zuckerberg recently announced a hackathon to boost morale, employee response was reportedly tepid—they are not only exhausted, but some stated: "I'm not sure this company supports hackathon culture anymore. I don't feel secure enough to spend time on innovation not tied to performance metrics."
In an internal memo, CTO Bosworth offered a rare, candid diagnosis of the core issue: "We have broken your trust—your trust that your expertise and contributions would be valued, that you could grow and advance, that this is a place where you can truly make an impact."
The Cost of Division and an Uncertain Path to Repair
Currently, Meta has introduced a series of measures to placate employees, including promises of more detailed management support, greater stability, no large-scale layoffs this year, increased team budgets, and improved office infrastructure. A widely criticized hot-desking policy is also being scaled back.
But some observers suggest this might signal a deeper inflection point—the steadily eroded voice of employees in the post-pandemic era may be at a nascent stage of resurgence. The Meta case demonstrates that treating employees as replaceable parts rather than partners ultimately backfires, harming the business itself.
The core paradox of this K-shaped divide is this: AI was supposed to flatten organizational structures, yet in reality, it is fostering more rigid power hierarchies. How to drive technological leaps while maintaining basic trust within the organization is becoming an unavoidable management challenge for large technology companies.
Comments