Meta’s bold, relentless push into AI has turned 2025 into a high-stakes sprint. When CEO Mark Zuckerberg told staff to brace for an “intense” year in January, he wasn’t speaking in hyperbole. Over the past twelve months, Meta has radically reshaped its organization, redirected capital toward AI initiatives, and plunged billions into competition with OpenAI, Google, and other tech giants. The leadership shift accompanied a broader cultural reset, including a public emphasis on what Zuckerberg described as more “masculine energy.”
This upheaval came alongside a tempered ambition for the metaverse, a raised bar for performance, and widespread layoffs as Meta pursued Zuckerberg’s audacious concept of “personal superintelligence.” The moves have delivered faster productIteration and greater efficiency in some areas, but they’ve also sparked internal frictions: AI-driven reorganizations have prompted tensions between longtime staff and new hires, and several ex-employees have criticized the atmosphere and strategy. Some observers also argue that labeling certain displaced workers as “low performers” was demoralizing and politically risky.
Meta’s revival mirrors a broader industry shift among major tech companies: pare back costs, tighten management layers, and gamble on AI as the defining frontier of the next decade. After a mid-2025 cultural reset, internal sentiment among Meta’s workforce has shown signs of improvement, according to a company spokesperson referencing a recent survey.
Investors, however, remain cautious. Meta’s plan to invest tens of billions of dollars into AI infrastructure and talent has raised questions about timing, scope, and return. While Meta’s stock has risen in 2025, it has trailed the broader market and several of its Magnificent 7 peers. Industry analysts emphasize the need for a clear, cohesive vision that connects every piece of Meta’s elevated AI strategy to steady, visible growth.
This portrait of Meta’s so-called “year of intensity” draws on conversations with more than a dozen current and former employees, analysts, and researchers who spoke with Business Insider.
The AI overhaul in focus
In mid-2025, Zuckerberg aimed to shift perceptions by signaling that Meta would not be outpaced in AI. The company spent about $14 billion to acquire Scale AI and brought on its founder, Alexandr Wang, as chief AI officer. Shortly after, Meta rebranded its AI-focused unit as Meta Superintelligence Labs (MSL), signaling a strategic reorientation toward high-impact AI work.
As leadership sought to consolidate efforts and lure top talent from competitors, some former staff described the effort as lacking a unified, coherent AI playbook. A former MSL engineer, Joena Zhang, stated in a November LinkedIn post that the organization suffered from unclear ownership in the first half of the year, with numerous meetings but few concrete decisions. In a July Substack piece, ex-Meta researcher Tijmen Blankevoort argued that Meta’s vision appeared unsettled and difficult for teams to rally behind.
To attract rival AI stars, Meta offered substantial compensation packages, drawing talent from OpenAI and Google’s AI divisions, among others. This talent influx intensified competition within MSL, as insiders highlighted disparities between new hires and veteran staff in pay and prestige, plus the race to access scarce computing resources.
In August, Meta executed its fourth major reorganization in six months, splitting MSL into four groups: a new TBD Lab (short for “to be determined”), a product team overseeing the company’s AI assistant, an infrastructure group, and FAIR, Meta’s long-standing Fundamental AI Research lab. The reshuffle left ownership of projects murky and prompted some reassignment, with concerns about uneven information flow between TBD and MSL.
Meta responded by pointing to a post from a communications executive as a rebuttal to internal coverage of the restructuring.
Attrition followed the organizational shakeups. At least eight MSL staffers—including researchers, engineers, and a senior product leader—left within two months of the formation of MSL. Meta noted that many were long-tenured and that some level of turnover is normal in a company of Meta’s size. Two months later, another wave of job cuts hit: roughly 600 positions were eliminated as part of a broader MSL reorganization aimed at expediting decision-making.
Industry observers have been mixed on the impact. Some, like Futurum’s Shay Boloor, believe Meta’s changes have accelerated model releases and integration across Facebook, Instagram, and WhatsApp, placing Meta among the few firms actively training frontier-class models at scale for billions of users.
Reality Labs, Meta’s reality- and VR-oriented division, also saw leadership changes. There has even been talk of potential budget reductions within the metaverse unit within Reality Labs, with Meta stating it would shift some investment from Metaverse toward AI glasses and wearables to match momentum, while denying broader restructuring plans.
Intense performance reviews and a leaner organization
The layoffs and reorganizations were part of a broader push to tighten operations. Meta trimmed management layers and implemented a stricter performance-review system. Zuckerberg had warned in January that the company would raise the bar for performance and move quickly to dismiss “low performers.” By February, roughly 3,600 people were let go from a workforce of about 78,450.
By May, managers were instructed to place a larger share of employees in the bottom performance tier for teams of 150 or more—roughly 15–20% below expectations, up from prior-year levels. This shift intensified competition and encouraged project-by-project risk management as teams sought to avoid ending up in the lowest tier.
Internal sentiment appeared to improve in the latter half of the year, with Meta citing a second-half survey showing gains in optimism, pride, and confidence in leadership. Participation was high, lending weight to the reported uptick in morale.
Departures and loyalty in flux
The confluence of policy shifts, reorganizations, and aggressive performance standards sparked a wave of departures in 2025. Some leaving employees cited misalignment with Meta’s evolving governance and political posture, while others pointed to a sense that the company’s values were shifting too far from their own. An engineer who exited after eight years described Meta in 2025 as a very different creature from the Oculus and Facebook era, noting conflicts between personal principles and the company’s alignment with government posture.
Other farewell messages highlighted ongoing pressure, perceived unfairness, and concerns about scope and narrative control. Meta’s spokesperson stressed that resignations represented a small portion of its 78,450-strong workforce, noting that headcount had grown year over year.
Feedback channels and culture under strain
Some departing staff also reported reduced opportunities to provide candid feedback to leadership on topics like DEI and the company’s “masculine energy” framing, with questions in Q&As sometimes being preselected or posts critical of leadership being removed from internal channels.
In a 2025 internal poll about workplace fear, many respondents indicated significant concern about speaking openly about working conditions. Still, other employees described Meta as a rewarding place for high-performing individuals who can thrive under pressure, particularly those who enjoy a frontier-level focus in AI, wearables, and robotics—and who appreciate strong compensation and perks.
Key takeaways
- Meta’s 2025 was defined by aggressive AI investment, structural overhauls, and a controversial leadership style aimed at accelerating innovation.
- The company faced internal friction from talent recruitment wars, unclear project ownership, and a tougher performance culture that some viewed as cutthroat.
- Despite tensions, Meta reported improved employee sentiment late in the year and remains committed to pushing AI capabilities across its ecosystem, even as it recalibrates investments toward new wearables and AI-enabled devices.
What do you think about Meta’s approach to AI leadership and cultural change? Does the emphasis on rapid model development and heavy performance scrutiny help or hinder long-term innovation? Share your perspective below.
