Meta Ends Kenya Partnership As 1,100 Workers Face Layoffs
On 1 May 2026, Meta quietly ended its partnership with Sama, a Nairobi-based company that provided data annotation services for the tech giant's artificial intelligence systems, including content used to train features in Meta's Ray-Ban Smart Glasses. The decision set off a chain reaction: Sama proceeded to terminate approximately 1,100 workers in Kenya. The workers, many of whom spent years labeling data to train AI models that generate billions of dollars in revenue for the Menlo Park company, were left without work. Some workers told local media they lost training opportunities and income with little notice or recourse.
What happened in Nairobi is not an anomaly. It is a feature of how platform companies organize their global operations. Meta, Alphabet, Amazon, and Microsoft have spent years building international labor supply chains for the unglamorous work that makes AI systems function — work that requires human judgment to label images, moderate content, and train algorithms. The profits of that labor accrue in California; the conditions of that labor are determined in Nairobi, Manila, or Hyderabad, far from the boardrooms where strategy is set.
Sama's relationship with Meta was not publicly disclosed until it became a story. The company, headquartered in San Francisco with operations in Kenya and Uganda, has previously handled content moderation work for social media platforms before shifting toward AI annotation tasks. Workers in Nairobi processed images and videos to help train AI systems used across Meta's products. The scale of the layoff — roughly 1,100 positions eliminated — represents the bulk of Meta's contracted annotation workforce through Sama, according to sources familiar with the arrangement. Meta declined to provide additional comment on the contract termination or the resulting job losses.
The workers now facing unemployment have limited leverage. In prior disputes over compensation and working conditions at similar outsourcing arrangements, employees have faced significant barriers to collective bargaining or legal recourse. A TIME investigation in 2023 documented systemic concerns about pay, mental health support, and job security in similar data annotation operations in Kenya. The companies that benefit most from this labor — and that hold the intellectual property produced by it — have remained largely insulated from the downstream consequences of their contracting decisions.
The structure of platform labor is designed to produce exactly this outcome. When a company like Meta decides to restructure, the risk is externalized onto contracted workers rather than absorbed by the firm itself. The arrangement allows technology companies to expand their AI capabilities without bearing the full social cost of the workforce required to build them. Workers in Nairobi performing this labor face lower wages, limited benefits, and minimal protections relative to direct employees of the platforms they serve.
The timing of Meta's contract termination coincides with a broader push by the company to expand its AI offerings. Meta has invested heavily in AI infrastructure, partnered with manufacturers like Luxshare Precision for hardware, and deployed AI features across its social media ecosystem. The Ray-Ban Smart Glasses, a product of those broader ambitions, rely on AI systems trained in part by data annotated by the workers now facing dismissal in Kenya. The financial returns from those products flow primarily to shareholders in the United States. The consequences of the labor that enabled them now fall on Nairobi residents with little connection to those gains.
The episode illustrates a durable asymmetry in how technology companies organize their global operations. AI systems are presented as products of American innovation and engineering — and to a significant degree they are. But the labeled datasets that make those systems functional require millions of hours of human labor, disproportionately performed in countries where labor costs are lower and worker protections are weaker. The workers who perform this labor are essential to the technology yet largely invisible in the public narrative around AI development.
What remains unclear from the available reporting is precisely when Meta made its decision to end the contract, whether any transition period was offered to affected workers, and what recourse, if any, Sama provided. Meta has not issued a public statement addressing the layoffs directly. The sources consulted do not indicate any company-funded transition assistance or formal commitment to the workers who lost positions as a result of the contract termination. This publication was unable to reach Sama for comment as of publication.
The 1,100 workers in Nairobi who helped train Meta's AI systems are now navigating that uncertainty without the institutional support that might cushion the transition. The platform company that depended on their labor has moved on to its next development cycle. The question of who bears responsibility for the people who make AI development possible — and at what point — has not been settled by any regulatory framework currently in place in Kenya or elsewhere. Until it is, episodes like this one will continue to recur whenever cost pressures or strategic priorities shift for the companies at the top of the supply chain.
Desk note: Monexus has followed this story through local reporting and public-source verification. We did not have access to a press release or official company statement; our account draws on available reporting and publicly observable labor market consequences. We welcome corroboration from the companies involved.
Wire provenance
This editorial synthesis draws on the following public wire/social posts:
- https://x.com/pirat_nation/status/1919012347929710592