Meta's Kenyan Contractor Cuts 1,100 Jobs After AI Training Contract Ends

Meta has cut its contract with Sama, a Nairobi-based data annotation firm, triggering the immediate layoff of roughly 1,100 workers, according to a 1 May 2026 report. The contract had covered the labelling of video data gathered through Meta's Ray-Ban smart glasses — a product marketed to consumers in Western markets but whose training footage passed through a supply chain of low-wage workers in Kenya.

The job losses landed without warning on an already fragile labour market. Kenya has become a significant node in the global AI industry's data-preparation chain, with firms like Sama and others providing the human annotation work that makes machine-learning systems functional. That work is cheap, often stressful, and — as this episode shows — renewable only so long as the next contract lasts.

A documented history

The relationship between Meta and Kenyan data firms is not new territory for investigators. A 2023 investigation published by TIME documented that workers at Sama facilities — the same company now shedding staff — had reviewed content including graphic material with inadequate psychological support. The reporting prompted at least one major tech client to revise its practices. Meta's own role in that supply chain drew scrutiny at the time. Workers reported that the conditions which later became public had persisted for months before any intervention.

When the contract ended on or around 1 May 2026, the layoffs followed immediately. Workers who spoke to researchers and advocates said they received little explanation and no transition support. The speed suggests that the Kenyan workforce had been held in a standing-by arrangement — present when needed, dispensable when not.

Who bears the cost

The structural logic of this arrangement is clear in both directions. For platform companies, contracting annotation work to firms in Nairobi or Manila offers cost savings that Western labour markets cannot match. For the workers themselves, even difficult gig-style annotation jobs represent formal employment in economies where formal employment is scarce. That asymmetry is what makes the leverage so one-directional.

When Meta ended the contract, it absorbed no direct cost. The severance — or absence of it — was Sama's problem. Sama's revenue shock became 1,100 Kenyan families' income loss. The arrangement is typical of how global technology supply chains operate: profit and risk are distributed along a gradient that runs, with very few exceptions, from wealthy consumers and shareholder returns at the top, to low-wage workers in the Global South at the bottom.

Platform companies routinely describe themselves as innovative employers or connectivity providers. The reality of how their products are assembled — including the human labour that makes AI systems functional — rarely appears in investor presentations or product launches.

The accountability vacuum

What makes this episode notable is not its uniqueness but its transparency. The contract termination and resulting layoffs were not hidden. The @pirat_nation Telegram account reported them on 1 May 2026, with the layoffs confirmed at approximately 1,100 workers. The information exists. The question is what follows from it.

Regulatory frameworks governing AI supply chains remain underdeveloped across most jurisdictions — including in the United States, where Meta is headquartered, and in the European Union, which has led on platform regulation more broadly. Kenyan labour law covers formal employment, but enforcement capacity is strained and the legal relationship between a Kenyan contractor and a foreign platform company involves questions that national frameworks were not designed to answer. No multilateral standard exists that would require Meta to provide transition support, advance notice, or disclosure of its AI labour supply chain.

What happens next

The workers who lost jobs in this round face the same market that produced their employment in the first place: demand is driven by the annotation needs of global AI companies, and those needs fluctuate with product cycles and contract renewals. For Kenya's data annotation sector, the episode reinforces both the sector's real growth and its vulnerability to decisions made in Menlo Park or Dublin without consultation.

For the broader AI industry, this episode is a reminder that the systems marketed as transformative technology depend on labour arrangements that would be recognisable in any commodity supply chain — workers called in when demand rises, sent home when it falls, with minimal protection either way. That structure has remained largely invisible to consumers and regulators. Whether it stays that way may depend on whether the next layoff cycle — and there will be one — generates the same coverage.

This publication covers Big Tech's Global South operations as a desk priority. Major wire services have not yet carried the specific contract-termination details reported here; Monexus is tracking the story as it develops.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://t.me/pirat_nation/8474