Sama to lay off over 1,100 Kenyan workers after Meta

Kenya's artificial intelligence outsourcing industry has encountered a significant inflection point. Sama, one of Africa's largest data annotation and AI training companies, is laying off over 1,100 workers—approximately 40% of its Kenyan workforce—following the termination of a major contract with Meta. This development represents far more than a single company's operational adjustment; it signals fundamental vulnerabilities in East Africa's emerging AI services economy that European investors have increasingly targeted over the past three years.

Sama has positioned itself as a cornerstone of Kenya's tech ecosystem since its founding in 2014. The company specializes in training machine learning models through human data labeling—work that involves thousands of Kenyan employees annotating images, text, and video to improve AI algorithms for global technology companies. Meta's contract represented a substantial revenue pillar, and its loss creates immediate financial stress. However, the deeper concern for investors lies in the sector's structural dependency on a handful of major clients and the volatile nature of outsourced AI work.

The timing is significant. Kenya has aggressively marketed itself as Africa's AI hub, with government policies explicitly designed to attract foreign AI companies seeking low-cost, English-speaking labor pools. Between 2020 and 2023, the sector experienced explosive growth, with multiple outsourcing firms expanding operations and creating tens of thousands of jobs. European investors—particularly from Germany, the UK, and the Nordic countries—viewed Kenya's AI sector as a strategic play on both the continent's digital transformation and the economics of nearshoring non-core AI development work.

This narrative has now become precarious. Sama's layoffs reveal critical risk factors that prospective investors should reassess. First, client concentration remains dangerously high. When a single contract represents such a material portion of revenue that its loss triggers mass redundancies, the business model lacks resilience. Second, the commoditization risk is real—as more countries develop AI outsourcing capabilities, and as automation improves, the value of human data annotation may compress faster than anticipated. Third, regulatory pressure on major tech clients (including Meta) is increasing scrutiny of labor practices in outsourcing arrangements, potentially introducing compliance costs and contract volatility that weren't previously priced into projections.

For Kenya's broader economy, the implications extend beyond Sama. The company employs some of the country's most educated workers in a sector that has become symbolically important to Kenya's positioning as a tech leader. Job losses here have reputational consequences that ripple through the entire ecosystem, potentially affecting recruitment for competing firms and dampening government enthusiasm for sector-specific incentives.

European investors currently exposed to Kenya's AI outsourcing sector face a critical juncture. The thesis wasn't wrong—demand for AI training data remains genuine and growing—but the execution risks were underestimated. Companies with diversified client bases and less commodity-like service offerings (those adding proprietary methodologies, domain expertise, or specialized training) are likely to weather this contraction. Pure-play data annotation firms face margin compression and consolidation pressure.