More than 700 Covalen workers in Dublin have been told their roles are at risk on Meta projects, showing how the AI buildout is starting to reshape the contractor layer that keeps big technology platforms running.
Covalen, a subsidiary of CPL, has opened consultations over roughly 700 potential redundancies at its Dublin operation, where staff work on projects for Meta. More than 500 of the affected roles are understood to be tied to AI annotation, with other jobs in content moderation, quality analysis and management also caught in the process.
The Communications Workers' Union has called for negotiations that deliver fair redundancy packages for staff. CWU deputy general secretary Ian McArdle said the decision showed workers on vendor contracts were being asked to absorb the cost of Meta's new spending priorities, as the Facebook and Instagram owner pours money into artificial intelligence while reducing headcount elsewhere.
According to The Journal, workers were called to a 3pm meeting at Covalen's Sandyford offices and told the roles were at risk because of changing business conditions, not individual performance. The development comes only days after Meta told employees it would cut about 10% of its global workforce, around 8,000 jobs, and leave thousands of open roles unfilled.
Meta's AI bill explains the pressure. The company has guided for capital spending of as much as $135 billion in 2026, a figure that broadly matches its combined AI investment over the previous three years. That money is going into data centers, chips, model development and high-priced AI talent. It is also forcing a harder look at teams and suppliers that no longer fit the company's new cost structure.
For Ireland, the question is not only how many direct Meta employees may be affected. The larger concern is the ecosystem around the company. Meta has about 1,800 employees in Ireland, but its footprint extends through vendors such as Covalen, which provide services that sit close to the platform's daily operations. When Meta changes direction, the impact can travel quickly through those contracts.
Contractor Vulnerability
Covalen has around 2,000 workers in Sandyford providing services connected to Meta. The latest consultation follows an earlier round in November, when 400 roles were put at risk and about 200 workers later left the business. For employees, that history matters. It suggests this is not a one-off adjustment, but part of a broader reset in the type of work Meta wants its suppliers to support.
Contract workers often feel these shifts first because vendor agreements are easier to change than permanent workforces. A platform can slow a project, reprice a contract or move work between suppliers without the same visibility that comes with direct layoffs. That makes the Covalen case a useful warning for other service firms tied to large technology clients: AI strategy is not just a product decision, it is becoming a procurement decision too.
Global Pattern
Meta's latest planned cuts would be its largest workforce reduction since 2023, when the company moved through a much deeper efficiency drive after years of rapid hiring. The difference now is the source of pressure. Meta is not pulling back from growth. It is redirecting spending from broad headcount toward infrastructure and AI systems that management believes can lift productivity across the business.
That pattern is visible across the technology sector. Companies are still hiring for AI research, infrastructure and applied model work, but they are cutting or freezing roles in areas where automation, slower growth or supplier consolidation can reduce costs. Layoffs.fyi has tracked tens of thousands of technology job cuts in 2026, and AI is increasingly part of the explanation, even when companies avoid making it the headline reason.
Entrepreneurship Implications
For startups, there are two practical lessons. The first is that talent may become more available in areas such as trust and safety, AI data operations, content quality and platform support. These are not marginal skills. They are exactly the operational skills many AI startups need as they move from demos to products that must work reliably in the real world.
The second lesson is more uncomfortable. If a young company depends heavily on one enterprise client or one platform contract, it inherits that client's strategic swings. Meta's AI spending may create opportunities for infrastructure suppliers and model builders, but it can also compress margins for service providers whose work is seen as replaceable, automatable or outside the next investment cycle.
The immediate issue is whether Covalen, the union and Irish officials can limit the damage for workers in Dublin. The wider story is bigger than one vendor. As AI spending moves from promise to balance-sheet reality, the next phase of the market will be defined by who captures that investment and who is asked to pay for it.
Also read: Archbald's council implodes as six AI data centers threaten to swallow a Pennsylvania town • Google grabs the Pentagon's AI business after Anthropic drew its red lines • AI agents are killing fixed app interfaces, turning software into disposable pixels
