On 24 June 2026, enterprise-AI services firm EXL announced it would acquire iMerit — one of the best-known names in data annotation — for up to $310 million: $170 million upfront and up to $140 million more in earnouts tied to performance over the next two years. The deal is expected to close in Q3 2026.
What EXL is really buying isn't just a labeling workforce. It's iMerit's Ango annotation platform and its Scholars network — vetted subject-matter experts used for reinforcement learning from human feedback (RLHF) and model evaluation. In other words, EXL paid a premium for expertise in the training loop, not for raw throughput.
Why this matters
For years, "data annotation" was treated as a commodity — cheap, offshore, undifferentiated. This acquisition is more evidence that the market has split in two. At one end, low-cost bulk labeling is being squeezed by automation and pre-labeling models. At the other, specialist, high-accuracy, domain-expert annotation is now valuable enough for a public company to pay nine figures for it.
That's the same shift we see in our own work. Annotating a potato crop for disease detection, or a field scene for a laser-weeding robot, is not something you hand to an anonymous crowd. It needs people who understand the domain, a quality process with measurable inter-annotator agreement, and traceable provenance for every label. Those are exactly the capabilities EXL just paid for.
What it means for buyers
If you buy annotation, three things follow from this deal. First, expect the big consolidators to focus upmarket on enterprise and foundation-model clients — which can leave focused, mid-scale, domain-specific work underserved. Second, "who actually does the labeling, and how is quality measured?" becomes the question that matters, not headcount. Third, the vendors who win will be the ones who can plug into a larger AI stack as the precision layer — rather than competing on price at the bottom.
For a specialist studio like H2L Robotics India, that's not a threat — it's the market moving toward what we already do: precise, auditable, domain-aware annotation for computer vision and physical AI.
The facts above are drawn from EXL's public announcement; the analysis and opinions are H2L Robotics India's own. Deal terms are subject to closing conditions and may change.
← Back to Annotation Radar