What Innodata Inc. told the SEC could break it.

Innodata's AI-data-engineering revenue is dangerously concentrated: one customer in its Digital Data Solutions segment — one of the 'Magnificent Seven' big-tech companies it serves for LLM training/evaluation data — generated approximately 58% of total revenues in 2025 (up from 48% in 2024), and as of December 31, 2025, 63% ($29.2M) of accounts receivable was due from one customer. A pullback, in-housing of data work, pricing reset, or payment failure by this single dominant customer would be severe and immediate. It warns an even greater revenue share could concentrate further. The customer is not named, so this is captured as a concentration risk rather than a named edge. A very high single-customer concentration — the dominant risk in its profile.

“One customer in the DDS segment generated approximately 58% and 48% of the Company's total revenues in the years ended December 31, 2025 and 2024, respectively.”

Innodata delivers its data-annotation and AI-data services primarily from offshore production centers in India, the Philippines and Sri Lanka (with additional operations in Israel, Canada, Germany and the UK), while its customers are in North America and Europe. It warns these operations are in countries vulnerable to disruptions from political uncertainty, political unrest, terrorist acts and natural calamities, and is exposed to foreign-currency fluctuations on offshore production costs after pricing is set. Concentrating its delivery workforce in a few South/Southeast Asian jurisdictions creates labor-supply, geopolitical and FX exposure; disruption there would impair its ability to deliver. A genuine offshore-labor geographic concentration.

“Our operations located in India, Israel, the Philippines and Sri Lanka are in countries that remain vulnerable to disruptions from political uncertainty, political unrest, terrorist acts, and natural calamities.”

As a provider of training, evaluation and model-related data for AI systems, Innodata is exposed to a broad and rapidly evolving set of U.S. federal/state and foreign laws governing data protection, privacy, cybersecurity and the use of data in technology-enabled services — including the development, evaluation and deployment of AI (e.g., the EU AI Act, training-data/copyright rules, cross-border data-transfer restrictions). It explicitly lists 'evolving regulation of artificial intelligence' among its international risks. New AI/data rules could raise compliance costs, restrict the data it can process for clients, or dampen demand for AI-data services. A specific, structural AI/data-regulatory exposure tied to the core business.

“Our business is subject to a broad and evolving set of U.S. federal, state, and foreign laws and regulations governing data protection, privacy, cybersecurity, and the use of data in technology-enabled services”