Innodata is a global data engineering company providing AI training data, data annotation, and AI systems services primarily to large technology companies. It operates three segments: Digital Data Solutions (DDS), which is the dominant revenue driver focused on AI-related data work; Synodex, a medical data digitization business; and Agility, a media intelligence SaaS platform. The company has repositioned itself around generative AI data pipelines over the past two years.
Revenue and operating income growth of 48% and 63% respectively are offset by a single-customer concentration of 58% on short-notice-terminable contracts, a 179% surge in stock-based compensation that inflates non-GAAP metrics, and deteriorating margins in two of three segments.
The company delivered 48% revenue growth and 63% operating income growth in 2025 with stable gross margins, demonstrating that its AI data engineering services are in strong demand. If the dominant customer continues to expand its AI training data requirements, Innodata has a scalable offshore labor model that can grow revenues with only modest incremental fixed cost. The Agility SaaS segment provides a small but recurring revenue cushion.
A single customer at 58% of total revenue, up from 48% the prior year, on contracts terminable with as little as 30 days notice represents an existential concentration risk. The Synodex segment is deteriorating with gross margin falling from 27% to 18% and a customer contract already terminated. Stock-based compensation nearly tripled year over year to $11.1 million, inflating non-GAAP metrics and representing real dilution not captured in the headline Adjusted EBITDA figure of $57.9 million.
The filing does not contain explicit forward-looking revenue or earnings guidance for fiscal year 2026. Management references ongoing investment in new capabilities to pursue growth and notes that new customer volumes may not meet initial expectations, but provides no quantitative targets.
Innodata competes in the highly fragmented AI data services and annotation market against larger, better-capitalized vendors as well as in-house teams at technology companies. The company has limited disclosed proprietary moat beyond its workforce scale and existing relationships, and acknowledges that there are relatively few barriers preventing companies from entering its markets. Its competitive position is heavily dependent on a single unnamed AI platform customer that now represents 58% of revenue.