I think so. For me, it is $INOD (-11,68Â %) one of the most interesting stocks in the AI sector. There are also other assessments that underpin this.
Aktienwelt360 - When people talk about artificial intelligence, many people first think of Nvidia, OpenAI or Microsoft. But while these giants dominate the headlines, there is a company working in the background that is providing the foundation for all their advances: Innodata (WKN: 907651). I've looked at the numbers, the strategy and the risks - and one thing is for sure: if you really want to understand the AI revolution, Innodata is a must-have.
From niche service provider to key AI player
Innodata has been around for over 35 years, but it was only with the rise of generative AI that the company found its place in the spotlight. Today, it provides data engineering services that are essential for training and fine-tuning AI models. It sounds technical - but at its core, it's simple: Innodata transforms huge amounts of raw data (text, images, videos, sensor information) into 'clean', AI-ready datasets.
This work is crucial, because without high-quality training data, even the best algorithms remain blind. This is precisely why five of the Magnificent Seven now rely on Innodata.
The business model is bearing fruit
Turnover rose from 86.8 million US dollars (2023) to 170.5 million US dollars (2024) - a growth of 96%. Even more impressive: a loss of USD 0.9 million turned into a net profit of USD 28.7 million.
This trend will continue in 2025. In the second quarter alone, turnover amounted to USD 58.4 million, an increase of 79% compared to the previous year. Net profit climbed to USD 7.2 million, following losses in the previous year. The EBITDA margin is now over 20% - a strong signal of operational efficiency.
The secret
What makes Innodata special is its human-in-the-loop approach. This combines automated processes with human expertise. More than 6,000 specialists check and correct data to make AI systems more robust, more precise and less prone to error.
This is not an old-fashioned approach - on the contrary. Especially when it comes to bias, ethical issues or highly complex content, human control remains indispensable. Innodata has developed a scalable business model that combines efficiency and quality.
Strategic key accounts and record contracts
Innodata has built up enormous contract volumes in a short space of time. A contract with a volume of USD 8 million began in 2023 - the same customer now generates USD 135 million annually. Further programs worth USD 20 million (April 2024) and USD 44 million (June 2024) were added.
However, this success also harbors risks: The two largest customers together contribute more than half. If one of these tech giants were to drop out, the consequences could be severe.
A market with enormous potential
The market for AI training data is growing rapidly. Current studies estimate annual growth rates (CAGR) of between 20 and 25 % - from just under USD 2.7 billion at present to up to USD 16 billion by 2032. The submarket for synthetic data is particularly exciting and is forecast to grow by over 40 % annually.
Innodata is well positioned in both areas. The company is evolving from a pure data provider into a smart data specialist that analyzes AI models, identifies weaknesses and provides targeted data sets for improvement.
The competition is intense
Scale AI, Appen, Labelbox, SuperAnnotate and iMerit are all vying for market share. Scale AI in particular is considered a powerful rival - the company was recently valued at USD 29 billion and received a USD 14 billion investment from Meta.
Innodata, on the other hand, is pursuing a platform-independent strategy that enables cooperation with various players such as Alphabet, OpenAI and Meta. This reduces the risk of conflicts of interest and increases flexibility.
The next step
While many competitors are still working with traditional training data, Innodata is already preparing for the next generation of AI: Agentic AI. These systems can initiate and execute tasks independently - without constant human intervention.
This creates completely new requirements in terms of data quality, security and control. Innodata wants to play a leading role here - especially in ensuring that AI agents act reliably and ethically.
My conclusion
For me, Innodata is not a classic AI hype title, but a true infrastructure champion in the industry. Without companies like this, neither ChatGPT nor Copilot or Gemini could function reliably.
Anyone who believes in the long-term success of artificial intelligence will find a profitable, agile and intelligently managed company here that is in exactly the right segment.
