Innodata (INOD) Q3 2025: $68M Pre-Training Data Wins Signal Transformative AI Growth
Innodata’s Q3 showcased record revenue and margin expansion, but the core story is the company’s rapid pivot to high-value pre-training data and federal AI contracts, now underpinning a 2026 growth inflection. With $68 million in new pre-training data programs and a $25 million federal win, INOD is increasingly central to the generative AI supply chain and positioning for sovereign and enterprise AI demand.
Summary
- AI Data Expansion: Pre-training data programs and federal contracts are reshaping Innodata’s revenue base.
- Margin Upside: Operating leverage is visible even amid elevated investment in new growth vectors.
- 2026 Growth Setup: Multi-year, multi-segment deal momentum sets the stage for accelerated expansion.
Performance Analysis
Innodata delivered record Q3 revenue and margin expansion, with revenue up 20% year-over-year and adjusted EBITDA margin reaching 26%. The company’s cash position strengthened to $73.9 million, reflecting both core business momentum and disciplined capital management. Importantly, these results came despite $9.5 million in incremental capability-building investments, signaling robust underlying profitability.
Growth is increasingly powered by new AI data initiatives, notably a $68 million pipeline in pre-training data contracts and a $25 million federal project win. While legacy post-training data work remains stable, the business is shifting toward higher-value, less commoditized segments. Operating leverage was evident as adjusted EBITDA rose 23% quarter over quarter, even as SG&A and CapEx ramped to support new verticals. The company’s largest customer relationship remains stable, with verbal confirmation of further expansion, and new big tech and sovereign AI customers are poised to contribute materially in 2026.
- Pre-Training Data Ramp: $68 million in new pre-training data contracts—signed or near close—represent a major revenue vector for 2026.
- Federal AI Entry: The $25 million initial federal win and direct government award validate Innodata’s new InnoData Federal unit.
- Margin Expansion: Adjusted EBITDA margin rose to 26% despite absorbing $9.5 million in growth investments.
With both core and emerging businesses scaling, Innodata is positioned to capture rising demand from hyperscalers, sovereign AI initiatives, and enterprise AI transformation.
Executive Commentary
"We delivered record revenue of $62.6 million, representing a 20% year-over-year organic growth and a 7% sequential quarterly growth. Adjusted EBITDA was $16.2 million, or 26% of revenue, up 23% sequentially, showing margin expansion even after factoring in growth investments... We anticipate potentially transformative growth in 2026."
Jack Abelhoff, CEO
"Looking past 2026, over the medium and long term, we believe the work we do with frontier model builders will expand and will become more complex. The next generation of models won't just need more data. They'll need more smarter data... We are at the very beginning of the generational technology shift that InnoData is at the center of and poised to capitalize on."
Rahul Singhal, President and Chief Revenue Officer
Strategic Positioning
1. Pre-Training Data as a New Revenue Engine
Innodata’s move into pre-training data, the foundational datasets that teach large language models (LLMs) language and knowledge, marks a strategic leap. Historically focused on post-training data—fine-tuning models for reasoning and task performance—the company invested $1.3 million to build pre-training capabilities. This investment has already yielded $42 million in signed contracts and $26 million in near-term pipeline, with the majority of revenue expected in 2026. This segment is now a core growth pillar, providing higher visibility and stickier relationships with AI model builders.
2. InnoData Federal: Securing Government AI Spend
The launch of InnoData Federal positions the company in the accelerating federal AI market. With a $25 million initial project from a high-profile defense customer and additional proposals in the pipeline, Innodata is now competing for mission-critical AI lifecycle projects—data collection, model deployment, and operational support. Regulatory tailwinds, such as executive orders streamlining AI procurement, further lower barriers for rapid federal expansion.
3. Sovereign AI and Enterprise AI Practice
Global sovereign AI initiatives and enterprise AI integration are emerging as high-potential verticals. Governments seeking national control over AI technology stacks are triggering large-scale, state-backed investments, and Innodata is in advanced partnership discussions in the Middle East and beyond. Meanwhile, the enterprise AI practice is gaining traction, with projects supporting content monitoring, monetization, and real-time analytics for major tech and hyperscaler clients.
4. Agentic AI and Model Safety
Agentic AI—autonomous agents for enterprise workflows—and model safety services are early but strategically significant bets. Innodata is piloting agent evaluation and safety benchmarking for leading chip and software companies, aiming to become the trusted partner for safe, reliable AI deployment as autonomous systems proliferate.
Key Considerations
This quarter marks a transition from point-solution data services to strategic, multi-year AI infrastructure partnerships. Investors should focus on the durability and scalability of these new vectors.
Key Considerations:
- Customer Concentration: While the largest customer remains stable, expansion into five additional big techs and sovereign entities is crucial for revenue diversification.
- Short-Cycle Investment Payoff: Modest capital outlays in new verticals are already generating outsized returns, validating the company’s rapid iteration approach.
- Federal and Sovereign Pipeline: Successful execution on federal contracts and sovereign partnerships could materially alter Innodata’s revenue mix and risk profile.
- Operating Leverage: Margin expansion amid investment underscores a scalable business model, but future SG&A and capacity absorption must be monitored as new projects ramp.
Risks
Execution risk remains elevated as Innodata scales new business lines and absorbs excess capacity ahead of contract ramp. Federal procurement cycles, while accelerating, can still be unpredictable. Customer concentration with big tech persists, making diversification efforts critical. Competitive intensity in AI data services is rising, and any delays in contract conversion or market adoption could impact growth visibility.
Forward Outlook
For Q4 2025, Innodata guided to:
- Continued sequential revenue growth as new contracts ramp.
- Margin stability with ongoing investment in capability building.
For full-year 2025, management reiterated guidance:
- 45% or more year-over-year revenue growth.
Management highlighted several factors that underpin 2026 acceleration:
- Contracted and near-contracted pre-training data programs with multi-million dollar annual potential.
- Federal and sovereign AI pipeline conversion, with initial wins expected to expand into multi-year, multi-project relationships.
Takeaways
Innodata’s Q3 execution validates its pivot into foundational AI data and government AI solutions, with new contracts providing multi-year visibility. The company’s ability to land and expand with hyperscalers, sovereigns, and federal agencies will determine the durability of its growth and margin story.
- AI Infrastructure Shift: The transition from post-training data to pre-training and full-stack AI lifecycle services is reshaping Innodata’s business model and competitive moat.
- Federal and Sovereign Leverage: Early traction in government and sovereign AI markets could unlock non-cyclical, high-margin revenue streams.
- 2026 Inflection Catalyst: Investors should watch for contract conversion rates, customer concentration shifts, and the scalability of new verticals as leading indicators for sustained outperformance.
Conclusion
Innodata’s record Q3 and $68 million in new pre-training data contracts signal a business at the center of AI’s next wave. With federal and sovereign AI tailwinds, the company is positioned for a potential step-change in scale and relevance, though execution and diversification remain key watchpoints.
Industry Read-Through
Innodata’s shift into pre-training data and federal AI projects highlights a broader industry trend: as generative AI adoption accelerates, demand is moving upstream from annotation to foundational data and lifecycle support. Federal and sovereign AI spending is emerging as a durable, high-growth vertical, with procurement reform unlocking new opportunities for agile, specialized vendors. Competitors in AI data, consulting, and model safety should expect intensified competition for large, multi-year contracts as governments and enterprises seek partners with scale, compliance, and end-to-end capabilities. The evolution of agentic AI and model safety offerings also signals future demand for trusted, repeatable benchmarking and remediation services as autonomous systems proliferate.