Innodata Inc. (INOD) Earnings

Management highlights

### Leadership Announcements - A planned leadership transition will take effect September 30: current President and CRO Rahul Singhal will become President and CEO and join the board of directors, while current CEO Jack Abuhoff will transition to Executive Chairman. - Jack Abuhoff will remain deeply engaged, focusing on developing new research-enabled capabilities for the federal government and enterprise markets to drive shareholder value. - Jayant Chauhan joined InnoData as the new CFO, with a mandate covering traditional finance responsibilities, strategic capital allocation, M&A, customer partnerships, and investor communications, while Marissa Espineli stepped into the role of Chief Accounting Officer. ### Financial and Business Highlights - Q2 2026 was a record quarter, with all key financial metrics exceeding analyst consensus, marking the 12th consecutive quarter of year-over-year revenue growth. Quarterly revenue now exceeds the company's full annual revenue from three years prior. - The customer base is continuing to diversify: the share of revenue from the largest customer decreased quarter-over-quarter while overall total revenue grew, and a new big tech customer scaled to become the second-largest customer in just one quarter. - The company delivered strong operating leverage: adjusted EBITDA grew 1.6x faster than revenue, as incremental new programs build on existing core operating infrastructure, leading to lower marginal costs. ### Innovation and Strategic Milestones - Research and innovation have become a core growth engine, driving differentiation across the full AI model training lifecycle, from pre-training to post-training, evaluation, and benchmarking. - The company established an early position in agentic reinforcement learning, a key AI frontier, with one large program scaling and a second new program awarded for desktop agentic task reinforcement learning environments. - The company positions itself as an AI deployment assurance layer for enterprise AI, combining observability platforms and reinforcement learning gyms, deepened delivery with one big tech customer, and launched new pilots for insurance and banking clients. - Two new public benchmarks for dynamic, long-horizon agentic model evaluation were released, designed to detect failure modes (such as grounding drift and instruction forgetting) that standard benchmarks miss, opening opportunities for scaled data generation engagements. - Expanded capability to deliver reasoning-enhancing training data across 5 frontier labs and 5 domains. Established new research partnerships for embodied/physical AI, including agreements with a leading university and a planned sub-millimeter precision motion capture lab for robot training data, with successful pilots completed with leading robotics companies and enterprise-scale multimodal programs currently scoped. - Launched the full AI Cyber Training Suite, which doubles the verified flaw repair rate of leading open-weight AI coding models after fine-tuning, addressing a key barrier to enterprise AI adoption.

Guidance

- Management maintains the prior full-year guidance of 40% or more year-over-year revenue growth, with no upward or downward revision. - The company has a pipeline of large, high-probability potential new engagements with both existing and new customers that have not yet been factored into the current guidance. Management will only add these opportunities to forecasts once contracts are finalized and revenue recognition timing is confirmed as a matter of prudence.

Segment performance

The transcript does not break out financial performance for distinct formal product segments. Overall company Q2 2026 revenue is $92.1 million, up 58% year-over-year and 2% sequentially. Adjusted gross profit is $45.4 million with an adjusted gross margin of 49%, up 2 points sequentially and 9 points above the company's 40% target. Adjusted EBITDA is $25.4 million (27.5% of revenue), up 92% year-over-year. Net income is $14.4 million, double the year-ago quarter. Fully diluted EPS is 41 cents. By customer contribution, the largest customer accounted for 37% of total revenue (down from 56% in Q1), and the second-largest big tech customer accounted for 34% of revenue (up from 17% in Q1). Higher-margin programs drove overall margin expansion in the quarter, with higher-margin offerings including pre-training programs and off-the-shelf datasets that the company owns and can monetize across multiple customers.

Risks & headwinds

No specific internal operational failures or new material risks were explicitly discussed during the call. Management noted that forward-looking statements are inherently uncertain, and actual results could differ materially from projections due to general market and business risks, which are detailed in the company's existing SEC filings.

Analyst Q&A

• Q: Can you describe the large unbooked opportunities not included in current guidance, and will you provide updates outside of quarterly announcements? /

  A: These unguided opportunities span all of the company's segments: government, enterprise, and frontier AI models, with agentic AI training and deployment capabilities making up a large share of the pipeline. Management maintains a disciplined approach of only counting opportunities once they are fully contracted, and will share more details as the opportunities progress through the second half of the year.

• Q: How do recent AI security incidents create new opportunities for InnoData's newly launched AI Cyber Training Suite? /

  A: Many security incidents stem from frontier models being trained on unpatched, flawed training data. InnoData's suite can detect existing code vulnerabilities, enable AI to generalize to novel threats, and help enterprises adopt generative AI more safely. The company is already in active discussions about this capability, which is an example of how the company's data layer innovation addresses unmet market needs.

• Q: Is the current quarter's higher mix of high-margin projects a sustained trend or just a one-off quarter? /

  A: Strategically, the company's ongoing innovation will produce higher-quality revenue with higher gross margins over the long term, but quarter-over-quarter margin will vary based on program mix. The company will still bid on and accept lower-margin large projects if they deliver compelling cash flow, which could lead to temporary weighted margin declines in some quarters.

• Q: Could sequential revenue decline in Q3 or Q4 2026? How should investors view that outcome? /

  A: A sequential revenue decline is possible within the company's business model (for example, after a large one-time project delivered across two quarters), but management does not view this as a failure. The company's top priority is long-term growth and sustained customer relevance, not steady quarter-over-quarter revenue expansion, and the current pipeline of large opportunities supports significant multi-year growth.