Innodata (NASDAQ:INOD) reported record second-quarter results for 2026, with revenue, adjusted gross profit, adjusted EBITDA and cash reaching new highs as the company continued to expand its work with AI model developers and large technology customers.
Revenue for the quarter was $92.1 million, up 58% from a year earlier and 2% sequentially, marking Innodata’s 12th consecutive quarter of year-over-year growth. Chairman and CEO Jack Abuhoff said the result exceeded analyst consensus by approximately $5.8 million.
Customer mix shifts as revenue base broadens
Abuhoff said Innodata’s customer concentration shifted during the quarter. Its largest customer accounted for 37% of revenue, down from 56% in the first quarter, following a change in program structure and service mix. The company said it still expects that customer to grow year over year for the full year.
Meanwhile, a big technology customer announced in the prior quarter increased from 17% of revenue to 34% and became Innodata’s second-largest customer. The company also said it added a new customer that it described as one of the fastest-scaling frontier AI labs.
“Our base continues to broaden in both customers and customer programs,” Abuhoff said.
The company reiterated its forecast for at least 40% year-over-year revenue growth. Abuhoff said Innodata has large potential engagements in its pipeline involving existing and new customers, but it has not included them in its forecast because the work has not yet been fully secured and the timing of revenue recognition is not yet known.
Research and innovation drive AI programs
President and Chief Revenue Officer Rahul Singhal said research and innovation are increasingly becoming a growth engine for Innodata, supporting work across model pre-training, post-training, evaluation and benchmarking.
Singhal highlighted the company’s work in agentic reinforcement learning, including a program with a large AI lab related to personalizing long-horizon agents and a separate program involving reinforcement-learning environments for desktop computer tasks. He said Innodata deepened delivery of these capabilities with one large technology customer during the quarter and began delivery with another.
Innodata also released two public benchmarks designed to assess frontier models on multi-turn, long-context and multimodal interactions. Singhal said the benchmarks are intended to identify failure modes that conventional leaderboards may miss, including grounding drift and instruction forgetting. The company said benchmark engagements can lead to data strategy recommendations and potential scaled data-generation work.
The company expanded training-data generation for reasoning capabilities across five frontier labs and five domains, Singhal said. It also signed two research agreements with a leading university and committed to a motion-capture lab expected to come online in coming months. The lab is intended to collect sub-millimeter-precision data for robots and physical-AI foundation models.
In addition, Innodata released the first stage of its AI Cyber Training Suite, consisting of 12 datasets and evaluation systems intended to help AI coding agents write more secure code and repair software vulnerabilities. Singhal said that, when testing leading open-weight models, the rate of repair for verified flaws more than doubled after one round of fine-tuning on a portion of the company’s data.
Margins, balance sheet and capital-markets flexibility
Abuhoff attributed the 49% adjusted gross margin partly to a higher mix of high-value pre-training programs and off-the-shelf datasets. He said some datasets are engineered around model weaknesses identified through Innodata’s benchmarking work, with the company retaining intellectual property and making the assets available across multiple customers.
In response to an analyst question, Abuhoff said gross margin may vary by quarter because Innodata could pursue larger projects with lower margins. However, he said the company’s long-term strategic focus is on higher-quality revenue, defined by both gross margin and recurring characteristics.
Innodata ended the quarter with $250.4 million in cash and short-term investments. Excluding customer prepayments, which Chief Financial Officer Jayant Chauhan described as pass-through items, cash and short-term investments were approximately $134 million, up $37 million sequentially. The company had no debt outstanding and remained undrawn on its Wells Fargo credit facility.
Chauhan said Innodata planned to file a prospectus supplement establishing an at-the-market equity program led by Goldman Sachs alongside a broader syndicate. He said the program would provide a supplemental capital-markets tool that the company could use selectively for growth initiatives, strategic opportunities and balance-sheet flexibility.
Leadership transition set for September
Innodata also announced a planned leadership transition effective Sept. 30. Singhal will become president and chief executive officer and join the board of directors. Abuhoff will transition to executive chairman.
Abuhoff said he will remain deeply engaged and focus on partnering with Singhal to build capabilities enabled by Innodata’s research team, particularly in federal government and enterprise markets. The company also recently appointed Chauhan as CFO, while Marissa Espineli was identified on the call as chief accounting officer.
Abuhoff said Innodata sees potential federal opportunities in model benchmarks, evaluations and red-teaming work, as government agencies seek AI solutions and frontier-model developers engage with government on regulation. He also cited enterprise demand for greater assurance around agentic AI deployment and cybersecurity as areas where Innodata believes its data engineering and evaluation capabilities can be applied.
About Innodata (NASDAQ:INOD)
Innodata Inc (NASDAQ: INOD) is a digital services and technology company that specializes in data engineering and artificial intelligence solutions. Founded in 1988 and headquartered in East Brunswick, New Jersey, the company provides structured content and digital transformation services to publishers, media companies, legal and compliance organizations, and other information-intensive industries. Innodata’s platform enables clients to convert unstructured text, images and multimedia into high‐quality, machine‐readable formats that support search, analytics and AI model training.
The firm’s offerings include content enrichment, metadata management, taxonomy development, digital asset management and data annotation services.
