NEW YORK, May 8, 2026, 10:12 (EDT)
- Innodata shares surged roughly 86% Friday morning, with the AI data company topping Q1 estimates and raising its revenue forecast for 2026.
- The company is projecting roughly $51 million in revenue this year from fresh work with a key Big Tech client.
- Customer concentration still looms large—one client accounted for 56% of first-quarter revenue, according to a filing.
Shares of Innodata Inc (NASDAQ: INOD) shot up Friday, jumping as much as 85.7% to $84.75, after the company posted record Q1 numbers, boosted its growth forecast for the year, and announced fresh Big Tech contracts that could pull in roughly $51 million in revenue come 2026. The stock spiked to $90.13 earlier in the session.
Why does this matter? Investors want to see AI vendors convert all that excitement around model-building into actual numbers—revenue, profits, cash flow. For Innodata, first-quarter revenue jumped 54% to $90.1 million. Adjusted EBITDA hit $25.0 million, landing at 28% of revenue—this metric excludes interest, taxes, depreciation, amortization, and a handful of other items.
This angle is especially relevant as the company looks to prove it’s not overly reliant on a single major client. CEO Jack Abuhoff pointed out that a Big Tech customer, which brought in zero revenue for Innodata a year earlier, is on track to be its second-biggest account this year. Revenue from other Big Tech customers shot up 453% for the quarter.
Net income shot up to $14.9 million, or 42 cents a share on a diluted basis, compared with $7.8 million, or 22 cents, a year ago. At the end of March, Innodata was sitting on $117.4 million in cash, cash equivalents and short-term investments. The company also noted its expanded $50 million credit line with Wells Fargo hasn’t been tapped.
Abuhoff described the quarter as “record-setting,” bumping up the company’s 2026 revenue growth outlook to “approximately 40% or more”—an increase from the “about 35% or more” guidance issued just 10 weeks prior. He stuck with the word “prudent” for the new forecast, citing several sizable programs still missing from the numbers. The Motley Fool
Abuhoff broke down the fresh Big Tech project for analysts: Innodata is handling everything from pre-training and mid-training all the way through post-training and evaluation. In practice, that’s data sourcing and testing for building, fine-tuning, and assessing major AI models. He also mentioned talks are underway with the same client for more contracts not reflected in the $51 million tally.
Rahul Singhal, who serves as both president and chief revenue officer at Innodata, pointed to a shift in focus: the company is now targeting “high-quality pretraining data” as well as model evaluation for frontier AI labs—those working on advanced foundation models. According to Singhal, one of the big cloud-computing players, a so-called hyperscaler, has tapped Innodata as a global trust-and-safety partner to test models before launch. The initial statement of work is projected to bring in roughly $3 million in annual run-rate revenue. The Motley Fool
The company’s making moves on software-style offerings, too. Innodata reported a beta launch for its Evaluation and Observability Platform, plus a $1 million platform deal signed with a hyperscaler customer. Another 15 firms are testing it out. Agentic systems, for reference, are AI setups meant to handle multi-step work with reduced human input.
The landscape remains in flux. Last year, Reuters said Google intended to end its relationship with Scale AI after Meta acquired a 49% stake in the data-labeling outfit. That move gave competitors like Labelbox, Turing, and Handshake a shot, as AI labs started to reconsider the importance of supplier neutrality.
Still, risks stand out. Innodata’s Form 10-Q revealed that one customer pulled in roughly 56% of first-quarter revenue—last year, that figure was 61%. Another customer kicked in 17%. As of March 31, the filing showed a single customer owed 65% of accounts receivable.
Segment detail has gotten thinner. Starting in the first quarter, Innodata now reports as one operating segment—no more separate numbers for DDS, Agility, or Synodex. Management attributed the move to its focus on agentic AI technologies and a push for a tighter, integrated model.
So far, investors are reading the quarter as confirmation that AI model spending is reaching further than just chipmakers and cloud giants, with data, evaluation, and safety firms now in the mix. The question now: Will the Big Tech projects ramp up at the promised pace — and can the company attract enough new clients to balance out the heavy concentration still evident in the filing?
