NEW YORK, May 7, 2026, 7:04 PM EDT
Innodata Inc. raised its 2026 revenue growth forecast Thursday after reporting first-quarter results that blew past expectations on both revenue and adjusted profit. Revenue jumped 54% to $90.1 million. New deals with a major tech client may add around $51 million to this year’s top line, the company said, pointing to continued appetite for AI training and evaluation services among smaller data-services players.
It’s catching investor attention now as the focus shifts past just chips or cloud servers to the behind-the-scenes labor essential for building, testing, and fine-tuning AI. Innodata calls itself a partner for AI development, training, post-training tasks, evaluation, and deployment—they work with both tech firms and frontier AI labs.
INOD shares slipped 1.9% to end the regular session at $45.64, but after hours they surged 29.1% to $58.91 as of 6:38 p.m. Eastern, according to MarketBeat. The company put out its report after the bell, with the earnings call set for 5 p.m. ET.
Net income jumped to $14.9 million, or 42 cents per diluted share, compared with $7.8 million, or 22 cents per diluted share, in the same period last year. Adjusted EBITDA came in at $25.0 million—28% of revenue—after adding back taxes, interest, depreciation, amortization and stock-based compensation to net income, per the company’s non-GAAP metric.
Chief Executive Jack Abuhoff described the period as a “record-setting quarter” and pointed to gains in scale, margin, and cash flow coming from the company’s strategy. Innodata now expects full-year 2026 revenue growth of roughly 40% or higher, an increase from the 35% or more it forecast just 10 weeks ago. ACCESS Newswire
The major variable here is the new big tech client. According to Abuhoff, Innodata saw zero revenue from this customer a year ago. Now, though, “we expect it to become our second-largest customer” this year. ACCESS Newswire
The focus now extends past simple data labeling. During the call, Innodata disclosed that the major contract covers pre-training, mid-training, and post-training data—datasets essential for training and refining large language models, or AI systems built on extensive text, code, images, and more. Innodata also mentioned it’s handling model evaluation and trust-and-safety work, checking if AI systems function properly ahead of launch.
The field’s crowded: Innodata’s annual filing flags Appen, TELUS Digital, and Scale AI as competitors in AI data engineering, training, and evaluation. Broader consulting and outsourcing players pitch similar services. The company points to service quality, technical depth, scale, and security as its differentiators.
The story isn’t without caveats. On the call, BWS Financial’s Hamed Khorsand pressed about any one-off drivers this quarter. Abuhoff responded, “things do start and stop” in model-training work, but he maintained that management doesn’t see the quarter as a fluke. MarketBeat
Customer concentration and project turnover stand out as the biggest risks. In its annual filing, Innodata disclosed that a single client represented roughly 58% of 2025 revenue. Many of Innodata’s customer deals are project-based, with most able to be ended with just 30 to 90 days’ notice. The company’s latest quarterly report also flagged that anticipated projects, sales pipeline activity, and talks with clients might never generate the revenue management expects.
Right now, investors seem to be reading the quarter as proof that Innodata’s footprint in the AI supply chain runs deeper than its market cap implies. But the real question is if those fresh programs will actually turn into ongoing revenue streams—or if this is just a spike in after-hours action.
