FEATURED

Innodata (INOD): The AI Data Engine Wall Street Nearly Slept On

Nobody was talking about Innodata six months ago. Now it’s the most interesting data story in the AI infrastructure trade — and the numbers explain why.

On May 8, 2026, INOD shares surged over 92% in a single session. That kind of move doesn’t happen on momentum alone. It happens when a business inflection finally catches up to what investors were still sleeping on.

Here’s what actually happened. Q1 2026 revenue came in at $90.1 million — up 54% year-over-year and 24% sequentially, blowing past the analyst consensus of $76.5 million by approximately $13.6 million, or 18%. Adjusted EBITDA nearly doubled to $25.0 million, with a 28% margin. Net income hit $14.9 million — up from $7.8 million in the prior year period — with diluted EPS of $0.42 against a consensus estimate of just $0.08. Operating cash flow jumped to $37.3 million, and the company closed the quarter sitting on $117.4 million in cash, up from $82.2 million at year-end 2025.

Management didn’t just celebrate the beat. They raised full-year 2026 revenue growth guidance from approximately 35% or more to approximately 40% or more. That’s not a rounding error — that’s a signal the pipeline is real and growing faster than the Street had modeled.

What Innodata Actually Does

Slight tangent, but it matters: most investors still think of data labeling as a commodity business. Low margins, high turnover, easily disrupted. Innodata doesn’t fit that mold anymore.

The company provides AI data engineering and model training services — pre-training datasets, fine-tuning, annotation, and increasingly, agentic AI evaluation. It’s the kind of work that determines whether a frontier model is actually reliable or just impressive in a demo. Revenue from non-primary Big Tech customers collectively surged 453% year-over-year in Q1 — a sign the customer base is broadening well beyond just one hyperscaler relationship.

A new engagement with a major Big Tech client is expected to contribute roughly $51 million in 2026 revenue alone. Management also noted that a Big Tech customer which generated zero revenue 12 months ago is now on track to become the company’s second-largest customer this year. Add a newly expanded $50 million revolving credit facility with Wells Fargo — expanded from $30 million on a three-year term, currently undrawn — and you have a company quietly building the financial infrastructure to execute at scale.

Wedbush reaffirmed Outperform and raised its price target to $100 (lifted from $80 to $100 on May 14 following initial post-earnings commentary). Maxim Group holds a $100 target. BWS Financial carries a $110–$111 target with a Buy rating. Per S&P Global, the 4-analyst consensus sits at $100.25, with a high of $111 and a low of $80. The stock recently traded near $95 — roughly in line with where consensus sits today, which means the initial re-rating move is largely behind us, but the fundamental story is still building.

The Part People Skip

Innodata’s new Evaluation and Observability Platform for agentic AI systems is the most underappreciated piece of the story. As autonomous AI agents proliferate across enterprise workflows, the demand for stress-testing, safety evaluation, and model monitoring will only grow. That platform is already supporting new customer engagements — and it’s barely in early innings.

The real risk here is concentration. Per the Q1 10-Q filing, one customer accounted for 56% of total revenue and a second customer accounted for 17% — meaning the top two customers represent 73% of the revenue base combined. If the primary relationship shifts or scales down, the model gets stressed — fast. That’s the number worth watching going into Q2.

Worth flagging separately: insider selling has been significant. CEO Jack Abuhoff sold over 243,000 shares across three days in mid-May, generating roughly $22.8 million in proceeds. Multiple other insiders — including the COO and two directors — also sold following the earnings surge. Insiders selling into strength isn’t automatically a red flag, but the scale of it warrants attention alongside any long position sizing.

INOD’s 37% EPS CAGR projected through FY2027 implies a PEG ratio under 1x at current levels — not a screaming value, but not the speculative bubble the bear case assumes either. The stock has traded between roughly $33.44 and $114.77 over the past 52 weeks. Volatility is part of the deal here. The stock carries a beta of approximately 2.40, meaning moves in both directions tend to be sharp and fast.

This is a company that was a quiet niche operator 18 months ago and is now embedded in the pre-training pipelines of multiple frontier AI labs. Whether it can convert that positioning into durable, diversified revenue — that’s the question the next two quarters will answer.

Worth a closer look before Q2 results hit on July 30.