Innodata Is Making AI Models Smarter. Buy the Stock.

(Illustration by Lynne Tuttle/Barron’s)

Buying Innodata stock offers an opportunity to invest alongside a leading player in data services for training AI.

INOD

Innodata Inc.

1-Year Price Chart

Created with Highstock 2.1.8

$36.68

as of market close September 3, 2025

Market Cap

$1.2 B

NTM P/E

43.7

Div Yield

0%

Beta

2.62

52 Week Range

$13.02

$71.00

Advancements in artificial intelligence can seem like magic, but the real secret behind the latest breakthroughs isn’t in the chips or the algorithms running AI models—it’s in the quality of the underlying data used to train them.

That is the message from Ridgefield Park, N.J.–based Innodata , a leader in AI data engineering. The company is a critical partner to numerous technology giants, offering software that transforms raw data such as text, images, video, and sensor information into high-quality, AI-ready data sets. This work is increasingly important as developers build more-complex and sophisticated models.

With a modest $1.2 billion market capitalization, Innodata’s financial performance tells the story. In the past three years, revenue has tripled, accompanied by a surge in profitability that has powered the company’s stock to a spectacular 11-fold increase over the period. Despite the large gains, there’s a sense that Innodata’s growth is just getting started. Following a sharp correction from highs earlier this year, the stock looks like a good bet for investors ahead of a rebound.

Innodata’s success has been a long time in the making. Leveraging a 35-year history in niche data services, Innodata was early in recognizing the explosive potential of AI and machine learning, making key investments over the past decade. In 2023, the company found itself at ground zero of the AI boom, pivoting to focus on servicing generative-AI development and deployment.

The core business idea is simple. Innodata cleans up inaccurate, incomplete, or irrelevant data, combining a proprietary technology platform with over 6,000 expert consultants to annotate and validate information for various AI applications. This human-in-the-loop framework provides specialized services, such as fine-tuning and preference optimization to teach large language models to control their tone, avoid biases, and ultimately reduce errors.

Playing coy to avoid naming its specific customers, Innodata notes that five of the Magnificent Seven tech leaders are using its services. Business from its biggest customer has increased from an initial $8 million contract in 2023 to a $135 million annualized revenue run rate, validating Innodata’s AI capabilities.

Demand for its services is widespread. In the second quarter of 2025, Innodata’s revenue reached $58.4 million, up 79% year over year, with management citing a robust pipeline of deals. The surge in earnings was also impressive, with net income of $7.2 million reversing a loss in the prior year. The results were strong enough that management hiked its 2025 guidance, forecasting annual revenue growth of 45% or more.

Maxim Group analyst Allen Klee points out in a recent research note that emerging computing techniques and hardware optimization “should create greater demand for data training” as attention turns to the quality of data, adding to Innodata’s tailwind of operating leverage. Klee has a Buy rating on the stock with a price target of $75, implying more than 100% upside from Wednesday’s close of $36.68.

As Innodata President and CEO Jack Abuhoff explained in a letter to shareholders earlier this year: “Enterprises can turn to Innodata when their in-house resources or competitor vendors fall short—especially for high-volume, high-complexity projects that require domain expertise, rapid scaling, or stringent quality standards.”

In other words, as AI evolves, the trend is for these types of data solutions and AI services to become more specialized, opening the door for Innodata to consolidate its leadership position and diversify its customer base.

That latter point is critical. As strong as the company’s growth momentum has been, its two largest customers contributed to more than half of total revenue over the past year. The dynamic raises questions regarding Innodata’s reliance on a select number of accounts, which the company is addressing by targeting expanded AI applications and what it sees as the future of the field: agentic AI.

This category of applications can autonomously initiate and carry out complex tasks with minimal human input. Innodata sees this new generation of AI unlocking the full value of large language models, creating new use cases, including robotics, for a range of industries.

That future would demand new layers of data safeguards to keep agents on task, underscoring the company’s opportunity in AI trust and safety. Its platform-agnostic approach may allow it to succeed regardless of which AI models—from Alphabet, Meta Platforms, OpenAI, or other players—come out on top. Innodata should grow as new enterprise customers build out agents specific to their businesses.

The allure of the company’s stock is clear. Its price/earnings ratio of 30 looks attractive, especially in view of the company’s growth trajectory. In contrast to other small-cap AI disruptors like C3.ai or BigBear.ai Holdings, which are struggling to reach profitability, Innodata is not only making money but also sports a clean balance sheet, with zero debt and growing free cash flow.

Common risks cited by Wall Street analysts include the possibility of materially weaker growth or a scenario in which AI training data loses its importance, either of which would pressure shares to the downside. The company also faces competition in the nascent industry from both larger tech companies offering alternative data solutions as well as other specialized players like Scale AI. Yet Innodata’s ability to exceed expectations should help it shed any doubts about whether its competitive edge is sustainable.

For investors confident that we’re still in the early innings of the AI revolution, Innodata stock is a high-quality small-cap to buy and hold for the long run.

The Technical View

Innodata is 47% off its 52-week highs and on a seven-week losing streak. Up just four sessions in August and trading just below the 200-day simple moving average—but not far beneath it—the stock could resume the uptrend, with a break above that average. If that happens, it could really jump, given that it went from $5 to $71 between April 2024 and February this year. The stock deserves some credit for still making a series of higher lows over the last 17 months.— Doug Busch

Write to Dan Victor at dan.victor@barrons.com