Innodata boosted its full-year revenue growth outlook by five points in Q2, now forecasting more than 45% growth this year, driven by strong demand, significant new deal wins and a large pipeline geared toward the second-half of the year.
However, digging into the details reveals that Innodata’s largest customer has not expanded over the last two quarters, and remains at the $135 million annualized rate as stated in Q4. This comes despite Innodata discussing expansion opportunities with this customer and a second SOW signed in Q1, with growth not yet materializing.
The CEO offered an update in Q2 that they’ve won “several new projects” with their largest customer and have others in the pipeline:
“We recently won several new projects with our largest customer and we have others in pipeline that are not yet included in our forecast, but which we think are reasonably likely. Several of these new projects are under the second SOW we reported signing with this customer last quarter. We believe that the second SOW potentially gives us access to an even larger generative AI revenue pool with this customer.”
Innodata is a $1.5B market cap stock highly dependent on the announcement of new customer wins and expanding its projects with the largest customer. Our strategy is to watch price closely for the best risk/reward on an entry; and any entry will have a tight stop.
We’ve discussed in the past that Innodata appears to have a low valuation when compared to peers. Our previous analysis also went into the details of Innodata’s products, market positioning and can be found here: “Innodata: Early-Stage AI Data Engineering; Lumpy Growth”
Discovery Members: You are invited to join Knox Ridley on Monday, August 18th at 4:30 pm EST for a special Discovery webinar where he will discuss technical setups on Discovery stocks.a special Discovery webinar where he will discuss technical setups on Discovery stocks.
Revenue up 79.4% YoY
Innodata reported a slight revenue beat in Q2, with revenue coming in 3.6% ahead of estimates to $58.4 million, up 79.4% YoY. This did mark a more than 40 point deceleration in growth sequentially, yet that was expected coming into the quarter.
However, there have been some shifts in forward estimates for Q3 and Q4. For Q3, revenue estimates have been revised nearly $2 million lower since June, now projecting revenue of $59.8 million for growth of just 14.5% YoY. Q4 revenue has been revised more than $5.5 million higher to $70.9 million, for growth of almost 20% YoY. This suggest Q3 is now expected to see be a bottom of sorts before rebounding and reaccelerating into 2026.
For the full-year, Innodata raised its revenue growth guidance from 40% to >45%, now implying revenue of >$247 million, up from $239 million as of Q1. It is important to keep in mind the fluid nature of Innodata’s business, and that any new contractual agreements or expansions could have a large and/or immediate impact on revenue.
Management stated that the guidance hike was driven by “significant new deals that have been finalized since our last call as well as several deals that we believe are highly likely to close in the near term.” Any of these new deals could easily boost growth depending on how quickly they scale.
One key point that could be behind the post-earnings sell-off was some comments made by management around customer engagement and deal sizes. CEO Jack Abuhoff said that with one Big Tech customer, Innodata was “recently awarded a number of significant engagements… enabling us to forecast $10 million of revenue from this customer in the second half of this year,” up from $0.2 million over the last four quarters. Given the size of the revenue guide raise at just $8 million implied, this comment raises some questions about growth in 2H from other core customers.
Largest Customer Accounts for 58% of Revenue
Innodata provided only a brief update on customer progress this quarter, one with its largest customer and the other being the $10 million revenue opportunity discussed above.
With its largest customer, Innodata said that it won “several new projects,” with some of these being under the second scope-of-work (SOW) signed last quarter. Management added that there are other projects in the pipeline with the customer that are not yet included in its forecast, but are “reasonably likely” to be signed in the future.
Innodata’s largest customer accounted for 58% of revenue in Q2 and 59% of revenue for 1H, implying contributions of ~$33.9 million in Q2 and $68.9 million in 1H. This represents no change from Q4 (ie. no expansion) when it was stated the customer was at a $135 million annualized run rate, or ~$34 million quarterly.
Innodata also had no other >10% customers, implying the remaining customers are not spending more than $5.8 million per quarter with the company. The forecast for $10 million in 2H from another customer would likely make said customer Innodata’s second largest and a second >10% customer.
AI Segment Grew 99% YoY on Top of Tough Comps
DDS remains the key driver of Innodata’s growth due to its role in handling AI data preparation, labeling and annotation, AI training and related services.
DDS revenue grew 99% YoY to $50.58 million, though this was slightly down sequentially from $50.83 million in Q1. Although the segment’s growth barely dropped out of the triple-digit range, what’s impressive is that revenue still practically doubled YoY against a 93% growth comp.
- Synodex revenue rose 4% YoY to $2.06 million, slowing from nearly 8% growth in Q1.
- Agility revenue rose 11.5% YoY to $5.75 million, flat with growth from Q1.
Margins and Adjusted EBITDA expand significantly YoY
Innodata’s margins have significantly expanded YoY, which helps set Innodata apart from stocks in the $1B or $2B market cap range (Innodata is at $1.5B market cap). There was a marginal sequential decline QoQ yet not enough to matter in terms of the overall improvement seen from strong margin and adjusted EBITDA expansion over the past few quarters.
For example, GAAP operating margin was up 14 points YoY from 1.3% to 15.3% in the current quarter.
- Q2 GAAP gross margin was 39.4% in Q2, down nearly half a point sequentially but up nearly 11 points YoY. Adjusted gross margin was 42.9%, down slightly from Q1 but up more than 9.5 points YoY.
- Q2 GAAP operating margin was 15.3%, up 1.1 points sequentially and up more than 14 points YoY.
- Q2 GAAP net margin was 12.4%, down 1 point sequentially but up from approximately 0% in the year ago quarter.
Turning to adjusted EBITDA — Innodata saw a slight sequential improvement in adjusted EBITDA margin in Q2, though on a YoY basis, adjusted EBITDA margin expanded more than 14 points. Innodata had guided for YoY growth in adjusted EBITDA with no further clarity, and has currently reported nearly $26 million for 1H ’25 versus $34.6 million for all of FY24.
- Consolidated adjusted EBITDA margin was 22.7%, up from 21.8% in Q1 and 8.6% in the year ago quarter.
- DDS adjusted EBITDA margin was 24.2%, up from 22.7% in Q1 and just 5% in the year ago quarter. Tracking DDS’ adjusted EBITDA is important considering the segment accounts for more than 92% of consolidated adjusted EBITDA.
- Synodex adjusted EBITDA margin was 22.3%, up from 20.8% in Q1 but down from 26.3% in the year ago quarter
- Agility adjusted EBITDA margin was 9.7%, faring the worst out of the three segments as this was down 4 points from Q1 and down nearly 10 points YoY.
EPS Beat by 80% yet EPS expected to decelerate in coming quarters
Innodata handily beat on EPS in Q2, reporting $0.20 in GAAP EPS versus consensus estimates for $0.11. Looking ahead, Q3 EPS estimates are following revenue in moving slightly lower, while Q4 estimates have moved slightly higher. Since our last report, Innodata: Early-Stage AI Data Engineering; Lumpy Growth, here’s how estimates have changed:
- Q3 EPS has been revised $0.03 lower, from $0.17 to $0.14, for a YoY decline of more than (73%), though as a reminder the year-ago comp is technically inflated from a $5.9 million income tax benefit.
- Q4 EPS has been revised $0.02 higher, from $0.19 to $0.21, for a YoY decline of (31.2%).
For FY25, Innodata is expected to report EPS of $0.76, down (14.6%) YoY before rebounding to 34.5% growth in FY26 to $1.02. To note, FY26’s EPS estimate is still unchanged since June’s writeup.
Cash Flow Declines 60% QoQ
Another blemish in Q2’s report were cash flows, which declined more than (60%) sequentially. Innodata’s balance sheet remains healthy, and the company still has the entirety of its $30 million credit line available.
- Operating cash flow was $4.23 million, down (61%) QoQ. OCF margin was 7.3%, down more than 11 points QoQ. For 1H, operating cash flow was $15.1 million, up more than 139% YoY.
- Free cash flow was $2.5 million, down more than (70%) QoQ. Free cash flow margin was 4.3%, down more than 10 points QoQ. For 1H, free cash flow was $11.0 million, up nearly 5x YoY.
- Cash and equivalents totaled $59.8 million, and debt remained zero.
- Deferred revenue was $6.5 million, down from $8.0 million in Q1.
Earnings Call Q&A
There were two topics on the earnings call Q&A session worth highlighting. The first was questions about Scale AI and if Innodata will see more customers onboard as a result of Meta’s large investment of $14.3 billion. The second was why agentic AI will drive more uses cases for the simulation data solutions such as what Innodata and its competitors offer.
Scale AI’s investment is a potential/speculative tailwind:
We covered why Scale AI’s investment could be a potential/speculative tailwind in our prior analysis stating: " Following Meta’s investment, it was rumored that Google, OpenAI and Tesla are looking elsewhere to avoid strengthening Meta at the cost of their proprietary data. Although it’s speculative, the exodus of major players from Scale AI could become a tailwind for Innodata. “
An analyst asked something similar on the Q2 earnings call. The CEO remained vague yet did state that something could materialize in the next couple of months:
“George Frederick Sutton, Craig-Hallum Capital Group:
Nice results. Congratulations. So I wondered if we could talk about during the quarter, your largest competitor, Scale AI was a large majority purchased by Meta. And we've had a few of the large tech companies come out and say they would no longer work with Scale AI. These ostensibely would be tech companies that you have statements of work with. So I'm just curious if you can kind of give us the after effect of that acquisition as you've seen it.
Jack S. Abuhoff, President, CEO & Director
[…] We have, in light of this stepped up that effort with certain companies and there are certain conversations that are going on and are now planned to be happening over the next couple of months that I think could be very exciting for us. I don't know that I can get into particulars much beyond that, but I'll reiterate that we do see an opportunity to accelerate our market presence.”
Agentic AI to drive forward major use cases:
Although agentic AI may still be a few years out before it’s commercially viable, the CEO pointed out the future of large language model (LLM) improvements lies in the quality of data. In order to have agentic AI that can tackle multivariant problems, training data will need to be supplemented with simulation training data.
The CEO stated the following on the call:
“We believe agent-based AI is going to serve as the cornerstone technology that unlocks the full value of large language models and generative AI for enterprises. Moreover, we believe that progress on Agentic AI is likely to soon result in a ChatGPT moment for robotics. Within the next several years, we believe Agentic AI will be served at the edge in hardware devices with which we will commonly interact in many respects in our lives. We believe the market for simulation data services and evaluation services to drive Agentic AI and robotics is likely to dwarf the market for frontier model post-training data.”
Conclusion:
We pointed out in our original analysis that Innodata has many competitors, and we provided an overview of how Innodata must find a path to compete with AI-native startups for data-as-a-service for AI training data. As proven by Scale AI, the market for data labeling and supervised learning is only going to grow in importance – but will Innodata be able to compete? That is a question this earnings report did not answer for investors. For an opportunity like this, we rely heavily on technicals.
Discovery Members: You are invited to join Knox Ridley on Monday, August 18th at 4:30 pm EST for a special Discovery webinar where he will discuss technical setups on Discovery stocks.a special Discovery webinar where he will discuss technical setups on Discovery stocks.
Damien Robbins, Equity Analyst at I/O Fund contributed to this analysis.
Please note: The I/O Fund conducts research and draws conclusions for the Fund’s positions. We then share that information with our readers. This is not a guarantee of a stock’s performance. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis.
Recommended Reading: