The Company
Innodata, Inc. (NASDAQ: INOD) is a data-engineering and AI services company that sources, curates, annotates and delivers high-quality training data and data-engineering solutions for AI and machine-learning applications. Its offerings span large-scale data collection (speech, image, video, text, sensor, code and documents), annotation/labeling, synthetic data generation, model evaluation and applied-AI automation/low-code deployment for enterprise and large tech customers.
Financials (latest reported / headline metrics)
- Q2 2025 revenue: $58.4 million (organic growth +79% YoY).
- Q2 2025 adjusted EBITDA: $13.2 million (≈23% of revenue); GAAP net income: $7.2 million; cash & short-term investments: ~$59.8 million as of June 30, 2025.
- Management raised full-year 2025 organic revenue guidance to ~45%+ growth following the quarter.
These results reflect a sharp acceleration in revenue and margin expansion driven by AI/data-services demand. Innodata+1
Bookings
| Customer | Revenue / Value | Contract Details (duration, start/end) |
|---|---|---|
| Leading cloud infrastructure & platform services company | ~US$1 million | Win announced Jan 17 2023 for large-scale data collection for AI computer vision initiative. |
| Leading multinational bank | ~US$11 million total over 5 years (≈US$2 m/year initial 2 yrs, then ~US$2.4 m/yr) | Five-year contract announced Sept 23 2021. Duration: ~5 yrs. |
| New “Big Five” technology company | Up to US$8 million spend for remainder of 2023; potential annual run-rate US$15 million or more by end of 2023. | Agreement announced July 18 2023. Specific start/end not detailed. |
| Existing “Magnificent Seven” Big Tech customer | Additional ~US$20 million annualized run-rate revenue from new programs. | Announcement April 24 2024. Amendment expected. Duration/term not clearly detailed. |
| Same “Magnificent Seven” Big Tech customer | Additional ~US$44 million annualized run-rate from two new LLM programs. | Announcement June 3 2024. Contract “contains early-termination upon notice” — term not fixed. |
| Same largest Big Tech customer (over time) | Total account value approx US$110.5 million annualized run-rate as of June 3 2024. | Ramp-up over ~1 year; term not explicitly fixed. |
If you like, I can search deeper and
TAM / CAGR (market the company operates in)
Innodata primarily competes in the AI training-data / data-labeling / data-engineering market. Recent market estimates (different analysts) put the addressable market in the low single-digit billions today with rapid expansion toward tens of billions within the decade: for example, one estimate values the data-labeling market at ~$6.5B in 2025 with ~25% CAGR to 2030, while broader “data labeling / services” or “AI training dataset” market estimates range from ~$2.6B–$18.6B (recent baselines) and project 20–25%+ CAGRs into the late 2020s depending on scope (pure annotation services vs. full training-dataset solutions). Taken together, a conservative working TAM for Innodata’s target addressable market (data collection + labeling + AI training datasets + applied data engineering) is multiple billions today with high-teens to mid-20s percent CAGR over the next 5 years. Mordor Intelligence+2Grand View Research+2
Products / Services (with approximate revenue split)
Innodata reports revenue across core business segments; using the company’s segment reporting (latest 12-month/LTM figures), the split is heavily concentrated in its Digital Data Solutions business. The table below shows the common segment names and an LTM revenue percentage approximation.
| Segment / Product area | What it is (brief) | Approx. % of revenue (LTM) |
|---|---|---|
| Digital Data Solutions (DDS) | Large-scale data collection, annotation, synthetic data, model evaluation and managed AI training-data services for LLMs, vision, speech, etc. (enterprise & big-tech customers). | ~86–87%. StockAnalysis+1 |
| Agility | Low-code / automation / deployment platforms and AI-enabled content/data workflow solutions (platform & automation services). | ~10%. StockAnalysis |
| Synodex / Legacy Media & Analytics | Media-monitoring, information products, legacy content services and related analytics (smaller, more stable business). | ~3–4%. StockAnalysis |
Notes: percentages are derived from the company’s segment revenue disclosure/LTM figures and rounded for presentation; Innodata has emphasized DDS (AI training data and related services) as the primary growth engine. StockAnalysis+1
Business Model
- Managed services + platform hybrid: Innodata combines high-volume outsourced managed services (human and automated labeling, quality control, data collection) with proprietary and third-party tooling (low-code platforms, model evaluation tooling) to deliver enterprise-grade training datasets and data pipelines.
- Contract structure: Revenue comes from long-term contracts, program expansions and per-project engagements (fixed-price and time & materials), often with multi-year engagements and renewable scopes as AI models require continuous dataset refresh.
- Unit economics & scale: The model leverages global delivery footprint, automation (model-assisted labeling/synthetic generation) and economies of scale to improve gross margins as volumes increase; recent quarters show strong operating leverage. Innodata+1
Customers
Innodata serves large technology and enterprise customers building generative AI and other ML systems, including major cloud and AI companies, software vendors and selected enterprise verticals that require high-quality labeled datasets and data-engineering. Management has highlighted multiple “big tech” customer wins and renewals that materially contributed to recent growth. Customer relationships tend to be a mix of a few large strategic accounts plus a diversified set of enterprise customers. Innodata+1
Competitors (top 3 with directly competing products)
- Appen (and related AI-data providers) — large global provider of data collection, annotation, speech and text datasets that directly competes with Innodata’s managed labeling and data-collection services. MarketsandMarkets+1
- Scale AI — platform + managed labeling at scale (widely used by big tech and enterprise customers); competes on high-volume annotation, tooling and ML-assisted labeling. (Note: Scale’s ownership/customers have shifted market dynamics in recent years, increasing competitive flux.) Business Insider+1
- Lionbridge / TELUS International (AI Data Solutions) — established provider of multilingual annotation, data collection and content services; competes on language, compliance and scale for global datasets. Twine+1
(Other relevant competitors and alternatives include iMerit, CloudFactory, Labelbox and a growing number of specialist startups; competitive intensity is high and evolving.) Encord+1
Founding History
Innodata was founded in 1988 and built its initial business around digitization, content conversion and information-product services. Over the years it expanded into global delivery and content/data enrichment services; more recently (last several years) the company strategically repositioned and invested to serve the emerging market for AI training datasets, data engineering, and applied AI solutions—shifting its center of gravity toward the Digital Data Solutions segment that now drives the majority of revenue. Management and the company narrative emphasize a multi-decade legacy combined with recent product and go-to-market evolution for the generative AI era.
Invest Alpha 