Innodata (INOD) Q2 2026 earnings review
Massive Operating Leverage Underpins Record Quarter, But Unexplained Prepayments Raise Eyebrows
Innodata delivered a blowout Q2 with revenue accelerating to $92.1M (+58% YoY) and Adjusted EBITDA surging 92% YoY to $25.4M. The company is successfully transitioning from a lower-margin data service provider to a high-margin intellectual property partner for Frontier AI labs, driving Adjusted Gross Margins to 49% (900 bps above target). However, beneath the pristine top-line metrics lies a highly unusual cash dynamic: $116M in unexplained 'pass-through' prepayments flooded the balance sheet. Furthermore, the company announced a planned leadership transition with founder Jack Abuhoff shifting to Executive Chairman, handing the CEO reins to President Rahul Singhal in September.
🐂 Bull Case
Gross margins hit 49%, fundamentally changing Innodata's margin profile. By retaining IP on off-the-shelf datasets and monetizing the same asset across multiple Big Tech customers, the company is breaking the linear link between headcount and revenue.
The newest Big Tech customer scaled from 17% of revenue in Q1 to 34% in Q2. Capturing massive share of wallet from the world's most demanding AI labs validates the company's technical capabilities.
🐻 Bear Case
Management claims 'diversification has been delivered.' The data disagrees. In Q1, the top 2 customers were 73% of revenue. In Q2, they are 71%. The company traded one giant whale for two, remaining dangerously top-heavy.
Reported cash tripled sequentially to $250.4M, but $116M of this is from customer prepayments for 'pass-through costs.' This is a massive, unexplained figure exceeding an entire quarter's revenue that obfuscates true cash generation.
⚖️ Verdict: 🟢
Bullish, but proceed with caution. The fundamental transition into high-margin, scalable AI data IP is undeniable and accelerating. However, the extreme concentration risk and the opacity surrounding the pass-through costs require strict monitoring.
Key Themes
Margin Profile Reversing Historic Norms via IP Ownership
Innodata is accelerating its transition from a pure service provider to an IP owner. Adjusted Gross Margin reached 49%, a staggering 9 points above management's stated target. This is being driven explicitly by a mix shift toward 'off-the-shelf datasets' where Innodata retains the IP and licenses it to multiple AI labs, allowing revenue to grow without linear cost increases.
Frontier Lab Demand Fueling Top-Line Hypergrowth
The macro tailwind of generative AI capital expenditure continues unabated. Innodata is successfully capitalizing on the race between Frontier Labs (the 'Mag 7'). The company landed a new frontier lab in Q2, while its newest Big Tech customer doubled its share of total revenue sequentially. Demand for complex, reasoning-based data is accelerating.
Product Innovation: Agentic Reinforcement Learning
Management specifically called out an early position in 'agentic reinforcement learning' as a primary catalyst for new program wins. The release of the AI Cyber Training Suite (12 datasets designed to train AI coding agents to repair vulnerabilities) proves Innodata is moving aggressively into complex, post-LLM agentic data architectures.
Concentration Narrative Contradicts the Math
CEO Jack Abuhoff stated, 'The diversification we planned for has now been delivered.' This is statistically misleading. In Q1, the largest customer was 56% of revenue, and the second was 17% (total 73%). In Q2, the largest is 37%, and the second is 34% (total 71%). While the weight between the two shifted, Innodata is still acutely vulnerable to the capital expenditure cycles of just two companies.
Massive 'Pass-Through' Prepayments Obscure Cash Dynamics
The balance sheet shows cash skyrocketing to $250.4M. However, management disclosed that roughly $116M of this is 'customer prepayments related to pass-through costs,' reducing core cash to $134M. A $116M pass-through mechanism—larger than the entire quarterly revenue of the company—is highly unusual. If these pass-throughs flow through the income statement in future quarters, they could heavily distort revenue and crush gross margins.
Founder Transitions to Executive Chairman
Effective September 30, Jack Abuhoff will step down as CEO to become Executive Chairman, while President Rahul Singhal assumes the CEO role. This is a classic maturity transition, moving from a founder-led build phase to a professionally managed scale phase. Abuhoff will focus on Federal and Enterprise recurring revenue, while Singhal manages the core Big Tech engine.
Other KPIs
Accelerating. Up 92% YoY compared to 58% revenue growth, showcasing severe operating leverage. The business model is proving it can scale without a proportionate increase in headcount or overhead, stabilizing EBITDA margins in the high 20% range.
Stable sequential growth. Up from $117.4M in Q1 2026. The true cash-generating power of the business remains strong, even after stripping out the $116M optical illusion created by the pass-through prepayments.
Guidance
Stable (Reiterated). However, this guidance appears aggressively sandbagged. H1 2026 revenue sits at $182.2M (+56% YoY). To merely hit 40% growth for the full year, H2 revenue would only need to be ~$170M, implying an average of $85M per quarter. Given Q2 just hit $92.1M, the 40% guide implies a severe sequential deceleration in H2, which directly contradicts management's bullish commentary on new project scoping.
Key Questions
The Pass-Through Cost Mystery
You received ~$116 million in customer prepayments for pass-through costs. What exactly are these costs (e.g., compute, third-party hardware)? More importantly, when this cash is deployed, will those costs flow through the P&L as revenue, and if so, how will that impact gross margins?
Sandbagged Guidance?
You reiterated 40% full-year growth despite delivering 56% growth in H1. Given your Q2 run rate of $92M, 40% implies a sequential decline in H2. Are there specific project roll-offs you are modeling, or is this just extreme conservatism?
Concentration vs. Diversification
While you noted diversification away from your largest customer, your top two customers now account for 71% of revenue. Are you comfortable with this level of concentration, and what is the realistic timeline for the rest of the business to dilute this top-heavy dynamic?
M&A Strategy Under New CFO
With Jayant Chauhan joining as CFO with a mandate that includes M&A and capital allocation, and with $134M in core cash on the balance sheet, should investors expect a shift toward inorganic growth to supplement the organic engine?