CuriosityStream Inc. (CURI) Earnings

Management highlights

### Core Financial Performance - Q2 2026 delivered record results across all key metrics: total revenue of $23.2 million (up 22% YoY), net income of $8.9 million (up 1133% YoY), EPS of 15 cents, adjusted EBITDA of $11.4 million (up ~300% YoY), and 6 consecutive quarters of positive adjusted EBITDA - Gross margin expanded to 73% from 53% YoY, and adjusted EBITDA margin reached 49% from 16% YoY, driven by operating leverage from higher-margin licensing growth and aggressive cost reduction - Total operating expenses declined 24% YoY via AI productivity tool adoption and talent alignment to high-priority initiatives, with further expense reductions expected in H2 2026 - The company ended the quarter with $10.9 million in cash and securities and no outstanding debt; it paid a $5 million regular dividend, repurchased $600,000 in shares, and prepaid $2 million to fully acquire its German JV, closing July 1 2026 ### Strategic and Operational Updates - The company maintains a diversified monetization model built on three distinct licensing pillars: traditional premium video licensing to global distribution partners, structured video/audio datasets for AI training, and an 880 billion token private code corpus for LLM training, reinforcement learning, and fine-tuning - The company has productized 17 off-the-shelf structured video dataset products tailored to specific AI developer requirements, including collections for sports, entertainment, science, HDR, character tracking, and raw footage, to reduce sales friction and shorten sales cycles - The company remains focused on disciplined subscription business growth, prioritizing durable economics over unprofitable customer acquisition at any cost, and continues to build long-term value for the Curiosity brand - Long-term strategic goal is to grow the company to $100 million+ of reliable, recurring annual revenue

Guidance

- Full year 2026 total revenue is guided to a range of $77 million to $82 million, with H2 2026 revenue guided to $38 million to $41 million - Full year 2026 adjusted EBITDA is guided to a range of $18 million to $22 million, with H2 2026 adjusted EBITDA guided to $6 million to $10 million; the lower H2 guidance reflects intentional conservatism around licensing deal closing timelines, not a change in underlying cost or margin expectations - End of year 2026 cash and investments balance is expected to land in the range of $17 million to $22 million - Management expects an additional 18% to 20% reduction in operating expenses in H2 2026 compared to H1 2026

Segment performance

CuriosityStream operates two core revenue segments: 1) Licensing: Q2 2026 revenue of $14.1 million, representing a 48% year-over-year increase vs. Q2 2025. This segment accounts for 60.8% of total Q2 2026 revenue, and was the primary driver of the quarter's overall revenue growth. 2) Subscription: Q2 2026 revenue of $8.9 million, which is roughly flat year-over-year vs. Q2 2025 and saw a slight sequential improvement from Q1 2026. This segment accounts for 38.4% of total Q2 2026 revenue.

Risks & headwinds

- Forward-looking statements are inherently uncertain, and actual results may differ materially from expectations due to a variety of unstated material risks; management refers participants to the company's SEC filings for a full discussion of risk factors - Licensing revenue can be lumpy/chunky quarter-to-quarter based on large deal closing timelines, introducing near-term revenue volatility - The emerging AI training data market is competitive, and sustained success depends on maintaining sufficient sales and operational velocity to keep up with market changes

Analyst Q&A

• Q: Can you provide details on your licensing pipeline and its potential to drive future AI and streaming growth? /

  A: The licensing pipeline is the most robust it has ever been, supported by three distinct, diversified licensing pillars that reduce risk and allow the company to serve customers across multiple large, growing markets with varying use cases and buying cycles. Productization of off-the-shelf datasets is expected to accelerate sales cycles and reduce operational friction, which should support more predictable pipeline conversion. While licensing revenue can be lumpy quarter-to-quarter, the broad scope of the company's IP offering minimizes the risk of large dips and supports material upside potential.

• Q: Can you share your plans for international expansion following the full acquisition of your German JV? /

  A: The acquisition of full ownership of the German JV gives the company full control of two 24/7 pay TV channels and FAST distribution across German-speaking Europe, which is already CuriosityStream's largest non-English market. The company's evergreen content travels well internationally, and a large share of current subscribers already come from outside the U.S. Ongoing rollouts of new localized currencies and payment systems will allow the company to capture more international subscription growth moving forward.

• Q: Why is your H2 2026 adjusted EBITDA guidance lower than Q2 2026's actual result? Is it due to higher costs or conservative forecasting? /

  A: The lower guidance is entirely a product of intentional conservatism around forecasting when large licensing transactions will close, not due to expected increases in costs or lower margins. The first half run rate would put full year adjusted EBITDA around $25 million, above the current guided range, and management intends to meet or exceed the published guidance, with updates to guidance possible if performance warrants changes in the second half.

• Q: What has the market reception been for the new 17 off-the-shelf AI dataset products, and how have they impacted the H2 pipeline? /

  A: Productization organizes the company's existing 3+ million hour library into clearly defined, heavily annotated datasets tailored to specific AI developer needs, rather than introducing entirely new content. This simplification of the purchasing process has already started to accelerate sales and operational cycles, and works across the full customer base: it is particularly helpful for smaller AI companies that want smaller initial orders, while also benefiting large frontier model developers. The breadth of structured datasets available positions CuriosityStream uniquely relative to competitors in the AI training data market.