
Walking the tightrope: Building an open-source B2B infrastructure company
The Open-Source Advantage: Zipline AI’s Strategy for B2B Infrastructure Leadership
(This article was generated with AI and it’s based on a AI-generated transcription of a real talk on stage. While we strive for accuracy, we encourage readers to verify important information.)
Mr. Varant Zanoyan, CEO of Zipline AI, shared insights on building a B2B infrastructure company around an open-source project. Leveraging his 12 years in data infrastructure for AI/ML at Palantir and Airbnb, Mr. Zanoyan co-founded Zipline AI. The company is founded on Chronon, an open-source data platform he helped develop at Airbnb to power its critical machine learning initiatives.
Chronon’s core function is to transform organizational data into features for AI/ML applications. It was first deployed at Airbnb to combat payments fraud, a significant challenge for the company’s high-volume international transactions. Its success led to broader adoption for trust-based ML, search personalization, recommendations, and even Generative AI for customer support, becoming a general-purpose data platform.
Motivated by Chronon’s internal success, the team open-sourced the project. Stripe became an early partner, integrating Chronon for its similar payments fraud detection needs. Today, Chronon powers 100% of Stripe’s charge path fraud prediction models, demonstrating robust capabilities in critical enterprise environments and validating the open-source approach for large-scale operations.
The open-source Chronon has gained significant traction across various industries. OpenAI utilizes it for personalizing the Sora 2 feed, recommending content based on user interactions. Uber Eats leverages Chronon for search personalization and recommendations, while Netflix employs it for content recommendations. This widespread adoption underscores Chronon’s versatility for diverse AI/ML applications.
The strong momentum of the open-source Chronon project led to the founding of Zipline AI a year ago, aiming to evangelize and support the platform. Mr. Zanoyan highlighted key challenges for B2B infrastructure as a service: solving critical, complex problems that companies prefer not to build entirely in-house, given demanding procurement processes and the need for trust.
Building trust as a startup is crucial, addressed by Zipline AI through open-source traction, case studies, and an experienced team. Data security is paramount, especially for financial clients, ensured by a full Bring Your Own Cloud (BYOC) deployment where no customer data leaves their account. For business continuity, Chronon’s open-source nature provides a self-hosting path.
This self-hosting option creates the “tightrope” dilemma: balancing proprietary value with open-source accessibility. Zipline AI’s solution maintains Chronon as an Apache 2 licensed open-source core with a healthy community. While self-hosting requires a team of about five people and a clunky user experience, Zipline AI offers a proprietary UI.
Zipline AI’s UI streamlines observability, governance, and management, significantly reducing developer and maintenance overhead to roughly half a person. This balanced approach makes technology adoption comfortable while encouraging partnerships for accelerated deployment. Switching from Zipline to self-hosted open source incurs additional headcount and slower iteration, but guarantees zero downtime and continuous project improvements from the growing open-source community.

