Runway and Holy Water: Two AI Video Companies Trying to Upend Entertainment | StrictlyVC Athens 2026

Runway's Early Generative AI Bet

Anastasis explains that Runway started in 2018 when generative AI was considered a "toy," anticipating that models would evolve from low-resolution images to generating entire films and fundamentally transforming creative workflows.

Holy Water's Pivot Culture

Anatoli describes launching utility apps in Ukraine before executing three strategic pivots toward content, starting with books and evolving into AI movies while maintaining 2x annual growth by adapting quickly to technological shifts.

Intentionally Lean Team Philosophy

Runway keeps teams smaller than conventional wisdom suggests to force high conviction in ideas, using rapid prototyping and demo culture rather than lengthy specifications to validate concepts before full investment.

AI-Native Operating Model

Holy Water structures teams as internal startups with equity stakes, allowing 2-3 person groups to launch products in weeks using AI agents, effectively aiming to make 200 employees perform like 2,000.

Dual-Use World Models Strategy

Runway shares 90% of compute infrastructure between creative and robotics applications but maintains completely different go-to-market motions, embedding artists with Hollywood studios while post-training models for specific robotics hardware.

Content Distribution Platform Vision

Holy Water leverages Runway's models to produce AI movies and short dramas, serving over 100 million users with ambitions to become a Roblox-like ecosystem where anyone can create and monetize Hollywood-quality content.

Structural Reorganization for Physical AI

Runway faces the challenge of evolving from a pure video generation company to a world modeling company serving robotics, requiring new research priorities and commercial structures for industrial applications.

Workforce AI Transformation

Holy Water struggles to transition 200+ employees from simply using AI tools to becoming "operators of AI agents" who build digital twins and automated knowledge systems rather than focusing on manual input hours.