Avataar AI Launches Varya at ₹0.48 a Second, 20x Cheaper Than Rivals
Updated
Updated · TechCrunch · Jun 12
Avataar AI Launches Varya at ₹0.48 a Second, 20x Cheaper Than Rivals
1 articles · Updated · TechCrunch · Jun 12
Summary
Avataar AI released Varya as an open-weight video model on India’s AI Kosh portal, pitching it for India with training tuned to local festivals, food, clothing and architecture.
Built by distilling Alibaba’s Wan 2.2 rather than training from scratch, Varya runs in four steps instead of 50, generating a 5-second 720p clip in 45 seconds versus 1,230 seconds on an NVIDIA H200 GPU.
Pricing is the main hook: Avataar plans to charge ₹0.48 ($0.005) per second on its hosted service, versus roughly $0.10 or more per second for models such as Veo, Kling, Luma and Runway.
The launch fits India’s AI strategy of emphasizing applications and developer ecosystems over frontier foundation models, as limited compute and high-quality data have slowed domestic model development.
Avataar is one of 12 startups backed by the government’s roughly $1.2 billion India AI Mission, which offers subsidized compute in exchange for public model releases as India targets $200 billion in AI investment by 2028.
Is India’s low-cost, localized AI a clever market strategy or a compromise that will lag behind frontier model development?
With its national AI Mission underfunded, can India's subsidized startup strategy truly challenge global tech giants in the long run?
India’s Varya AI Video Model: Open-Weight, Ultra-Low Cost, and Built for 1.4 Billion People
Overview
The launch of Varya by Avataar AI on June 11, 2026, marks a pivotal moment for Indian AI video generation and signals a strategic shift in the nation’s approach to artificial intelligence. Instead of focusing solely on foundational model development, India is emphasizing practical AI applications, reflecting a fundamental strategic decision within its broader AI ambitions. Industry experts believe India’s strength lies in building a robust ecosystem of applications and developers. This pragmatic approach is crucial for India to carve out its unique niche and establish a strong presence in the rapidly evolving global AI landscape.