The India AI Impact Summit 2026 established a global “AI for All” framework while securing $200 billion in infrastructure goals amid notable logistical challenges.
Sarvam opened Indus in limited beta, powered by its 105B sovereign model, and outlined new products across document intelligence, speech, on-device AI, dubbing, and agent infrastructure.
OpenAI and Paradigm have released EVMbench, revealing that AI agents can now successfully exploit more than 70% of tested smart contract vulnerabilities — as real-world DeFi hacks make the stakes impossible to ignore.
Gemini 3.1 Pro enters preview in February 2026 with adjustable thinking modes, a verified 77.1% ARC-AGI-2 score, and expanded rollout across Gemini and developer tools.
OpenAI launches OpenAI for India and partners with Tata Group to deploy 100MW of AI-ready data center capacity, with plans to scale to 1GW and expand enterprise and youth AI programs.
Humain invests $3 billion in xAI Series E before SpaceX acquisition, converting the stake into SpaceX shares in the $1.25T merged company.
World Labs raised $1 billion, including $200 million from Autodesk, to scale AI world models that generate and reason about 3D environments.
Google released its 2026 Responsible AI Progress Report, outlining lifecycle governance, Gemini 3 safety testing, agentic AI safeguards, provenance tools, and security frameworks as EU AI Act enforcement approaches.
OpenAI hires OpenClaw creator Peter Steinberger. OpenClaw will move to a foundation-backed open-source structure as companies debate security risks.
