Google launches AI initiatives for developers, startups in India

Bengaluru (PTI): Tech giant Google on Tuesday announced the launch of its new initiatives, including AI Research Foundations curriculum and ATL Saathi, in a bid to help India's developers, startups and enterprises embrace frontier and agentic AI to drive the next wave of innovation.

The company also highlighted that, based on third-party evaluations, the Google Play and Android ecosystem generated Rs 5.3 lakh crore in revenue for app publishers and the wider economy in India in 2025, growing by 28 per cent from 2024.

Announcing the launch of its new initiatives at the Google I/O Connect India 2026 here, Manish Gupta, Senior Director for India and APAC at Google DeepMind said, the ultimate metric of AI progress isn't just model parameters, but also in the positive transformation it enables. India is championing this as it adopts AI across every tier of the economy - from local merchants to national health initiatives.

"In bringing our frontier AI, on-premise capabilities, and commitment to safety, we aim to accelerate this momentum, and look forward to the country's AI learners, educators, builders, and innovators leading India's AI ambition from the front," he said.

In a statement, Google said that Google DeepMind is bringing its rigorous AI Research Foundations curriculum - a free 56-hour programme that equips learners to build and fine-tune Large Language Models, and drive high-impact AI research.

"Learners can earn industry-recognised Google Cloud Skill badges and certificates upon completion," it said.

Along with being available on the Google Skills platform, Google has partnered with NASSCOM and IISc Bangalore to scale the programme with institutions across India, while AVPN, a network of social investors, will work with local partners to bring it to learners under the Google.org APAC AI Opportunity Fund.

With the launch of ATL Saathi, the company said it would deepen its collaboration with the Atal Innovation Mission.

"ATL Saathi is a desktop web application that gives teachers a Gemini-powered assistant to help deliver the Atal Tinkering Labs' curriculum and curate hands-on experiments for students. ATL Saathi is being rolled out to 100 schools this year, with an aim to eventually reach 10,000 schools served by the Atal Tinkering Labs programme," it said.

To proactively help mitigate potential threats that might emerge with the misuse of agentic AI, Google said it is bringing Sec-Gemini v3, its specialised cybersecurity agent, to trusted government and enterprise testers, including Flipkart.

"Sec-Gemini can reason across complex security data and help teams investigate incidents at machine speed," it said.

Highlighting how Google AI is helping strengthen India's healthcare ecosystem, the company said that researchers at AIIMS Delhi are now leveraging Google's multimodal MedGemma open models to develop new India-specific models for leprosy and sexual and reproductive health.

"These models have the potential to help patients and healthcare professionals identify and manage conditions based on images and text inputs. AIIMS will make the outcomes of al these localised clinical health models available to the Indian developer ecosystem," it said.

It also noted that Gemini Live now supports 25 Indian languages and dialects, including Sanskrit, Bhojpuri, and Maithili, stating that this is the company's latest measure in ensuring AI addresses India's linguistic and cultural diversity.

Google said it had recently expanded its commitment to India's digital sovereignty by making Gemini 3.5 Flash accessible to Indian enterprises and startups via the Gemini Enterprise Agent Platform and the Gemini Enterprise app.

"Starting now, this model will be available with strict in-country machine learning processing commitments," it said.

The company said it has also been enabling Indian enterprises, including regulated industries and the public sector, to run Gemini on Google Distributed Cloud from within Indian data centres.

"Enterprises and public sector organisations can now build with frontier AI entirely within their own perimeter, with supporting services also fully disconnected from the public internet," it said.

The company further said that Google is open-sourcing a secure runtime environment known as CAPSEM (Capabilities Security for Agents) to help startups build securely with agentic AI.

"CAPSEM places each AI agent inside an isolated virtual machine. If an agent is compromised or encounters a malicious prompt, the wider system remains fully protected," it said.

To ensure that agentic AI continues to support interoperability as it is built on the foundations of trust and security, Google said it is driving open industry standards.

"The Device Bound Session Credentials (DBSC) standard makes stolen session cookies instantly useless to bad actors, while the Agents-to-Payments, operating alongside the open Agent2Agent (A2A) protocol, helps make authorised, low-value agent-led financial transactions (under $100) highly secure and accountable," it said.

Google said it has also established new collaborations with IIT Delhi and IIT Madras to drive research in agentic safety, including early threat detection and the development of next-generation Guardian Agents that help mitigate institutional risks, it said.