China's Open-Source AI Strategy and Implications for India

#GS-3 #Science & Technology #Artificial Intelligence #Current Events #International

Key takeaways

At the 18th BRICS Summit, China proposed establishing a shared BRICS open-source AI community to bypass US chip sanctions and expand software influence.
India's data center capacity is projected to expand from 1.6 GW in 2026 to 6 GW by 2029, requiring USD 110 billion in infrastructure investment.
India remains 90% to 95% dependent on imported semiconductors and relies entirely on foreign foundries like TSMC for sub-5nm AI chips.
To protect its digital autonomy, India is implementing the ₹10,300 crore IndiaAI Mission to build sovereign models and indigenous GPU computing power.

During the 18th BRICS Summit held in New Delhi in September 2026, China proposed creating a shared BRICS open-source AI community.

China offered to lead the BRICS AI initiative by providing shared large language models (LLMs) and a dedicated cloud platform.
Strict US semiconductor bans restrict China from dominating chip hardware, so Beijing focuses on dominating the software layer instead.
Chinese tech firms like Alibaba are releasing advanced foundational models like Qwen for free to attract global developers.
By sharing open-weight models like DeepSeek, China mimics the Android strategy to establish its software as default global infrastructure.

Joining a China-led AI network risks shifting India's reliance from US proprietary models to Chinese systems with hidden data practices.
India promotes democratic digital values, whereas China mandates that AI models align strictly with state objectives.
Free foreign open-weight models could harm the commercial viability of Indian AI startups like Sarvam and BharatGen.

Expanding India's data center capacity from 1.6 GW in 2026 to 6 GW by 2029 requires USD 110 billion in funding.
Indian data centers consume 150 billion liters of water and 0.5% of national power, which could double by 2030.
Mumbai holds 50% of India's data centers, causing high latency and delays in smaller Tier-2 and Tier-3 cities.
India imports 90% to 95% of its semiconductors, making its AI ambitions vulnerable to conflicts in the Taiwan Strait.
The India Semiconductor Mission (ISM) attracted USD 20 billion, but India still relies on TSMC for sub-5nm AI chips.
The Digital Personal Data Protection (DPDP) Act, 2023 creates regulatory unpredictability for cross-border data flows and foreign cloud investments.
The Copyright Act, 1957 lacks clear rules regarding fair use of copyrighted data for training generative AI models.

India should fast-track the ₹10,300 crore IndiaAI Mission to build sovereign models on diverse datasets through initiatives like Bhashini.
The government must expand public-private partnerships to build domestic GPU clusters and reduce reliance on foreign cloud servers.
India should extend its Digital Public Infrastructure (DPI) model, similar to UPI, to offer democratic open-source AI tools globally.
India can secure chip supply chains by strengthening partnerships like iCET with the US and the Quad semiconductor initiative.
India should use global platforms like G20 and GPAI to set ethical governance rules for risks like agentic misalignment.
Updating NEP 2020 and upgrading Atal Tinkering Labs (ATLs) will help build an AI-ready workforce.

India must balance rapid AI adoption with technological self-reliance by building sovereign compute capacity and transparent digital infrastructure.