# India’s $30 Billion AI Push: DeepSeek Hosting, AI Compute, and Techno-Legal Governance

India has officially entered the global AI race with an ambitious **$30 billion investment**, announced by **Ashwini Vaishnaw, Minister of Electronics and Information Technology**. The announcement highlights India's commitment to AI infrastructure, foundational models, and applications, with a **four-year subsidy program** aimed at boosting growth. This move positions India as a key player in AI while ensuring data privacy, security, and respect for intellectual property (IP) rights.

### AI Compute & Infrastructure

A major highlight of this initiative is the establishment of **Common Compute**, with a total of **18,693 GPUs**, making India a competitive force in AI computing resources. This initiative will significantly lower the cost of AI compute, bringing it down to **₹115.85 per GPU hour**, compared to the global standard of **$2.5-$3 per GPU hour**. For high-end computing, the cost is projected to be around **₹150 per GPU hour** for running the best LLM.

To further enhance India’s AI capabilities, the government will **host DeepSeek on Indian servers**, similar to how LLaMA was previously run. This ensures data privacy and security while making India a self-reliant hub for AI research and development.

### Foundational Models & AI Applications

The government has issued a **call for proposals to create foundational AI models**, with six major players expected to take the lead in this initiative. This move is expected to reduce dependency on foreign technology and drive local innovation.

Additionally, **18 AI applications** have been developed across key domains, including **climate change, agriculture, and learning disabilities**. These applications showcase India’s focus on leveraging AI for social and economic impact.

### AI Safety & Regulatory Measures

To ensure the safe and ethical development of AI, the government is launching a **dedicated AI Safety Institute** and supporting **eight research projects** focused on AI risk mitigation. This aligns with India’s **techno-legal approach**, balancing technological advancement with robust legal frameworks.

By **respecting IP rights** and fostering a **regulated yet innovation-friendly environment**, India is establishing itself as a **trusted AI destination**. This approach aligns with global efforts to address AI governance and safety concerns.

### Support for Startups, Researchers, and Students

A **40% subsidy** will be provided to **startups, researchers, and students**, further reducing the cost barrier and encouraging innovation in AI development. This subsidy will play a crucial role in making advanced AI tools accessible to a wider talent pool.

### A Government That Runs Like a Tech Startup

One of the most intriguing aspects of this initiative is the vision to **run the government like a tech startup**. This means agile decision-making, rapid execution, and a strong focus on innovation. With AI deeply integrated into governance, India aims to enhance efficiency and transparency across various sectors.

### Final Thoughts

India’s AI strategy is a game-changer, providing a **structured and subsidized** environment for AI development while addressing **privacy, security, and legal concerns**. By building its own foundational models, investing in AI compute, and fostering innovation through a techno-legal framework, India is setting a strong precedent for responsible AI growth.

This initiative marks a significant step in India’s journey to becoming a **global AI powerhouse**, leveraging its **trusted status in technology and IP protection**. The next few years will be critical in seeing how these policies translate into real-world impact, but the groundwork has been firmly laid.

### Sources

* **Official Announcement:** [YouTube Video](https://youtu.be/l5hnclcIaRc)
    
* DeepSeek Source Paper : [https://github.com/deepseek-ai/DeepSeek-R1](https://github.com/deepseek-ai/DeepSeek-R1)
    
* DeepSeek Coder : [https://github.com/deepseek-ai/DeepSeek-Coder](https://github.com/deepseek-ai/DeepSeek-Coder)
    
* gpt4free open source: [https://github.com/xtekky/gpt4free](https://github.com/xtekky/gpt4free)