Union Minister for Railways and Electronics and IT, Ashwini Vaishnaw, stated on Thursday that India is likely to introduce its own indigenous AI model at a reasonable price in the near future, adding that it is already safe and secure.
This is a timely step, as India is a trusted nation, and therefore, it will help the country emerge as a more reliable technological powerhouse for ethical AI solutions in the days to come, the minister told the media.
The IndiaAI mission, now bolstered by a high-end common computing facility, is closer to customizing indigenous AI solutions for the domestic context using Indian languages.
Scientists, researchers, developers, and coders are working on multiple foundational models in this regard, and with the current pace, the Union Minister expressed hope that the Indian AI model will likely be ready within six months.
The AI model is starting with a computational facility of roughly 10,000 GPUs. Soon, the remaining 8,693 GPUs will be added. It will largely benefit researchers, students, and developers initially.
They have decided to offer it to users for less than Rs 100 per GPU, after subsidizing 40 percent of the cost.
Compared to global models that cost $2.5 to $3 per hour of usage, India's AI model will cost less than Rs 100 per hour after the 40 percent government subsidy.
It is also informed that the half-yearly and annual plans, which are quite attractive, would further make it affordable.
This is about nine times what the open-source model China's DeepSeek has and about two-thirds of what ChatGPT offers.
The safety and ethical deployment of the AI model remain the government's top priority. This commitment was underscored as the Union Minister announced that India is establishing an AI Safety Institute, adopting a techno-legal approach.
The Indian AI model will promote innovation and develop citizen-centric governance tools, including many industrial uses that have immense potential to harness technological benefits for the people at large.
Over the past year and a half, India has been developing a robust AI ecosystem framework to support its own foundational AI model.
Indian Servers to Host China’s DeepSeek AI to Address Privacy Concerns: Ashwini Vaishnaw
Indian servers will host the new Chinese artificial intelligence platform, DeepSeek, while also addressing privacy concerns surrounding it, Union Minister of Railways and Electronics and IT, Ashwini Vaishnaw, said on Thursday. The arrival of DeepSeek has raised concerns about user safety as the AI models were created by a Chinese AI company.
Union Minister of State for Electronics and IT, Rajeev Chandrasekhar, addressing a press briefing here, said that hosting DeepSeek on Indian servers would address issues related to cross-border data transfers.
"DeepSeek is open source. We will host it on Indian servers very soon. That itself addresses the privacy concerns of cross-border data transfer," the Minister said.
Founded by quant fund leader Liang Wenfeng, DeepSeek has triggered a sell-off in tech stocks across the world.
Vaishnaw also mentioned that the country had already surpassed its first set of targets for GPUs under the IndiaAI Mission.
It has now offered 18,693 GPUs. The initiative will provide resources to researchers, startups, and businesses across the country. Notably, 15,000 high-end GPUs have been procured, including 1,480 H200 GPUs.
At the same time, models like DeepSeek and ChatGPT were trained using 2,000 and 25,000 GPUs, respectively.
In the next two to three years, the Minister said, an estimated $30 billion will be invested in India for hyperscalers and data centers.
"Approximately 10,000 GPUs are now available for use starting today," Vaishnaw said.
The minister also disclosed that a common computing facility for researchers will start functioning within two days. This will be a very useful resource for various AI projects underway in India.
DeepSeek is being touted as an alternative to OpenAI’s ChatGPT, amid reported probes by OpenAI and Microsoft regarding whether DeepSeek copied their APIs.
Meanwhile, Euroconsumers, a coalition of consumer groups in Europe, has filed a complaint with the Italian Data Protection Authority regarding how DeepSeek handles personal data under GDPR.
The Italian DPA stated that "the data of millions of Italians is at risk" and has given DeepSeek 20 days to respond.
Read also| OpenAI Investigates Potential Use of Its Models by China's DeepSeek for AI Training
Read also| DeepSeek Disrupts AI Cost Assumptions, Sends Shockwaves Through US Tech Giants