Introduction

India is hosting the AI Impact Summit from 16th to 20th February, 2026. The government’s objective is to bridge the AI gap between the global north and the south and advocate for inclusive development of the technology. One of the key propositions which has emerged during the Summit as a solution is to promote Open Source AI.

What is Open Source AI?

Open Source Initiative (OSI) has given the first stable definition of open source AI. It says that the system could be used for any purpose without asking for any permission, allow for its technical know-how, be adopted with or without modification and could be shared with anyone openly. More broadly, models deploying open weights could also qualify as open source, provided they could be implemented at scale.

How can it bridge the global north and south divide?

One of the key hindrances in large scale adoption of AI in the global south is the high cost of technical infrastructure. Open Source AI simply brings down that cost by up to six times with minimal performance reduction. Such systems are interoperable thereby further underlining the utility potential.

What are its limitations?

One of the major drawbacks in the development of open source AI is the lack of incentives. Open Source AI is everything ‘tragedy of the commons’ since it is centred around community development. There is a general lack of awareness about the performance and overall cost-effectiveness of these models which results in even lesser adoption.

What lies ahead for the technology?

Technology governance has now become more about digital sovereignty. The governments are increasingly realising that the development of technology is becoming a state instrument. Some governments are taking the regulatory route, wherein they are putting strict checks on AI development, while others are focusing on building capacity. The Indian government is adopting a mix of these measures where it is regulating through a light-touch approach and, at the same time, investing in digital public infrastructure based on open source. The success of these measures, however, could only be measured in the coming time.