InMobi Says Power in Tech Is Shifting From Software to Proprietary AI Models

Proprietary Data to Define Next Phase of AI Leadership, Says InMobi Official​

New Delhi, February 22 The balance of power in the technology sector is moving away from traditional software toward artificial intelligence models trained on proprietary data, a senior official at InMobi said on Sunday.

The remarks come at a time when the Indian government is backing open source frameworks to build sovereign AI capabilities, even as debate continues over how India’s IT industry should position itself in the global AI race.

According to Tewari, the future strength of technology companies will not be determined by software alone. Instead, competitive advantage will increasingly depend on proprietary models trained on exclusive data sets.

“The future of technology companies lies in the fact that software is no longer the primary source of power. The power is moving towards proprietary models that you are essentially training using proprietary data. Therefore, even the most advanced models are becoming commoditized. The value lies in the models that use proprietary data,” he said.

India Needs Vertical AI Models Built on Advanced Foundations​

Tewari emphasized that India must build its own proprietary vertical AI models tailored for regional needs. These models, he noted, should be developed on top of existing advanced foundation models to remain globally competitive.

Among the leading global AI systems currently in use are ChatGPT, Google Gemini, and Claude 3, which power a range of AI driven services across industries.

At the India AI Impact Summit 2026, Indian startup Sarvam unveiled an advanced AI model positioned as a competitor to established global players, signaling growing domestic capability in high end AI development.

Scale and Data as Strategic Advantages​

InMobi stated that its mobile advertising services currently reach 2 billion users across more than 150 countries. The company has also developed its own agentic commerce system operating on over 100 million devices worldwide, including mobile lock screens and smart televisions.

Tewari underlined that AI systems improve with scale. Larger data sets enable better fine tuning and training, making user reach and data access central to long term competitiveness.

“Any AI model needs a large user base to fine tune and train itself. The larger the data sets involved, the better it becomes,” he said.

He added that the historical advantage once enjoyed by Western internet technology firms has narrowed.

“Now, it's a level playing field. We can actually test products on a much larger number of consumers and then make them global. That's our big advantage. We can build products from here and take them into global enterprises. That advantage never existed before, and I think we have essentially made that happen,” he said.

AI Could Add Over $3 Trillion to India’s Economy by 2047​

During his presentation at the India AI Impact Summit 2026, Tewari said artificial intelligence could generate more than $3 trillion in incremental economic impact for India by 2047 by transforming how commerce operates.

He said the company is focused on building commerce models trained on proprietary consumer data, which helps differentiate its AI systems.

“We are essentially trying to create commerce models that are trained on proprietary consumer data, which essentially trains those models and makes them different. The world is moving very rapidly. What you saw in the last two, three years is no longer valid,” Tewari said.

The comments underscore a broader shift in the global AI landscape, where ownership of data and domain specific models is emerging as a critical strategic lever for long term growth.