India must urgently build its own frontier-scale artificial intelligence models instead of remaining a passive consumer of global technology if it wants to play a meaningful role in shaping the future of AI, according to Pratyush Kumar, co-founder of Sarvam AI.
Speaking at the CII Business Summit 2026, Kumar said Sarvam AI plans to train its first trillion-parameter AI model within the next nine months, positioning the company at the centre of India’s push to develop homegrown AI capabilities.
Also Read: Lockdown in India? Oil Minister Hardeep Singh Puri reacts amid PM Modi's 'save petrol, diesel' appeal
“We can rent it for now until we don’t have it, but you have to build it. You have to own the destiny around that,” Kumar said, stressing that the conversation should no longer be about whether India should build its own AI models, but how quickly it can create strategic long-term capability.
Kumar said AI would emerge as the core intelligence layer powering sectors ranging from governance and manufacturing to science and industry, making ownership of foundational AI models critical for economic and technological value creation.
He warned that India risks repeating mistakes made during previous industrial and technological shifts by remaining dependent on foreign innovation.
“What we saw with steam engine, steel making and the internet, in all these eras, we became users and frankly lost out on key value creation,” he said. “This is now the start of a new era, which is going to play out quarter by quarter.”
Kumar said Sarvam AI has already demonstrated a proof of concept showing India can build large-scale AI systems end-to-end using domestic capabilities across data, algorithms, infrastructure and research.
Also Read: Uday Kotak puts India's troubles in perspective as the economy swings between hope and despair in the wake of Iran war
He added that India would need far greater investments in computing infrastructure and research talent if it hopes to compete with global AI leaders.
“The intelligence layer will accrue the value,” Kumar said. “It requires infrastructure, but it requires R&D talent to build these models.”
He also cautioned that India still lacks a clearly defined national AI strategy, warning against policy drift at a time when countries across the world are rapidly accelerating investments in the technology.
Speaking at the CII Business Summit 2026, Kumar said Sarvam AI plans to train its first trillion-parameter AI model within the next nine months, positioning the company at the centre of India’s push to develop homegrown AI capabilities.
Also Read: Lockdown in India? Oil Minister Hardeep Singh Puri reacts amid PM Modi's 'save petrol, diesel' appeal
“We can rent it for now until we don’t have it, but you have to build it. You have to own the destiny around that,” Kumar said, stressing that the conversation should no longer be about whether India should build its own AI models, but how quickly it can create strategic long-term capability.
Kumar said AI would emerge as the core intelligence layer powering sectors ranging from governance and manufacturing to science and industry, making ownership of foundational AI models critical for economic and technological value creation.
He warned that India risks repeating mistakes made during previous industrial and technological shifts by remaining dependent on foreign innovation.
“What we saw with steam engine, steel making and the internet, in all these eras, we became users and frankly lost out on key value creation,” he said. “This is now the start of a new era, which is going to play out quarter by quarter.”
Kumar said Sarvam AI has already demonstrated a proof of concept showing India can build large-scale AI systems end-to-end using domestic capabilities across data, algorithms, infrastructure and research.
Also Read: Uday Kotak puts India's troubles in perspective as the economy swings between hope and despair in the wake of Iran war
He added that India would need far greater investments in computing infrastructure and research talent if it hopes to compete with global AI leaders.
“The intelligence layer will accrue the value,” Kumar said. “It requires infrastructure, but it requires R&D talent to build these models.”
He also cautioned that India still lacks a clearly defined national AI strategy, warning against policy drift at a time when countries across the world are rapidly accelerating investments in the technology.
