'Entire offering is just a repackaging': AI founder tears into OLA CEO Bhavish Aggarwal's Krutrim

'Entire offering is just a repackaging': AI founder tears into OLA CEO Bhavish Aggarwal's Krutrim

Positioned as India’s answer to global AI infrastructure players, Krutrim aims to offer cloud-based access to powerful large language models, all hosted on domestic servers.

Advertisement
OLA CEO Bhavish AggarwalOLA CEO Bhavish Aggarwal
https://akm-img-a-in.tosshub.com/businesstoday/2023-04/logo.png
Business Today Desk
  • Apr 9, 2025,
  • Updated Apr 9, 2025 1:14 PM IST

Aaditya Aanand, a tech entrepreneur who is building multibagg.ai, tore into Bhavish Aggarwal's claims after the Ola CEO announced that Krutrim had become "one of the world's first" to host Meta's Llama 4 models on Indian servers. "After raising 280 million dollars, he is excited for hosting an open source model? An intern having a good PC with GPU can do it in their bedroom," Aanand said in a post reacting to Aggarwal's announcement.

Advertisement

Aggarwal had pitched the launch as a major leap in India's AI capabilities, citing domestic deployment of DeepSeek's AI models — ranging from 8B to 700B parameters — and now Meta's Llama 4 Scout and Maverick models. He highlighted pricing between ₹7 and ₹17 per million tokens and framed it as a move toward "democratising access to cutting-edge AI for every Indian developer and startup".

"This is a crucial step as true technological independence requires not just building our own models, but also creating the infrastructure ecosystem that lets Indian innovators build without compromise,” Aggarwal wrote on Linkedin. "This is how we build India's AI future—by creating the foundation that lets our talent compete globally while keeping our digital sovereignty intact."

Advertisement

But Aanand, a former associate at Goldman Sachs, dismissed both the pricing and sovereignty narrative. "Disruptive pricing? I think he hasn't seen the pricing of Gemini Flash models or GPT 4o-mini. FYI, that's less than ₹10 per million tokens," he wrote. "Democratizing access to AI for Indian developers? As if Indian developers were waiting when Krutrim will host these models. Being delusional is the only solution for him."

He also took aim at Krutrim's positioning as a nationalistic tech breakthrough: "Technological independence and digital sovereignty? Seriously? All he did was host someone else's model on his own servers and pretending like he has put the Indian flag on the moon."

According to Aanand, the effort lacks originality. "The entire offering is just a repackaging. You are not building India's future, you are just renting it locally. Patriotic chest thumping is no real innovation." Positioned as India’s answer to global AI infrastructure players, Krutrim aims to offer cloud-based access to powerful large language models, all hosted on domestic servers.

Aaditya Aanand, a tech entrepreneur who is building multibagg.ai, tore into Bhavish Aggarwal's claims after the Ola CEO announced that Krutrim had become "one of the world's first" to host Meta's Llama 4 models on Indian servers. "After raising 280 million dollars, he is excited for hosting an open source model? An intern having a good PC with GPU can do it in their bedroom," Aanand said in a post reacting to Aggarwal's announcement.

Advertisement

Aggarwal had pitched the launch as a major leap in India's AI capabilities, citing domestic deployment of DeepSeek's AI models — ranging from 8B to 700B parameters — and now Meta's Llama 4 Scout and Maverick models. He highlighted pricing between ₹7 and ₹17 per million tokens and framed it as a move toward "democratising access to cutting-edge AI for every Indian developer and startup".

"This is a crucial step as true technological independence requires not just building our own models, but also creating the infrastructure ecosystem that lets Indian innovators build without compromise,” Aggarwal wrote on Linkedin. "This is how we build India's AI future—by creating the foundation that lets our talent compete globally while keeping our digital sovereignty intact."

Advertisement

But Aanand, a former associate at Goldman Sachs, dismissed both the pricing and sovereignty narrative. "Disruptive pricing? I think he hasn't seen the pricing of Gemini Flash models or GPT 4o-mini. FYI, that's less than ₹10 per million tokens," he wrote. "Democratizing access to AI for Indian developers? As if Indian developers were waiting when Krutrim will host these models. Being delusional is the only solution for him."

He also took aim at Krutrim's positioning as a nationalistic tech breakthrough: "Technological independence and digital sovereignty? Seriously? All he did was host someone else's model on his own servers and pretending like he has put the Indian flag on the moon."

According to Aanand, the effort lacks originality. "The entire offering is just a repackaging. You are not building India's future, you are just renting it locally. Patriotic chest thumping is no real innovation." Positioned as India’s answer to global AI infrastructure players, Krutrim aims to offer cloud-based access to powerful large language models, all hosted on domestic servers.

Read more!
Advertisement