CIO Insider

CIOInsider India Magazine

Separator

Krutrim to Launch AI Chip, Bodhi 1 by 2026

CIO Insider Team | Friday, 16 August, 2024
Separator

According to reports, Krutrim, the artificial intelligence startup of Ola founder Bhavish Aggarwal, plans to launch its first AI chip, Bodhi 1, by 2026 and a superior Bodi 2 two years later.

“Krutrim’s chips will be custom-built to handle complex AI workloads. They will enable the development of faster and more efficient AI systems,” says Bhavish.

The Bengaluru-based company is working on four chips: Bodhi 1, Bodhi 2, Sarv 1, and Ojas.

The company announces a partnership with global semiconductor companies Arm and Untether AI on the development of CPU and AI chips, platforms, and systems.

“While Bodhi 1 will be introduced in 2026, it will be designed for frontier LLMs, AI inference, and fine-tuning, and will have best-in-class power efficiency. Bodhi 2 will come by 2028 which will be able to support more than 10 trillion parameter models,” says Bhavish.

Sarv 1, its cloud-native CPU, will be introduced in 2026 while Ojas, an Edge AI chip will power next-generation Ola Electric vehicles, as per reports.

The timeline for chips looks ambitious, and beyond that, Ola needs to make the entire stack competitive to attract developers and be ready for high capital with an aggressive business model to be viable and compete well with the top hyperscalars

"The announced vision and roadmap look appealing, and the timing is apt to impress investors talking about AI and indigenous development of silicon," says Neil Shah, partner at Counterpoint Research and a technology industry analyst.

“We need to understand more about IP and foundry tape-out plans which is the first key step. The next challenging step is building the AI software stack where hyperscalars such as AWS, and Azure who have also designed their chips, are still majorly dependent on Nvidia and AMD for AI workloads and are years ahead. The timeline for chips looks ambitious, and beyond that, Ola needs to make the entire stack competitive to attract developers and be ready for high capital with an aggressive business model to be viable and compete well with the top hyperscalars,” Adds Neil.



Current Issue
AI Use Cases To Watch For In 2025



🍪 Do you like Cookies?

We use cookies to ensure you get the best experience on our website. Read more...