At the recent ‘Sankalp 2024’ event in Bengaluru, Ola founder Bhavish Aggarwal made several announcements about AI and technology. The event’s highlight was the launching of India’s first homegrown AI chip by 2026.
But why this focus on AI and homegrown tech?
If you’ve followed Aggarwal’s journey, this isn’t a new theme. He has repeatedly stressed the importance of building India’s own technology infrastructure.
Back in June, he laid out his concerns and said, “Our data is not truly ours—it’s owned and controlled by American platforms. Even our cloud infrastructure is run by companies like AWS and GCP. The chips we rely on are designed by giants like Intel, AMD, and Nvidia. Even our developer tools, from GitHub to PyTorch, are rooted in Silicon Valley. We don’t have control over any of this.”
He’s driven by the belief that if India wants to lead in the AI revolution, it needs to own its tech stack, from the chips powering AI to the cloud that stores our data.
According to Aggarwal, India must innovate at the hardware level too. He’s not interested in building just another gadget; he’s focused on crafting AI-driven edge devices with custom silicon to take on global tech giants.
So, Ola is stepping into the AI hardware game with its Bodhi-1 chip, designed for LLMs and inferencing tasks. It is expected to hit the market by 2026; it promises “class-leading power efficiency” making it a contender for widespread AI adoption.
It also announced Bodhi-2, a next-gen chip built for high-end AI workloads. Set for a 2028 release, this chip is poised to handle exa-scale computing.
Then there’s the Ojas chip, which Ola calls India’s first edge AI chip. While details are sparse, the buzz around Ojas suggests it will play a pivotal role in Ola’s future electric vehicles, offering AI-native architecture that could revolutionise the EV ecosystem.
As Aggarwal puts it, “There’s no middle ground, either we do or we don’t.”
For him, data sovereignty is about more than just keeping data in India; it’s about full control—from the cloud infrastructure to the chips themselves.
As the country is home to the world’s largest population and generates more data than others, Aggarwal believes that if India can master AI, using this vast data to its advantage, it could position itself as the most intelligent country on the planet.
Building an AI Chip
Coming to working on AI chips, Ola has already revealed a suite of AI chips designed for different applications. Though details remain under wraps, the Bodhi series of AI chips, the Sarv-1 cloud-native CPUs, and the Ojas edge AI chips are all part of the company’s vision for a more self-reliant India.
Partnering with companies like Untether AI, which provides next-gen AI acceleration solutions, Ola is gearing up to make its mark on the global stage.
While manufacturing AI chips involves the same process as other semiconductors: design, fabrication, assembly, and packaging, Ola is adopting a fabless model. This means relying on external foundries, like TSMC or Samsung.
Not to forget, Bhavish Aggarwal is in the good books of the central government. In early September, he met Union ministers Nitin Gadkari and Piyush Goyal to discuss the future of electric vehicles, green energy and ONDC in India.
But perhaps an important interaction must have been when he sat down with finance minister Nirmala Sitharaman, expanding the conversation on AI, another area where he seems eager to make a mark.
The Capital Story
Ola IPO’s recent success and its investments in technology, reflect its growing capital. Since debuting on August 9 to August 19, Ola Electric’s shares have surged by 92%, soaring well above their issue price of INR 76. With a market capitalisation around INR 63,000 crore, the company saw its IPO attract strong investor demand, oversubscribed by 4.45 times.
The company has allocated INR 1,600 crore for R&D, which could essentially contribute to AI chip development.
In its Q1 report, the company posted a net loss of INR 347 crore, compared to a INR 267 crore loss in Q1 FY24. However, revenues surged by 32%, reaching INR 1,644 crore for the same period.
So far, everything sounds promising for the AI chip development in India, but we need to wait until 2026 to see if these chips actually hit the market. And when they do, it’ll be exciting to understand how well they perform.
Until then, it is only about anticipation and hope.
The post Why Ola’s Bhavish Aggarwal is Bullish About Building ‘Made in India’ AI Chips by 2026 appeared first on AIM.
