India’s AI Hardware Dream Takes Off

“I really think chips are the new oil. While oil defined the 20th century, chips will define the 21st century and every country will want to own elements of manufacturing,” Krishna Rangasayee, the CEO and founder of SiMa.ai, told AIM.?

SiMa.ai has a significant presence in India, and around half of the startup’s innovation happens in India. Interestingly, the company has chosen India as a strategic market over China, the largest semiconductor market in the world.

The company is betting on India’s renewed focus on semiconductor manufacturing and fabrication ambitions. “I think India is going to be an AI product-focussed leader globally and a lot of innovation in AI will come from India,” said Rangasayee.

The company has chosen not to sell its chips in China, which is indeed a bold move for a startup. It has also outlined plans for a third generation of chips after the first one on 16 nm and the second generation on 6 nm nodes.?

Rangasayee mentioned the company’s potential to sell its chips to phone and PC makers, possibly for edge use cases. However, the startup has no intention of developing chips for data centres.

Speaking of Edge, Vicharak is on the Move

Starting a hardware company based in India is no easy task. However, Surat-based company, Vicharak, took on the herculean task of churning out hardware in-house, designed specifically for AI workloads.

Vicharak’s founder and CEO Akshar Vastarpara said that the company’s focus is not just on creating hardware, but redefining computing technology. “Our first target is to develop a GPU-like technology that can be used in mobile phones, laptops, and servers.”

This led to the creation of Vaaman – a complete packaged computing board that boasts a six-core ARM CPU and an FPGA with 112,128 logic cells.

The company recently secured funding of INR 1 crore, boosting its valuation to INR 100 crore.

Vicharak’s products are poised to revolutionise single-board computing. “We are in the same industry as Raspberry Pi, but our boards include FPGAs alongside processors, offering a complete AI infrastructure.”

The Silicon Valley of India

The CEO of Ola and Krutrim, Bhavish Aggarwal, is also building AI chips in India. Post his meeting with Synopsys CEO Sassine Ghazi, he announced “exploring the development of India’s first AI chip”. As per his post, further details on this can be expected on August 15.

This came after Aggarwal visited Taiwan for Computex 2024 and met Arm chief Rene Haas. Moreover, the company also silently acquired Bodhi Computing, a startup which builds and sells server-grade systems. Ola is developing advanced RISC-V and AI solutions for data centres and transportation.

What About the Others?

Not just hardware, Vicharak is also in direct competition with NVIDIA’s CUDA with its focus on software. Its flagship product, Gati, exemplifies this vision. “Gati is our AI exploration project. We’re writing our own infrastructure on top of FPGA, creating a stack similar to what NVIDIA does with CUDA,” said Vastarpara.

Every company wants to take a slice of what NVIDIA is doing, but it’s not just Indian companies. Most recently, San Francisco-based AI hardware company Etched unveiled Sohu, the first specialised chip (ASIC) built exclusively for Transformer models.?

The founder and CEO Gavin Uberti calls it the ‘biggest bet in AI’.?

According to the company, “Sohu is more than 10 times faster and cheaper than even NVIDIA’s next-generation Blackwell (B200) GPUs.” Still, the risk is that it is only made for Transformer-based architectures. In other words, the success of Etched depends on the future relevance of Transformers.

Groq, like Etched, is gaining traction with its LPUs. Even big-tech companies are making their own chips. Regardless, NVIDIA is undoubtedly on the top.?

Meanwhile, India, one of the biggest future markets for AI workloads for any hardware provider, has decided to make its own hardware.?

Is India’s hardware dream finally taking off? Only time will tell, but it doesn’t seem too far away.

TECH TALKS

SiMa.ai CEO and founder Krishna Rangasayee spoke with AIM about the startup's latest chips, fabricated by TSMC on 6 nm nodes. He said that these are engineered specifically to support multimodal generative AI applications on drones, autonomous vehicles, and humanoid robots.

However, despite being a big market, the startup is not engaging with China. He discusses how the convergence of semiconductors and AI is shaping current geopolitics.?

NEWS BYTES

• Nandan Nilekani revealed that Infosys is currently managing over 225 generative AI programs for its clients and also now boasts over 250,000 employees trained in generative AI.
• Lenovo has introduced a suite of AI solutions and the sixth generation of its Neptune liquid cooling technology to improve the practical AI deployment in Indian enterprises of all sizes.
• Hugging Face has unveiled a new Open LLM Leaderboard, which would rank open-source LLMs based on extensive new evaluations, including the MMLU-Pro benchmark.
• Less than two weeks after the launch of Lamini Memory Tuning, Lamini AI has officially partnered with Meta.
• ElevenLabs has recently released a new iPhone app that allows users to listen to anything from a complete novel to the contents of a website.
• Fractal has announced the merger of Eugenie AI, an AI company dedicated to providing AI-driven products for climate change and industrial sustainability.