Artificial Intelligence & Machine Learning

Nvidia Charts AI-Powered Computing Evolution in India

Nvidia founder and CEO Jensen Huang outlined his company's vision and plans for the next wave of AI at the Nvidia Summit in Mumbai, India, on Oct. 24. Engineers, data scientists, Indian business executives, analysts and celebrities gathered at the event to hear Huang's vision. The event, which included several major partnership announcements, marked a significant milestone in Nvidia's engagement with India's AI ecosystem.

Huang began by tracing computing's evolution, acknowledging the IBM 360 that introduced nearly every industry to IT. He explained how Moore's law - the doubling of transistors on a chip every 18 months - and the Windows operating system democratized computing through multitasking capabilities.

The Nvidia CEO described the "seismic change" and "tectonic shifts" ahead as computing approaches the limits of Moore's law. CPUs will be replaced by GPUs to drive the intelligent age of computing, and AI agentic models will enable everyone to be 20% to 30% more productive. Multilingual capabilities in language models will [commoditize this] and lift more people out of poverty. "Moore's law enabled the doubling of software capabilities and reduced costs every two years - which made it possible for society to use more IT … But Moore's law has ended [for scaling CPUs]," Huang said. "We have to do something different now, as computing depreciation will end, and we will move to computing inflation."

He detailed how Nvidia's CUDA architecture has accelerated applications "20 to 50 times" across industries. "In the age of Software 2.0, the machine is learning and writing the software. And more humans can now develop applications without coding knowledge - using natural language prompts, thanks to generative AI," he said.

In a moment reminiscent of Steve Jobs' famous "one more thing" presentations, Huang lifted a heavy array of Blackwell GPUs. He described Blackwell as a system designed to analyze data at unprecedented scale to help enterprises "discover patterns in relationships and learn the meaning of the data."

The computing paradigm shift brings new challenges. "Compute cycles are longer and have more permutations when computers start to think, as they solve complex problems. That's why speed and scaling are crucial. Technology is now moving four times a year," Huang said. To address these challenges, Blackwell architecture incorporates "acceleration levers" built on two key platforms: Nvidia AI Enterprise and Nvidia Omniverse.

Nvidia's India Strategy

Nvidia's presence in India runs deep, with more than 10,000 engineers working on GPU designs. The company has forged partnerships with leading Indian technology firms including TCS, Wipro, Yotta, RIL Jio, L&T Technology Services Limited, Tech Mahindra, Zoho and others to make gen AI and agents more inclusive for India's linguistically diverse population. This focus on India stems from its demographic advantages: a population of nearly 1.5 billion, with 65% of the population aged under 35 years, though the linguistic diversity - 25 languages and hundreds of dialects - presents unique challenges.

Huang underscored India's distinctive advantages in the AI race, pointing to its vast IT talent pool, extensive digital data resources and massive user base that can create an "AI flywheel."

Mukesh Ambani, chairman of Reliance Industries - RIL - joined Huang on stage, announcing plans to develop the best Hindi language model. He noted the serendipity that "Vidia" means knowledge in Hindi. "When you devote yourself to knowledge, prosperity will follow and drive the intelligence revolution. It will drive prosperity for 8 billion people," Ambani said.

RIL is developing extensive AI infrastructure powered by Nvidia's latest GB200 technology. Following Jio's mobile phone model, which offers data at 15 cents per gigabyte compared to the global average of $3.5, this partnership aims to make AI accessible and affordable to India's entire population.

Strategic Partnerships

The summit showcased several transformative partnerships reshaping India's AI landscape.

Wipro Limited unveiled initiatives using Nvidia's complete AI stack to help clients across healthcare, communications and financial services sectors develop AI-driven business strategies. The collaboration focuses on rapid implementation using Nvidia AI Enterprise.

Yotta Data Services introduced its Shakti Cloud platform with six AI services, becoming India's first provider of Nvidia NIM microservices. The platform will give organizations access to leading AI models such as Llama and Mistral for creating custom AI agents and copilots for clients in healthcare, manufacturing, telecommunications, financial services and retail sectors. Yotta also partnered with Sarvam AI to develop Sarvam 1, India's first open-source foundational model, trained on Shakti Cloud infrastructure.

Gnani.ai launched an innovative speech-to-speech large language model, powered by Nvidia's AI-accelerated computing platform. The system processes more than 10 million voice interactions daily, serving more than 150 enterprises across India and the United States.

Tech Mahindra established a center of excellence powered by Nvidia platforms, focusing on sovereign LLMs, agentic AI and physical AI. The center employs Tech Mahindra's Optimized Framework alongside Nvidia AI Enterprise software to deliver customized AI applications.

TCS expanded its five-year Nvidia collaboration, creating a specialized business unit within its AI.Cloud division to deliver industry-specific AI solutions at scale.

Sify Technologies introduced GPU Cloud, offering GPU-as-a-Service through its CloudInfinit+AI Platform. The service supports compute-intensive tasks including machine learning, deep learning and scientific simulations on a pay-as-you-go basis.

Zoho Corporation announced plans to integrate Nvidia-powered LLMs across its SaaS applications, committing $20 million to Nvidia's AI technology and GPUs over two years, benefiting its more than 700,000 global customers.

L&T Technology Services opened its Nvidia AI-powered Experience Zone in Bengaluru, India, creating an immersive environment for mobility and technology sector clients to engage with AI solutions.

Expert Analyses

Gartner Senior Director Analyst Naresh Singh shared his perspective on India's AI ambitions with ISMG. "The government is showing the way in building India as a technology powerhouse. AI is one aspect in terms of how the country progresses, not just in shaping the technology world but also in building India as a technology provider - moving beyond back-office services to developing new technologies," he said.

Singh acknowledged India's current position, contributing only 2% to 3% to the global IT industry. "These are still early days, but the startup ecosystem we've created in the last five years provides an excellent foundation for creating world-class technologies," he said.

He pointed out that there are many automation companies in India who already use AI. But gen AI "is completely new … It's a very new technology that has driven AI into a new era. So gen AI is certainly a good opportunity for Indian technology providers and not just the startups," he said.

Janakiram MSV, founder and principal analyst at Janakiram and Associates, highlighted the summit's key message: India's transition from software to intelligence exports. "Exporting software is fairly well-defined, straightforward and it is well-oiled machinery, but exporting intelligence is very different from exporting software. So he [Haung] wants India to invest in the entire [Nvidia] stack, all the way from accelerated computing and GPUs to the topmost layer, which does the intelligence inference and the use cases," Janakiram said.

Some summit attendees expressed measured skepticism about the AI revolution. An anonymous source drew parallels to cloud computing's evolution. Cloud technology promised cost savings and robust security, but CIOs now face unexpected costs and data residency concerns, leading many to pull back applications and data to on-premises solutions and adopt hybrid models.