Intel’s India vision: From AI acceleration to enabling electronics manufacturing for the world

It’s hard not to notice India’s steady rise in the global technology conversation – be it in AI, semiconductor manufacturing, or data center innovation, according to Santhosh Viswanathan, Vice President & Managing Director, Intel, India Region, who has a ringside seat to all the cutting-edge action and raw dynamism shaping India’s tech future. 

In an extensive interview, he discussed Intel’s AI hardware approach, the “Make in India” initiative, and insights into some distinct advantages of India both as a massive consumer market and a hub for future-looking R&D. Santhosh Viswanathan’s perspective sees India not merely as a place to do business but as a crucible for invention – especially in AI, where more than one-fifth of the world’s data is generated. And according to him, integrating all that data with real-world solutions, though, demands a flexible approach. For Intel, that means avoiding the trap of “AI = GPU,” a phrase Santhosh is quick to dispel.

Shattering the “AI = GPU” myth

“Intel recognizes that AI solutions must be tailored to meet the unique needs of different markets,” Santhosh says, highlighting how crucial it is to sidestep the assumption that AI is always about a GPU. The company’s strategy, he explains, hinges on one overwhelming truth, which is that AI can run on many kinds of hardware.

“For instance, Xeon processors are designed to handle a wide range of AI workloads, from edge to cloud, while Gaudi AI chips offer cost-effective AI acceleration for training and inference tasks.”

Also read: Intel Xeon 6 P-Core chips aim to outpace AMD and boost AI datacenter business

But what about GPUs and large language models (LLMs)? NVIDIA has become the darling of AI hype in recent years, leaving many to wonder how Intel competes or even coexists in that space. Santhosh acknowledges the heavy odds at stake but offers an alternative that’s not just necessary but demanded by tech companies.

“AI is not a one-size-fits-all solution. For many, AI is just another workload running alongside hundreds of others. Sometimes you don’t need the highest performance on a single function – just enough to meet the SLA and run efficiently.” He points out that many businesses use LLMs not to train from scratch, but to refine or fine-tune existing models – tasks well within the capability of a CPU, especially a modern Xeon processor.

So, where do Gaudi chips come in? “Our Gaudi AI chips provide competitive performance and cost-efficiency for large-scale AI training and inference, making them an attractive option for enterprises looking to scale AI deployments,” he says. At the same time, he doesn’t deny that smaller or medium-sized models can be effectively managed on general-purpose Xeons. The overarching theme is that Intel wants to offer a continuum of solutions, not just a single one.

That’s the mantra he keeps returning to – “AI does not equal GPUs.” In many enterprise scenarios, a CPU might be more than enough to handle fine-tuning a smaller large language model or to run certain machine learning tasks. Relying solely on a GPU for every job can be overkill – and an expensive one at that. The future, from Intel’s vantage point, is about heterogeneous computing: a CPU here, an accelerator there, and a flexible approach that ensures you’re using just enough horsepower to get the job done efficiently.

Also read: Complexities of Ethical AI, explained by Intel’s Lama Nachman

“Our recently launched Intel Xeon 6 processors with Performance-cores, for instance, allow for 5:1 consolidation of a five-year-old server on average, with potential for up to 10:1 in certain use cases,” Santhosh explains. “These features will enable deployment of AI in a frugal and sustainable way in India.”

Why does that matter so much? Because as AI expands beyond research labs and top-tier data centers, it starts trickling into smaller businesses, municipalities, and edge devices. If the hardware is too power-hungry, which is usually the case with GPUs (as any gamer will agree) you can bet it won’t scale seamlessly in a region where resources can be more constrained, especially outside major metros.

India’s AI ecosystem: Rich in data and talent pool

It’s not just about hardware. India is bursting at the seams with data – and data is, after all, the lifeblood of AI. According to Santhosh, “India generates nearly 20% of the world’s data,” an astounding figure that underscores why so many tech giants are eager to plant deeper roots here. He also believes the vast local talent base is a key advantage, and Intel is keen on nurturing it through global system integrators and independent software vendors.

He name-drops an interesting collaboration with Infosys on an AI-first suite of solutions called Infosys Topaz, leveraging Intel Xeon processors and Gaudi 2. Another one with Bud Ecosystem for making generative AI inference possible on CPUs. “Bud has also trained and optimized foundational models on Gaudi, creating real-time private agents that can run on AI PCs.” 

Also read: AI Laptops vs Traditional Laptops: What’s the real difference?

In addition, Santhosh also points out a couple of interesting local collaborations that highlight what Intel sees as a blueprint for the future. “There are many great AI examples in India, including Intel’s collaboration with Haystack Analytics to create an edge compute solution for faster tuberculosis testing.” He also mentions a Generative AI model designed to converse in many Indic languages, done in partnership with Tech Mahindra.

“We’re expanding AI education and digital readiness programs,” he adds. Indeed, Intel has taught 1.6 million students under its AI for Youth initiative and offered AI awareness to over 4.5 million citizens via AI for All – two programs that, while massive in scope, might just be scratching the surface of India’s potential.

As far as Santhosh or Intel India is concerned, these partnerships are more than just business deals – they’re seeds for an ever-expanding AI ecosystem that’s starting to flourish in India. Because the potential payoff is huge, as IDC’s Asia/Pacific AI Maturity Study 2024 suggests India’s AI spending could hit $5 billion by 2027. That’s why Intel’s strategic approach is focused on putting the right hardware in place, developing local competencies, and relying on a supportive regulatory environment to accelerate economic and technological gains.

Make in India and Intel’s evolving role

Now, with a national push for “Make in India,” the question is how Intel’s role might morph, especially since it’s had a local presence in the country for over three decades, particularly through design centre investments. In case you didn’t know, a big part of Intel’s India operation is about design and engineering – all through Intel’s software and hardware products and services stack – supported by 12,000 employees and state-of-the-art facilities. “We’ve collaborated with policymakers, government bodies, and ecosystem partners to grow the electronics manufacturing sector,” Santhosh says. They’re providing early access to products, reference designs, and even supporting local manufacturing of HPC servers.

But the path isn’t smooth. He laments how India has historically been services-focused, whereas scaling manufacturing means developing strong product skills. Tying in the dynamic regulatory framework – often essential but occasionally adding complexity – Santhosh stresses the need for stable infrastructure like power and logistics. He believes India needs to work in two key areas going forward in order to scale up production. “Two key factors are critical for manufacturing at scale. The first is demand – without a strong domestic market, local production struggles. The second is innovation. We need to strengthen design and innovation capabilities, going beyond cost optimization.”

Also read: AI chips and Indian innovation: Insights from Tessolve CEO Srini Chinamilli

Santhosh points out that cost disadvantages and the need for a robust component ecosystem also remain key stumbling blocks. “We’re committed to upskilling the workforce, driving tech advancements, and ensuring India-specific designs can compete globally,” he insists, painting a picture where Intel is both a beneficiary and an enabler of India’s aspirations to become a heavyweight in electronics manufacturing. “Intel remains committed,” he adds, referencing the company’s willingness to work closely with government and industry to expedite advanced technology manufacturing.

Additionally, Santhosh sees India as a vital node in a more balanced global supply chain, especially relevant in a post-pandemic world wary of single-region dependencies. “The Make in India initiative is pivotal in fostering a diversified and resilient electronics manufacturing ecosystem,” he says, citing record-breaking $115 billion electronics output in FY24, propelled by mobile phone production.

He calls out the revised IT hardware PLI scheme and new proposals for electronic components as strong incentives, aiming to capture $500 billion in electronics production by 2030. It’s a big vision, but not without hurdles, as mentioned above. If done right, though, “India could become a key player in the global supply chain, boosting resilience and self-reliance,” he emphasises.

India’s role in the global semiconductor supply chain

Given all the hype about local fabs, how does Intel, which doesn’t have a fab in India, plan to engage and attack this opportunity, I ask him? In response, Santhosh sees synergy in advanced infrastructure and high-tech clusters. “Policy reforms, expanded PLI schemes, and a supportive regulatory framework are crucial. Public-private partnerships will also play a big role in driving the semiconductor ecosystem.”

Even without a local Intel foundry, the company can still shape the ecosystem by providing early products, design expertise, and forging alliances. He mentions a collaboration with HCL Tech on custom silicon solutions. “We support skill development, innovation centers, startups – our efforts combined with infrastructure and policy can position India as a key player globally.”

Also read: Nvidia wants to develop an AI chip with India, all details

The National Semiconductor Mission (NSM) launched in late 2021 aims to reduce import reliance, attract global investors, and foster domestic manufacturing. Intel’s ongoing investment and academic ties are crucial. “We have large design and engineering centers here, plus the Intel Unnati program. These position India as a key semiconductor ecosystem partner.”

With government funding for infrastructure, workforce training, production- and design-linked incentives, the pieces are falling into place. The question is how quickly everything aligns in practice.

Finally, no forward-looking tech conversation can ignore quantum computing. Santhosh highlights Intel’s unique approach in this frontier tech journey, which focuses on the entire quantum stack – architectures, algorithms, and control electronics. The company’s collaboration with QuTech has produced a 12-qubit silicon chip called Tunnel Falls, offered to academic institutions for immediate experimentation.

“We’re committed to collaborating with additional research institutions globally to build the quantum ecosystem,” he says. For India’s Quantum Mission, that could mean direct synergy, enabling local labs to skip the capital-intensive overhead of manufacturing. If quantum gains traction in India, Santhosh believes the ripple effects across computing, cryptography, and high-level problem-solving could be monumental.

In short, Intel’s presence in India is about more than just hardware. It’s about strategic collaborations, developer ecosystems, and a strong belief that despite all the challenges India can be a global tech powerhouse if nurtured with the right technology and skill-building efforts. If Intel’s bet pays off, India may find itself not just a consumer of tomorrow’s breakthroughs, but a primary driver of them.

Also read: Fear vs Facts: Two years after the AI revolution began