L&T Semiconductor Aims to Be the NVIDIA India Never Had

The criticality of semiconductor self-reliance cannot be demonstrated better than the ongoing tariff wars between the United States, China, and India. Over the last decade, the Indian government has been pushing for the ‘Make in India’ initiative to reduce dependence on Taiwan, South Korea, and China. 

Yet, so far, nothing like an NVIDIA or Taiwan Semiconductor Manufacturing Company (TSMC) has been built in India. Nonetheless, companies like Larsen & Toubro, L&T Semiconductor Technologies (LTSCT), Tata Electronics, Vedanta Semiconductors and Reliance are in the process of establishing fab facilities across the country. 

Moving away from its identity as a services company, L&T launched a wholly owned subsidiary, LTSCT, in 2023. Since then, it has set its sights on becoming India’s first major semiconductor product company. 

Not a manufacturer-for-hire or a design house, but a full-fledged chipmaker—one that could eventually rival the likes of NVIDIA and STMicroelectronics from the West and MediaTek from Taiwan. 

Sandeep Kumar, CEO of LTSCT, is spearheading this ambitious vision. LTSCT formally started operations in November 2023, coinciding with the day Kumar joined L&T. However, this wasn’t a decision made on a whim.

In an exclusive interaction with AIM, Kumar revealed that he had been nurturing the idea since late 2020, driven by the belief that India must stop being the back office of the global semiconductor industry and start owning its chip IP.

“I had an overall business plan and strategy to build something meaningful in India and presented it to L&T, other large business houses, and the ministry,” Kumar recalled. “Most were only interested in fabs, but my idea was different—India needs to build product companies, not just factories.”

Eventually, it was L&T that recognised his vision and gave a green signal to the initiative.

Why Products Over Fabs?

LTSCT’s approach is simple, pragmatic, and rooted in the economics of semiconductors. It aims to generate revenue of $1 billion in the next four to five years and set up a fab within that timeframe.

According to the India Electronics and Semiconductor Association (IESA), India’s semiconductor market is projected to grow at a 13% CAGR, expanding from ₹4,50,164 crore ($52 billion) in 2024 to ₹8,95,134 crore ($103.4 billion) by 2030.

“In the semiconductor world, if you look at the value of product companies versus fabs and manufacturing, the ratio is about 70:30,” he said. “So, to me, bringing semiconductors to India means bringing Indian semiconductor products to market—not just making chips here that belong to someone else.”

For now, LTSCT is strategically betting on the fabless model, where the focus is on chip design and IP, while the manufacturing is outsourced. Notably, giants like NVIDIA and Qualcomm follow this strategy, and Kumar believes that with this approach, India is uniquely positioned to succeed.

“For over 30 years, Indian engineers have been designing chips for global companies. The talent pool exists. What we lacked was capital and a vision to build something of our own,” he said.

Now, with ₹830 crore in initial capital from L&T and access to top talent and customers, the foundation is in place. LTSCT is not targeting the Indian market alone. In fact, Kumar is quite clear that an India-first approach wouldn’t work in semiconductors—at least not yet.

Indian startups often focus only on the Indian market, which is neither large nor advanced enough. As a result, they lose out to international players.

Instead, LTSCT is developing a diverse range of chips—MEMS sensors, power ICs, analogue mixed-signal, RF products, and even silicon carbide and gallium nitride-based components—targeting mobility, industrial, energy, and telecom sectors globally.

The company already has a pipeline of over 50 customers, both Indian and international, many of whom are marquee names. “These are not startups,” Kumar emphasised. “They are premier, global customers.”

Sectors involving mobile handsets, IT, telecommunications, consumer electronics, automotive, aerospace, and defence are expected to drive steady growth. Notably, mobile handsets, IT, and industrial applications contribute nearly 70% of the industry’s revenue and are projected to remain the main drivers of this expansion.

This comes at time when union minister Ashwini Vaishnaw announced that India’s first chip design centres dedicated to developing end-to-end 3 nanometer chips will be built in Noida and Bengaluru. The facilities, set up by the India unit of Japanese semiconductor manufacturer Renesas Electronics, will offer products across automotive, industrial, infrastructure, and IoT sectors.

From Fabless to Fabs—But on Their Terms

This is when things become interesting. While the company’s focus is fabless for now, LTSCT eventually plans to establish its own fabs. However, unlike others racing to build foundries first, Kumar wants the demand to drive the need.

“Our thinking is different from Tata’s approach. They’re building the fab first, then trying to find customers to fill it,” he said. “We’re building products first. Once we hit $1 billion in revenue, which should take about five to six years, we’ll need a fab to keep up with production.”

The rationale is purely economic. An empty fab kills the margins. “If it’s 50% empty, your product cost doubles. So, we want to build a fab only when we have enough demand to fill it,” he said.

If all goes according to plan, fab construction could begin in the fourth year, allowing enough lead time for operations to start by the seventh.

Kumar said the choice of fabrication geometry—whether 7 nm or 55 nm—will depend on the products being manufactured. Right now, the focus seems to be on a range between 28 nm and 55 nm on the higher end, and potentially 7 nm on the lower end. The exact details will become clearer in a few years.

A Product Company from India, at Last

Regarding competition with China, Kumar remains clear-eyed. “A fab is just a factory,” he said. “The real competition is in the product.”

To explain, he offered an analogy: “A Toyota car factory doesn’t compete with a BYD factory. Toyota competes with BYD in the market. Similarly, we’re not competing with SMIC’s or China’s fabs—we’re competing with their chip products.”

For LTSCT, the question isn’t where the chip is made—it’s whether the chip is good enough to compete globally. India has long played a supporting role in the global chip industry, providing the talent but never owning the technology. That’s what LTSCT aims to change.

“There are global product companies from every major region—Intel and NVIDIA from the US, STMicro and NXP from Europe, Renesas from Japan, MediaTek from Taiwan, and iSilicon from China, but none from India,” Kumar said. “That’s what we want to build.”

LTSCT is aiming to do what no Indian company has done before—build India’s own NVIDIA. Moreover, if things go according to plan, the next globally recognised semiconductor brand might just be ‘Made in India’.

_{Images are for reference only.Images and contents gathered automatic from google or 3rd party sources.All rights on the images and contents are with their legal original owners.}