Chennai: Shashwath T. R., co-founder and chief executive of Chennai-based semiconductor startup Mindgrove Technologies, sounds hopeful.
“Our chip is out of the lab,” he says. “We are coming out of the lab and into the industry.”
Since late 2024, Mindgrove’s engineers have been chasing a singular goal: perfecting a dedicated ‘vision system-on-chip’ tailored for security hardware. That persistence scored its first breakthrough this June, when the five-year-old venture struck an alliance with Prama India to integrate its homegrown silicon directly into the domestic company’s CCTV cameras and surveillance networks.
What created an opening for Mindgrove was a policy change in CCTV rules. In April 2024, the Indian government introduced several requirements for security reasons, including disclosure of the origin of critical components such as the chip. From 1 April 2026, non-certified imported cameras could no longer be sold, effectively shutting Chinese brands such as Hikvision and Dahua from the internet-connected CCTV market.
But even a mandate can only do so much. The memorandum of understanding (MoU) is not an immediate commercial order. The prototype is expected by the end of the year, after which Prama will have to qualify the product before a commercial order can be placed.
Another startup, BigEndian, is building what its co-founder and chief executive Sunil Kumar calls a “camera silicon prototype”. The two-year-old is also stitching up an ecosystem of design and manufacturing partners in Taiwan.
Many mass-market electronics brands in India source components from China and assemble them into products. “They will not be able to shift to a new chip easily,” Kumar says. BigEndian’s Taiwan ecosystem is intended to make that transition easier by giving customers a configurable end product.
BigEndian started work on its chip architecture in September 2024 and sent its chip blueprint to a manufacturing foundry in Taiwan in January this year. The first silicon was expected in July and is now arriving in September; the prototype is expected to be ready around November. Any revenue, realistically, would only be generated around April 2027, roughly three years after work on the chip began.
Both Mindgrove and BigEndian are fabless chip companies—they design chips but rely on specialist manufacturers to produce them. The two companies illustrate the challenge facing India’s youngest crop of such startups. They have spent significant amounts of capital getting their chips from design to production. And now they must find customers that could help them scale.
The CCTV policy change was an exception that created demand—it was a ‘stick’ approach, not a carrot. The larger question is whether Indian original equipment manufacturers (OEMs) are genuinely interested in India-built chips. And does it even make economic sense? Won’t they be seduced by cheaper alternatives from Taiwan and China?
For the startups and their investors, timing is crucial. How long can young companies afford to wait for the market to make the switch?
Their first few years of building chips have exposed many gaps, from dependence on foreign-owned intellectual property (IP) and design software to reliance on global foundries and suppliers where they have little bargaining power. These add to the time, cost and uncertainty of getting an India-designed chip into a product.
While India has spent substantial policy capital helping create this generation of chip companies, the ecosystem they need to scale remains incomplete.
A short history
First, let’s take a peek at how chip policies evolved.
Around 2007, the India Electronics and Semiconductor Association, then led by Poornima Shenoy, started pressing the central government for a policy to make chips in India. The rationale back then was the same as you hear today—India needs its own chip ecosystem for strategic autonomy, and to save on precious foreign exchange. Over the next many years, consortiums to make fabs emerged and fizzled out, largely because they either couldn’t find customers or close the required funding.
But India kept building on its strength in chip design. The country accounts for about 20% of the world’s semiconductor design engineers, according to government estimates. But these engineers are mostly employed by global chipmakers such as Intel, Qualcomm, Broadcom and Nvidia. Over the decades, these companies built large design and R&D centres in India.
Industry associations, meanwhile, kept pushing for homegrown chip companies and production infrastructure. Apart from fabs, this infrastructure includes outsourced semiconductor assembly and testing facilities (OSATs).
The result was the India Semiconductor Mission, launched in 2021 and backed by a ₹76,000 crore framework covering manufacturing, packaging, testing and chip design. Its incentive scheme has supported 24 design projects, while 105 startups and small companies have gained access to tools used for designing, simulating and verifying chips.
About 12 manufacturing projects worth more than ₹1.64 trillion have thus far been approved, with three—from Micron, Kaynes Semicon and CG Semi—in commercial production. Tata Electronics is building a ₹91,000 crore fab in Dholera along with PSMC.
The India Semiconductor Mission 2.0 was next approved in July this year with a budget of ₹1.27 trillion. It expands the focus to chip design, equipment and materials, fabrication, packaging, R&D and talent.
“We must develop an end-to-end semiconductor ecosystem in India,” says V. Kamakoti, director of IIT Madras and an adviser to the government on its semiconductor strategy. “A fab is the only missing piece of the puzzle.”
For Kamakoti, establishing a commercially successful 28nm fab will create the manufacturing base and know-how needed to progressively move towards more advanced nodes.
There is little doubt that India’s fabless story will move faster given the country’s design capabilities. Startups, however, unlike entrenched multinationals, face several headwinds.
Foundational problems
From an OEM’s perspective, Indian-designed chips will be evaluated by the same global benchmarks of performance, reliability, lifecycle support, supply assurance and cost competitiveness, says Pankaj Mohindroo, chairman of the India Cellular and Electronics Association.
It is difficult to beat Chinese or Taiwanese chip makers on price because of the ecosystem the countries have created.
A chip is hardly designed from scratch. Companies license pre-designed building blocks or IP for functions such as memory interfaces, USB and PCIe connectivity. They also depend on electronic design automation (EDA) software to design, simulate and verify chips before they are manufactured.
The costs simply add up for Indian players.
BigEndian’s Kumar says the pool of IP needed to kick-start a chip can cost an Indian company about $5 million compared to less than $2 million for a Chinese company. Even before a chip project is off the ground, startups are at a competitive disadvantage.
China, Kumar argues, built up its semiconductor industry by developing these foundational blocks first. India, he adds, should put more energy and money into building IP along with EDA tools and AI workflows. The global EDA market is dominated by Synopsys, Cadence and Siemens EDA, whose commercial tool suites can run into millions of dollars.
The geopolitical implications are also becoming harder to ignore. The Trump administration’s 2025 restrictions on Chinese access to EDA tools showed how software and intellectual property can become instruments of geopolitical policy. If Indian companies were to similarly lose access, the country’s fledgling fabless chip industry could effectively be brought to a standstill.
Jyothis Indirabhai, co-founder and chief executive of Netrasemi, a company building edge AI system-on-chips that target applications like smart cameras for surveillance, says he has tried to reduce dependence by developing AI and vision accelerators, signal-processing and video-compression engines before building the chips. But it still had to buy critical advanced memory and interface IP.
S. Krishnan, secretary, ministry of electronics and information technology (MeitY), acknowledges that foundational IP remains a problem, but says EDA tools present an even bigger challenge. The government plans to support companies developing EDA tools, he says, but fabs that have established EDA flows will have to agree to use them once developed.
The cost of being small
For the foreseeable future, India’s chip builders are dependent on global manufacturing where they have little bargaining power.
Kumar says BigEndian was told it would take three months to secure a memory-chip requirement, while US-based companies he knew could get the same component in two weeks. Its own prototype was supposed to arrive in four months but took close to seven months. In another example, a component that BigEndian had budgeted at $1.80 ultimately cost about $8 after accounting for currency depreciation, logistics, insurance and other costs.
Indian startups are more likely to be deprioritised than large semiconductor companies or startups from the US and Europe, says Raja Manickam, founder and chief executive of fabless power semiconductor startup iVP Semiconductor. He previously co-founded Tessolve, a semiconductor engineering and testing company, and was formerly the chief executive of Tata Electronics’ OSAT business. A startup ordering 1,000 wafers a month cannot command the same “capacity and mindshare” from a foundry as a company ordering 100,000.
Put together, this makes Indian chips slower and more expensive to build.
MeitY’s Krishnan does not dispute that Indian companies are at a disadvantage. The government’s response? Aggregate demand from several Indian companies so the foundry gets a larger combined order which could result in better terms and priority.
“This is a request that the companies have raised with us,” Krishnan says. The approach is still being worked out.
But manufacturing is only part of the equation. Someone still has to buy the chips, and soon.
Customer first
Some semiconductor founders are starting with customers before they start designing.
Fabless semiconductor startup HrdWyr approached potential customers and signed up with audio and wearable electronics brand boAt before beginning chip design. It is building a custom chip for boAt's TWS earbud charging cases.
“India has the talent, the market and the policy momentum to build a globally competitive semiconductor ecosystem. The next step is turning that capability into commercially viable chips that meet global benchmarks on performance, reliability and scale,” says Gaurav Nayyar, chief executive of boAt.
Parag Naik, who co-founded Saankhya Labs and sold a majority stake to Tejas Networks in 2022 for ₹283.94 crore, is taking a similar approach with his new startup Sampige. “It helps having a customer first because we know upfront the price that we need to achieve, the volume commitment, the unique IP, the unique problems they are facing,” he says.
But even established semiconductor executives find it difficult to open doors as startups.
“The Intel badge opened doors for me. As a startup, it’s not easy,” says Varma S. Konala, chief marketing officer at Pradhi Semiconductor, who spent nearly 25 years at Intel.
The problem is a classic chicken-and-egg one. An OEM wants confidence that a new semiconductor company can deliver a reliable product at scale, while the startup needs customers and commitments to build a chip.
To make the value proposition more compelling, chipmakers are attempting to pack in more features.
Mindgrove, for instance, has positioned its chips in the middle of the market, offering more features and flexibility than low-end alternatives while claiming it would be about 20% to 30% cheaper than comparable chips available in the market.
Netrasemi is taking a similar approach with its A2000, a higher-end vision chip designed to process video and run applications such as number-plate and face recognition, rather than competing in the commodity end of the CCTV market. HrdWyr’s first chip, Indus 1011, developed with boAt, claims to combine the functions of a microcontroller and power-management chip to improve battery management in wireless-earbud charging cases.
But these differentiators still have to translate into orders. “This is the biggest issue,” says Konala. “They will sign an MOU, but nobody will sign a binding agreement with you at that point.”
Startups are therefore expecting the government to play a greater role, be it through industry connects, mandates or incentives.
No permanent crutch
The government does not want to create a protected market in which Indian chipmakers survive indefinitely on subsidies. Krishnan says Indian chips ultimately have to compete globally. “We are not a closed economy. We are not going to say you should use only us,” he says.
He also argues that industrial policy must be judged by global and export competitiveness. Indian companies cannot depend permanently on government support.
That said, Krishnan agrees that Indian chip startups need some help to reach scale before they become globally competitive—it could be procurement or subsidy support. The modalities are in the works.
The CCTV experience provides one possible model. Krishnan says the government is now looking at whether a similar approach could work in other areas such as smart meters, IoT, industrial IoT devices and e-passports.
Meanwhile, for India’s fabless chip startups, the clock is already ticking.
“For us, each month is expensive. Every delay is expensive,” says BigEndian’s Kumar.