But even though so much semiconductor design work is being done in India, it’s all happening in MNCs and not Indian companies.
Now, a new breed of deep tech hardware startups is emerging to fill this gap. Bengaluru-based Semiconductor Fabless Accelerator Lab (SFAL), for instance, was formed last year. Initiated by a group within the India Electronics and Semiconductor Association, it has support from both the state and central governments.
“We have 40,000-50,000 engineers in semiconductor design and related fields working in India for MNCs. So now, harnessing this skill, we can build our own innovative designs and create intellectual property,” says SFAL chief executive officer Muthukrishnan Chinnasamy, who started his career at Texas Instruments in 1988.
Typically, chip design companies leave the manufacturing to others, because this is a less capital-intensive path. Big global players like Intel Corp. and Qualcomm Inc. do both, but chip design IP is where India can play to its current strength, just like with software.
Although the talent pool is abundant, designing and developing chips is still an expensive proposition for startups, requiring anywhere from $2 million to $5 million to create the IP. Venture capital is scarce for deep tech hardware startups, who have to rely on grants, angels and a few early-stage investors willing to place bets on unproven technology.
Where SFAL makes a dent in the capital requirement is in providing access to a common pool of chip design tools. “Roughly 20-30% of the capital is required for these electronic design automation tool sets, which becomes an entry barrier for entrepreneurs. So, we have reduced the burden of entry,” says Chinnasamy.
This has in turn generated global interest from tech companies that want to tap Indian startups that complement what they do.
For example, US-based Applied Materials, which provides equipment for manufacturing semiconductor chips, has started an accelerator programme called Astra in Bengaluru. This creates a channel into global markets and sources of funding. Some early successes have emerged to indicate the potential of Indian startups. Bengaluru-based Saankhya Labs, for example, founded in 2006 by Parag Naik, who returned home after working in Silicon Valley, survived its early years with angel funding as well as strategic investors who were buying its semiconductor and wireless communication products and services. Then it got backing from Intel Capital in 2011 and began expanding its global presence.
India has two main advantages in this space. The first is availability of talent with MNC work experience—not just founders, but also the top deck of engineers at these startups. The second is a faster, lower-cost go-to-market, because it is less expensive to hire talent here than in developed markets.
But there are substantial challenges: dependence on foreign funds and lack of proximity to end customers in developed countries. “Because the end customers are not in India, startup teams often don’t know whom to collaborate with and how to take the solution forward,” says Satish Mugulavalli, partner at YourNest Venture Capital, which has a focus on deep tech.
Investing in deep tech startups in a nascent ecosystem is a tough proposition. “I’m an engineer and it’s still hard for me to judge how much deeper and wider the technology moat of one startup is from another’s. There are a number of firms using light for computing systems, but they may not be solving the same problem,” says Manish Gupta, principal at growX Ventures, an early stage deep tech fund.
Against all odds
These funds have come up despite the challenges, driven again by Indians who have gained from international engineering work. Now, the wait is for larger, global funds to take bigger bets as more startups prove themselves.
A semiconductor startup ecosystem has taken a long time to emerge in a country where MNCs have been getting their chip design work done for the past 35 years. But it’s finally catching headwinds.
The roll-out of higher bandwidth 5G in India next year could be an inflection point as it will create opportunities for semiconductor startups. “When 5G networks are combined with AI and IoT, that’s where we will see a lot of bottlenecks. You will no longer be able to scale on the cloud because it will involve huge data transfer. So, you will need smarter devices for edge computing and that’s where semiconductor innovations will be needed,” says Vivek Kumar, launch manager, Entrepreneur First.
Sumit Chakraberty is a Consulting Editor with Mint.