Can InCore fulfill India’s semiconductor dream with scalable processor IPs?

For decades, British company ARM dominated the intellectual property (IP) market in the semiconductor industry. However, the advent of RISC-V, an open instruction set architecture (ISA), a few years ago challenged Arm’s market monopoly, providing a much-needed transition to global semiconductor ecosystem.

Many leading fabless design companies such as Qualcomm and NVIDIA, and even Samsung and Google, which previously depended solely on SoftBank-owned ARM’s IP offering, have also started developing their own RISC-V processors for a variety of applications.

Domestically, IIT Madras sparked a major revolution with the country’s first home-grown SHAKTI microprocessor, which was open source and built on RISC-V.

Started around 2012, the SHAKTI program was led by the current director of IIT Madras, Professor Veezhinathan Kamakoti, who was soon joined by GS Madhusudan, Neel Gala, Arjun Menon and others.

After several years of collaboration, Madhusudan, Gala, and Menon decided to apply their SHAKTI insights to the development of scalable, commercial-grade IP addresses. Additionally, what fueled their motivation was the growing IP semiconductor market that was about to be cracked to USD 15 billion by 2032.

Realizing that more semiconductor companies would need vertically integrated, customizable IP cores in the near future, given the growing market for consumer electronics, automotive, artificial intelligence and various computing devices, they founded the company InCore semiconductors in 2018

The startup is currently focused on developing IPs for semiconductor processors based on the open source RISC-V architecture. Its product suite can be integrated with chips for a variety of embedded applications, including wearables, smart IoT devices, PoS terminals, and more. Notably, Gautam Doshi, an experienced industry expert, later joined the initiative.

The beginnings of InCore

InCore is not a chip manufacturer without factories, it is an IP company that serves system-on-chip (SoC) companies and helps them design chips according to their diverse customer requirements.

For the uninitiated, in semiconductor parlance, IP addresses are a collection of specifications, source codes, and other information required to produce a given semiconductor.

Going back to the story, after working on building the SHAKTI processor, the founders realized that scaling the program had its challenges as there were requirements for more independent research and efforts to get it into shape that could function as a plug-and-play IP play for SoC Company.

“We realized that open source software, at least in terms of hardware, was not scalable. It has a lot of nuances and ecosystem constraints that a commercial entity like InCore would want to adopt,” said co-founder and chief technology officer, Gala.

But just four years after the startup’s founding, in 2022, InCore felt confident that RISC-V would remain in the market as some of the top semiconductor companies began betting on the technology and the community behind it grew.

After almost five years of being in bootstrap mode, InCore has raised $3 million in an initial round of funding, starting at Summit XV in 2023 to make his dream a greater reality.

Currently, InCore has its own proprietary offering that includes azurite and calcite cores.

Talking about how the startup differs from SHAKTI’s academic research, Gala said, “We are more focused on productization, refining last-mile engineering and optimizations to make the actual processors shippable. They address actual workload, current issues… We are not specifically going back to SHAKTI or adopting SHAKTI beyond a certain point, rather SHAKTI has been the foundation of everything InCore is today.”

He added that despite the dependence on SHAKTI, the overlap is minimal (5-10%). The company claims to have rewritten most of the code and added new processors to the lineup.

What makes up the InCore core?

Gala noted that when it comes to IP addresses, InCore needs to be factory-agnostic so that SoC customers can choose any factory for final chip production.

It also required InCore to spend a lot of time characterizing its IP addresses for specific target segments.

“That’s why we don’t build one processor at a time. We create a code generator or processor generator that can be configured and customized. From a single line of code or from a generator, you could provide a spec and say you want the smallest possible code, or you want high-performance code, or you want to balance them, and the generator would spit out the required IP address,” Gala explained.

InCore’s Azurite is a tiny core equivalent to the ARM Cortex M0 and Cortex M4 processors. It is intended for applications such as energy meters, smart metering systems and smart card applications that operate at 50-250 megahertz computing operations.

Its Calcite is the core that runs Linux. Its target segments are computing requiring operations at frequencies up to 1-1.4 GHz, which include point-of-sale terminals, wearable devices and network systems.

InCore is currently piloting its IP offering with several companies in India and around the world. The startup is still in the pre-revenue stage and plans to start generating revenue by the end of 2024.

Meanwhile, the founders focused on minimizing the product line. However, this does not mean giving up the ability to address multiple segments or provide more solutions beyond cores.

Over the past few months, InCore has also started working on other design components such as network-on-chip, interconnects and fabrics, which is expected to increase its revenue stream in the long run.

The future of InCore

Gala emphasized that the maturing period for semiconductor manufacturers is too long, and for intellectual property providers it usually takes six months to a year to reach any transaction. This is because each process goes through a rigorous evaluation process, legal procedures and testing.

Meanwhile, Indian semiconductor startups face the difficult task of building trust in the global market. This particular issue proved painful as Gala and the other founders tried to raise funds.

Later, the founders realized that the semiconductor industry was incomprehensible to many, so the flow of patient capital in this ecosystem was skewed.

Despite the obstacles, the startup is committed to moving forward and is currently working on creating its next core offering, Dolomite, which will be suitable for mobile, storage and other high-performance embedded devices.

By mid-2025, this offer will be ready for implementation. With this, InCore claims to be able to cover all core offerings provided by ARM.

The startup claims to have already built a pipeline of 10-12 customers. Currently, InCore is focusing more on player capture in the Bay Area.

The startup also does not care about raising funds immediately. However, the next round will not be launched until next year to support the next level of growth after achieving several milestones, such as obtaining proof of effectiveness of the technology on silicon chips and introducing several additional product offerings.

The main plans are already being finalized. Now it remains to be seen how InCore will make the country’s dream of becoming a semiconductor hub a reality, all while working in the not-so-decorated IP construction segment, especially when much of the focus is on building semiconductor fabs and supporting design firms in the Indian semiconductor sector the market is expected to exceed $150 billion by 2030.

(edited by Shishir Parasher)