Behind Morphing Machines’ 20-Year Quest To Put India On The Global Semiconductor Map
semiconductor

Behind Morphing Machines’ 20-Year Quest To Put India On The Global Semiconductor Map

SUMMARY

Having existed for almost two decades doing research and development on reconfigurable computing, Morphing Machines has built a system-on-chip platform to meet diverse requirements across different industries

Operating in the country’s semiconductor market expected to breach the $150 Bn mark by 2030, Morphing Machines is looking at catering to domains such as AI and finance

Although the startup doesn’t have a steady source of revenue, it earns largely from the POCs it’s been conducting with its potential customers

As India’s electronics manufacturing industry gains momentum, it’s important to recognise the advanced technological capabilities and talent within the country’s semiconductor ecosystem.

Although India hasn’t been the brightest star in the realm of semiconductor manufacturing, largely due to China’s dominance and other geopolitical factors, the country has a long history of fostering innovative semiconductor companies with excellence in design and engineering.

Bengaluru-based Morphing Machines is an example of India’s rich scientific research and development capabilities. Having existed for almost two decades doing research and development (R&D) on reconfigurable computing, Morphing Machines has built a system-on-chip (SoC) platform to meet diverse requirements across different industries. 

Their intellectual property, REDEFINE, is said to be one of the few many-core processors in the world that integrates various domain-specific architectures (DSAs) on a single chip.

Now, before we delve deeper into the company’s history, it is critical to understand why semiconductors have lately become the talk of the town. Notably, these chips are essential in almost every modern device — from smart bulbs, smartwatches, and mobile phones to medical equipment, telecommunication gear, electric vehicles, aerospace technology, and satellites. 

Besides, with the advent of artificial intelligence (AI), the country’s semiconductor market is expected to breach the $150 Bn mark by 2030, growing at a CAGR of 24% from 2023.

Morphing Machines’ Two-Decade Quest To Chip Excellence 

Two Decades Later, Morphing Machine Cradles The Tech To Replace NVIDIA & Intel ProcessorsTwo Decades Later, Morphing Machine Cradles The Tech To Replace NVIDIA & Intel Processors

Founded by S K Nandy, a then-researcher at the Indian Institute of Science (IISc) in Bengaluru, Morphing Machines was incubated in 2005 at IISc. In 2006, his wife, Ranjani Narayan, also a former research scholar, joined Nandy’s endeavour.

The aspiration of the husband-wife duo, who came with a cumulative experience of almost 40 years as researchers in the field of Very Large Scale Integration (VLSI) design and high-performance computing, has always been to make a processor that would be futuristic and different from the ones designed by the likes of NVIDIA and Intel.

At the outset, the couple dedicated the first three to four years of their quest to working on the concept of the processor. In 2009, Morphing Machines started running simulations of the tech. This caught the attention of some defence sector laboratories, who found its technology intriguing. Research Centre Imarat (RCI), developing missile systems, was one among them.

This time around, the startup began getting small projects, nudging the founders to come out of the incubation mode and set up their first office with an initial staff of four to five individuals.

Morphing Machines made a giant stride with aircraft equipment manufacturer Safran in 2015 when the French multinational saw a massive use case for its processor (REDEFINE) in avionics. During its proof of concept (POC) phase, the Morphing Machines team started working on building a second and more enhanced version of its tech. This took another five years for the company. 

Speaking with Inc42, founder and CTO Narayan, said, “NVIDIA was all over the world at the time. So, we had to recognise our pitfalls and theirs – what is NVIDIA good at and where are they not able to level up? Similarly, with Intel.”

He added that until 2020, they did not approach any investors because they wanted the research to come to fruition first. Around 2021, Morphing Machines ran under the Semiconductor Fabless Accelerator Lab (SFAL) — an initiative by the Karnataka government to boost fabless semiconductor startups. The SFAL helped the company with all the required tools to start building the actual hardware for the processor.

As the startup got closer to shaping a full-fledged business, Deepak Shapeti joined the team with his years of experience as an entrepreneur and working with startups. Shapeti is currently a cofounder and CEO of Morphing Machines.

Currently working on its go-to-market initiatives and preparing for the commercial deployment of REDEFINE, Morphing Machines recently raised $2.76 Mn (around INR 23 Cr) in a seed funding round led by Indian deeptech VC firm Speciale Invest. IvyCap Ventures, Navam Capital, and CIIE Initiatives, too, participated in the round.

REDEFINE is still about a year away from being deployment-ready. The startup has started building its customer pipeline and is looking to make waves in an array of sectors and industries.

Morphing Machines’ Competitive Edge

Imperative to mention that the two decades spent by the cofounders researching and developing have not been without breakthroughs, as Narayan puts it. 

For starters, the startup’s processor, REDEFINE, is much different from the general-purpose processors. It also offers reconfigurable computing capabilities, unlike other fabless semiconductor companies that build separate SoCs for different applications and industries.   

In addition, REDEFINE, as the name suggests, redefines the relationship between software and hardware. “If you need to run a telecom application, it has its own specific software structures, which we map onto the hardware to create a seamless integration. If tomorrow you want to use the same processor for image processing, those software structures are re-mapped, creating a new integration with the same hardware, just in a different way,” the cofounder said.

She added that while the hardware is capable of doing the tasks, it’s still dumb without software. Moreover, Morphing Machines uses multiple open system software to build its own compiler, which enables this marriage between the software and the hardware, REDEFINE.

“In our case, since the hardware is specialised and has its own idiosyncrasies, we have built our own compiler, which takes the user application and compiles it into a language that the hardware understands,” said Narayan.

How Morphing Machines Is Forging Ahead?

Currently, Morphing Machines is looking at catering to domains such as AI and finance. CEO Shapeti said that in AI, the startup is particularly targeting some of the big techs working on neural processing units (NPUs) for building small-language models (SLMs). Therefore, the startup has a huge potential target customer base, including the likes of Microsoft, Qualcomm, Apple, and Intel. Further, the startup is also focussed on computational finance. 

Although the startup doesn’t have a steady source of revenue, it earns largely from the POCs it’s been conducting with its potential customers. 

To date, Morphing Machines has spent around INR 25 Cr as a bootstrap entity. Over the next two years, the company expects expenses to reach approximately INR 36 Cr. This amount is still significantly lower than what it would have cost if the company were operating outside of India.

According to Shapeti, the startup has received a significant boost from the Central government’s Design Linked Incentive (DLI) scheme. This programme has not only provided financial support but also enhanced the startup’s reputation in the market.

For now, as the startup looks to tap into the global semiconductor market, expected to surpass the $1 Tn market size by 2030, its journey from here on will be interesting to track.

[Edited by Shishir Parasher]

LEAVE A RESPONSE

Your email address will not be published. Required fields are marked *