Exploring TCS’s Major Investment in Semiconductors: Digital Twins, Chip Design, and AI Solutions
From deploying AI, internet of things (IoT) solutions, semiconductor research and development (R&D), chip designing and testing and validating chips, India’s largest software exporter, Tata Consultancy Services (TCS), is offering end-to-end semiconductor solutions to its customers, a top company executive told Moneycontrol.
TCS is also working with the Tata Group.
Tata Group is currently gearing up to set up semiconductor manufacturing in India in partnership with Taiwan’s Powerchip Semiconductor Manufacturing Corp (PSMC), driving a whopping cumulative investment of over Rs 1.18 lakh crore across its fabrication plant in Dholera, Gujarat; and assembly, testing and packaging (ATP) plant in Morigaon, Assam.
The Dholera plant is expected to produce 50,000 wafers per month and the Morigaon plant will be able to package and produce 48 million chips per day. Wafers are circular flat plate-like silicon structures from which chips are cut out.
There is also an upcoming pilot project facility in Bengaluru, where design prototyping work will be done, which then goes into fabrication, assembling and packaging.
Semiconductor design and AI are among TCS’ key emerging areas, where it expects to scale up business. While peers such as Accenture, Infosys, HCLTech and Tech Mahindra, among others, too are betting big on semiconductor R&D and has each acquired smaller companies to strengthen their capabilities in this space, TCS has built its team and these capabilities in house over the past couple of decades.
After focusing on semiconductor R&D, the next milestone for TCS will be to perfect its chip design capabilities and solutions, said Reguraman Ayyaswamy, Senior Vice President and Global Head of the Internet of Things (IoT) and Digital Engineering, TCS, in an interview to Moneycontrol.
The chip designing process involves defining product requirements for the chip’s architecture and systems. It will also include planning the physical layout of the chip’s individual circuits and transistors.
According to Ayyaswamy, the chip design process and tools are currently going through tremendous transformation. “This is because you’re now trying to build AI not on the software, but on the chip itself,” he said.
Speaking of the on-going work with the Tata Group, Ayyaswamy said, “As part of the ER&D work, TCS is involved in System on Chip (SoC) design activities. This is an emerging area with IoT and AI on the edge. TCS will, in addition to the design activities, potentially leverage Tata Group companies in electronics that offer end-to-end value chain covering assembly & testing and foundry manufacturing in the future.”
Without naming specific companies, he added that TCS has been working with many of the major global chipmakers over the years and helping them in VLSI (very large scale integration), analog and digital chip design, and setting up technical innovation centres.
Stages of collaboration on chip manufacturing
Not just on R&D and designing of the chips, TCS will be involved in the Tata Group project from the planning stages as well, as it looks to deploy IoT solutions driven by AI capabilities that will be integrated into the manufacturing plant from the construction level itself.
Ayyaswamy explained that bringing Industry 4.0 and AI-led manufacturing at the early stages itself will reduce human intervention and also help the plant get notified on any irregularity expected in the machinery weeks prior.
Beyond the foundry processes of front end design engineering, logic design verification etc., TCS will be actively working on the assembly and testing phase by providing infrastructure for machines, test cases and validation services using simulation to check the chip’s quality.
Even as this project helps TCS build the entire suite of semiconductor-related services, it will be looking to offer the same to existing and newer external customers as well.
According to Pareekh Jain, Founder and CEO of industry insights platform EIIRTrend, similar to other ER&D services, the semiconductor business will bring higher margins, by 200-300 bps, for TCS as compared to pure play IT services business.
Jain added, “At present, out of the overall $80 billion spent every year on semiconductor R&D by global chipmakers, about $1.5-2 billion goes to outsourced R&D services. With the increased demand for chip manufacturing, outsourced semiconductor R&D spends are poised to reach $5-10 billion over the next six years while overall R&D spend by chipmakers will grow to $100 billion.”
Using Gen AI to balance skill requirements
With the sudden increase in demand for semiconductors, there has been a shortage of niche talent building core products too, Ayyaswamy said. TCS has partnered with educational institutions and tied up with partner companies like Nvidia and Qualcomm, to name a few, to help with skilling.
“The readily available talent is not high, because it’s an emerging area. Through academic collaborations, we are getting some very good quality young people. And then we give them a garage-type of setting where they can come and work like a nerd, try new LLMs (large language models) and do some high performance compute work with algorithms,” he said.
He added, “There’s also a lot more interaction going on with the partners be it Nvidia, Intel, or hyperscalars, like AWS, we have very intense interactions with them on the products that they will make, that essentially will apply in a domain context. And how we can really use that domain context etc.”
The company is interestingly also leveraging generative AI to reduce dependence of skilled talent.
“Generative AI is able to allow a chip designer to do the right level of packaging, right level of decision making. So, we see how we can use generative AI to probably notch down the skill sets that are required, which will help design those chips much faster.”
Source: https://www.moneycontrol.com/
