The recent shift from centralised data centres to edge computing represents a significant change in industry trends and requirements. This shift comes at a critical time as India gears up for an AI upgrade to its chip infrastructure.
In an interview with AIM, Michael Hurlston, CEO at Synaptics Inc. – a leader in chipset solutions – and Rajesh Subramaniam, founder and CEO of embedUR systems – an expert in embedded software and AI modelling – outlined plans to make India a key hub for their partnership, backed by significant investments and upcoming expansions.
The companies expressed their commitment to enable Indian manufacturers and industries to develop energy-efficient and compute-efficient edge devices.
Both companies claim to have a big lead in terms of microcontrollers and microprocessors that actually run AI compared to their competitors. Hurlston emphasised their ambition to replace the whole market of traditional MCUs and MPUs with those from embedUR in order to provide the same product at more affordable prices while supporting AI capabilities.
Chennai plays an important role by providing a rich pool of engineering talent who are upskilled in the latest technology. This ensures the company operates with low attrition rates.
Driving Down Costs
Synaptics’ chipsets, coupled with embedUR’s efficient software solutions, claim to bring the cost of AI-enabled microcontrollers down to a range of $10 to $20. Each Synaptics chipset costs between $5 and $15, with the additional memory and components required for the solution adding less than $5.
In contrast, high-performance semiconductor solutions like NVIDIA’s Blackwell processors are priced at around $10,000 per chip, with complete systems costing hundreds of thousands of dollars.
This efficiency is made possible by shrinking AI models and optimising software to fit within minimal memory footprints, drastically reducing the overall bill of materials (BOM).
Moreover, many existing mid-range and lower-cost devices, such as washing machines, already include microcontrollers (MCUs). Upgrading these devices to incorporate neuro-processing units (NPUs) requires only a small incremental cost.
For instance, a washing machine that costs ₹5,000 could increase in price to around ₹6,200 while gaining significant AI capabilities.
Bengaluru and Chennai as Critical Hubs
According to Hurlston, about 25% of Synaptics’ global workforce is already based in India, primarily in Bengaluru. The city serves as the company’s engineering hub, and the company plans to double its workforce there over the next three years, increasing from 400 to 800 employees.
This expansion is driven by India’s rich pool of engineering talent, which has become increasingly vital as the global tech industry, particularly in the US, faces a shortage of skilled professionals.
Meanwhile, Tamil Nadu is preparing to become the biggest hub of this partnership. embedUR operates with low attrition rates and a strong focus on embedded software in Chennai.
Subramaniam highlighted that Tamil Nadu, with over 1,000 engineering institutions, provides a steady pipeline of skilled graduates. Over the past 13 years, the company claims to have built lasting relationships with key institutions, ensuring access to top-tier talent that can be quickly upskilled to meet the demands of cutting-edge technology.
embedUR’s expertise in connectivity and embedded software is further strengthened by its proprietary AI model framework, ModelNova. These lightweight, efficient models are designed to run on low-cost, resource-constrained devices.
Future Prospects for Indian Consumers
The leaders mentioned that the collaboration is about more than global ambitions; it also has direct implications for the Indian market.
Subramaniam said that anyone who builds products like Indian OEMs and manufacturers, such as Tata Electronics or TVS, can reach out to them. EmbedUR will help them build the product for the Indian market because it understands the use case and the applications.
He also emphasised the need for the Indian consumer industry and factories to upskill and uptool. “Our goal is that the Indian industry is able to refresh their products in terms of building energy-efficient as well as computing-efficient edge devices, internet of things (IoT) devices, or home devices.”
This action highlights the potential for India’s technology ecosystem to capitalise on local talent and resources while adopting globally competitive solutions.
AI for Everyday Devices and Mass Market
IoT has been a cornerstone of technology for over a decade, powering devices ranging from toasters to cars using embedded microcontrollers and processors.
However, the integration of AI is set to redefine IoT by enabling devices to make decisions independently without relying on cloud-based processing.
For example, future washing machines equipped with AI-enabled microcontrollers could optimise water usage, adjust programmes based on load type, or even recognise user habits to offer seamless automation, all at an affordable cost.
Furthermore, voice recognition capabilities could enhance user convenience by enabling commands like “start programme number three” or automatic adjustments for eco-friendly settings.
The collaboration between Synaptics and embedUR aims to democratise AI for mid-market and budget devices by leveraging low-power, low-cost NPUs and lightweight AI models.
This is especially useful considering how AI-enabled appliances like LG’s ThinQ washing machines, already on the market, often cater to high-end consumers, costing upwards of $3,000.
This shift is particularly significant for India, where such AI-enabled appliances, if affordable, could drive economic and societal changes, such as increasing women’s workforce participation by easing household chores.
How is Edge AI Advantageous?
As the tech industry shifts away from power-intensive, large-scale data centres dominated by companies like NVIDIA, Intel, and AMD toward efficient, localised edge AI solutions, India’s semiconductor ecosystem is positioned for significant advancements.
While powerful, traditional data centres face critical challenges, such as high energy consumption, complex cooling requirements, and the need for vast physical infrastructure.
Edge AI offers an alternative. “Now, instead of having to go back and forth to a data centre, you can actually run machine learning (ML) algorithms on a semiconductor device using software and other models,” Hurlston said.
This will not only reduce reliance on data centres but also minimise latency and energy use by allowing decisions to be made in real time at the device level.
Edge AI could redefine the way AI applications are deployed by combining hardware, software, and model development, offering India a unique opportunity to lead in this emerging domain.
The post AI Agents to Upgrade Chips in India appeared first on Analytics India Magazine.
