Tamalika Banerjee: Pioneering energy-efficient AI
“AI is a wicked problem. It can’t be addressed by one person or discipline. Interdisciplinary collaboration isn't optional; it’s essential.”
Published on October 31, 2025
Tamalika Banerjee, © CogniGron
As editor-in-chief, Aafke oversees all content and events but loves writing herself. She makes complex topics accessible and tells the stories behind technology.
AI is transforming industries at breakneck speed, yet its rapid expansion comes with a huge cost: energy consumption. One of the speakers at the first edition of Watt Matters in AI, Tamalika Banerjee, is confronting this challenge head-on. A physicist, tech innovator, and professor at the University of Groningen (RuG), Banerjee is the driving force behind IMChip, a startup developing neuromorphic computing hardware that could make AI more energy-efficient and sustainable.
Banerjee’s journey into AI and neuromorphic computing is rooted in a deep-seated concern for sustainability. “The way I see it is that we borrow this planet from our children. We must use technology responsibly so it benefits society rather than harming it,” she explains. Over the past 15 years, her work has evolved from fundamental physics to the forefront of energy-efficient computing, blending science, engineering, and societal responsibility.
Brain-inspired technology
IMChip, a Groningen-based startup, is developing computer chips inspired by the human brain. Traditional silicon-based processors are limited by their binary logic, constrained to ones and zeros. The human brain, by contrast, operates on a continuum, allowing far more nuanced computations with far less energy. IMChip’s technology bridges this gap.
Unlike traditional processors that consume vast amounts of energy to process and store data separately, IMChip’s neuromorphic chips combine these functions in one place, just like neurons and synapses do. This brain-inspired design enables far more efficient computing, dramatically reducing energy use. The result: chips that are up to 200 times more energy-efficient than conventional processors. Banerjee: “Neuromorphic computing is a game-changer. We can get more processing power for less energy, which is critical as AI scales.”
An interdisciplinary approach
The startup’s beginnings were unexpected. It all started at the Groningen Cognitive Systems and Materials Center (CogniGron), which promotes sustainable energy use and responsible resource management. “CogniGron matches people from materials science, AI, and mathematics in a truly interdisciplinary environment. Our group reported on the surprising enhancement of performance in memristors [key components in neuromorphic computing, red.] with area downscaling, and it caught the attention of the deep-tech market,” Banerjee recalls. “I hesitated at first, but we’ve been developing the science behind the company since 2018.”
Watt Matters in AI
Watt Matters in AI Conference, 26 November, Eindhoven, The Netherlands. Tickets via wattmattersinai.eu.
Going from lab to market: anything but straightforward
The path from lab to market has been anything but straightforward. IMChip received its first pre-seed funding in April 2025, enabling three core team members to begin full-time work. Banerjee acknowledges the challenges: “It’s a deeptech startup, not just software. You need infrastructure, hardware, and interdisciplinary know-how. Scaling is difficult, especially in Europe, where the risk appetite for investors is lower than in the United States.”
Yet the entrepreneur remains optimistic. IMChip has already demonstrated prototype devices that mimic synapses and neurons, a milestone that moves the company closer to its goal of delivering its first chip by 2028. Looking further ahead, the company aims to revolutionize the semiconductor landscape, creating hardware that not only supports AI development but does so in a sustainable and responsible manner.
‘AI is a wicked problem’
Banjeree’s vision is clear: AI development should not rely solely on software or isolated innovations. “AI is a wicked problem. It can’t be addressed by one person or discipline alone,” she says. “We need cohesion across science, engineering, and entrepreneurship. Interdisciplinary collaboration is not optional; it’s essential.” By reducing energy consumption and creating hardware designed for sustainability, IMChip addresses both environmental concerns and broader societal responsibilities.
Her work also highlights the importance of translating fundamental science into tangible impact. “Without science, there is no technology. But research must deliver outcomes that are relevant and beneficial to society,” she notes. This philosophy is at the heart of IMChip’s mission: to ensure that advances in AI benefit humanity without compromising the planet.
Keynote at Watt Matters in AI
At the Watt Matters in AI event, Banerjee wants to help the audience gain insight into how advanced materials, physics, and computing can converge to create AI technologies that are powerful, responsible, and sustainable.
For Banerjee, the goal is as much about inspiration as innovation. “We borrow the world from our children; we don’t inherit it from our ancestors. AI can help save the planet if we approach it responsibly,” she says. In an era where AI often outpaces our understanding of its societal impact, Banerjee’s work offers a compelling example of how science, technology, and ethics can intersect to shape a better future.
By championing interdisciplinary collaboration and energy-efficient AI, Tamalika Banerjee and IMChip are demonstrating that innovation and sustainability are not mutually exclusive—they are inseparable. Her work sets a new standard for what it means to develop AI for the benefit of both people and the planet.