News & Analysis

MIT alumni looking to reinvent how chips communicate

4th March 2024
Paige West
0

Lightmatter, a company founded by three MIT alumni, is reimagining the very essence of the microchip, transcending conventional reliance on electrical signals by integrating light, or photons, for data processing and communication.

The company's debut offerings, the Envise chip and Passage interconnect, stand as testaments to this innovative dual approach, merging photonics and electronics to enhance operational efficiency, particularly in AI applications.

Nicholas Harris, Co-Founder, and CEO of Lightmatter, elucidates: "The two problems we are solving are 'How do chips talk?' and 'How do you do these [AI] calculations?'" Through Envise and Passage, Lightmatter aspires to address these queries head-on, offering a glimpse into the future of high-efficiency computing.

In recognition of the critical role AI plays in contemporary and future computing landscapes, Lightmatter has secured over $300 million in funding in 2023, reaching a valuation of $1.2 billion. This financial endorsement underscores the company's potential to mitigate the burgeoning energy demands of data centres and AI algorithms by collaborating with leading technology firms worldwide.

Harris envisions a transformative ecosystem enabled by Lightmatter's interconnect technology, facilitating the integration of hundreds of thousands of avant-garde compute units. "That simply wouldn’t be possible without the technology that we’re building," he asserts.

Harris's journey from Micron Technology to MIT and subsequently to the helm of Lightmatter reflects a deep-seated desire to transcend the limitations of traditional computing paradigms. His explorations into quantum computing and photonics during his PhD under Dirk Englund at MIT, culminating in the development of silicon-based integrated photonic chips, paved the way for Lightmatter's inception. This technological foundation, initially intended for photonic quantum computing, found a new purpose in accelerating deep learning processes – a realisation that emerged amid the burgeoning interest in AI at MIT.

Lightmatter's Envise chip exemplifies this synergy between electronics and photonics, leveraging light's unique properties to perform concurrent calculations across multiple wavelengths. This capability significantly boosts operations per area while enhancing energy efficiency. Similarly, the Passage interconnect capitalises on light's superior latency and bandwidth to facilitate seamless communication between processors, akin to the role of fibre optic cables in long-distance data transmission.

The implications of Lightmatter's innovations extend far beyond technical efficiencies; they represent a crucial step towards sustainable computing. As Harris highlights, the projected energy consumption of data centres and AI systems by 2040 could account for as much as 80% of global energy usage. Lightmatter's technologies aim to curtail this trajectory, offering a more energy-efficient computing paradigm that could alleviate the environmental and economic pressures of escalating power demands.

As Lightmatter forges ahead, its mission extends to reimagining the entirety of computing infrastructure through the prism of photonics. With Envise and Passage marking the beginning, the company is poised to redefine computing, anchoring the future of the industry in the transformative power of light. This journey not only heralds a new era for computing but also underscores the pivotal role of innovation in addressing the intertwined challenges of technological advancement and sustainability.

Featured products

Upcoming Events

View all events
Newsletter
Latest global electronics news
© Copyright 2024 Electronic Specifier