Photo of a Lightmatter photonic chip. [Image: Lightmatter]
Lightmatter, a developer of integrated-photonics processors, interconnects and software targeting artificial-intelligence and high-performance computing (HPC) applications, recently announced that it had pulled in another US$154 million in venture-capital funding in a Series C investment round. The new investment brought the company’s total capital raise thus far to US$270 million.
Funds from the Series C round will be used, according to Lightmatter, to “arm some of the largest cloud providers, semiconductor companies, and enterprises with the power of photonic technology” for advanced AI and HPC.
Generative AI, again
As with another recent Series C announcement, by optical-interconnects supplier Ayar Labs, Lightmatter stressed the potentially game-changing dynamics of emerging generative-AI systems and large language models such as ChatGPT. Such models, Lightmatter maintained in a press release, “are both more lucrative and resource intensive than their predecessors,” and their steep computing requirements are starting to run up against limitations tied to energy demand and cost.
Lightmatter cofounder and CEO Nicholas Harris—who was profiled as one of OPN’s 2021 “Entrepreneurs to Watch”—suggested that the rapid progress in AI “is forcing computing infrastructure to improve at an unprecedented rate,” and increasing the associated energy costs at a significant, even “planetary” scale. Both generative AI and HPC will, Harris argued, “be transformed by photonic technologies in the coming years,” as light-based integrated circuits offer the prospect of reducing heat loads, increasing energy efficiency and thereby slashing operating costs.
Chips, interconnects, software
Lightmatter cofounder and CEO Nicholas Harris argued that both generative AI and HPC will “be transformed by photonic technologies in the coming years.”
To help engineer that photonic transformation, Lightmatter offers three products: Envise, a photonic AI accelerator chip; Passage, a wafer-scale programmable photonic-interconnect system; and Idiom, a software layer that allows the company’s hardware to talk to standard deep-learning frameworks. The result, Lightmatter says, constitutes “a full stack of hardware and software solutions to realize the benefits of photonic compute and interconnect technologies.”
In the past year, the company adds, it has also increased its management depth and expertise in areas ranging from software and hardware engineering to financial and product management. Lightmatter now “has over 20 active roles across product, R&D, and engineering, and holds over 150 patents worldwide,” according to the company.
Funding product delivery
Proceeds from the new US$154 million Series C round will, according to Lightmatter, be used to “fund the delivery of [the company’s] products to customers.” Participants in the Series C investment include the cross-border venture-capital firm SIP Global Partners, Fidelity Management & Research Company, Viking Global Investors, Google Ventures (GV), HPE Pathfinder and existing Lightmatter investors.
In the company’s press release, Jeffrey Smith, general partner at SIP Global, argued that “Lightmatter’s unique approach to harnessing the power of photonics in hardware trips will further the initial capabilities and use cases that we’re seeing from generative AI.” And GV general partner Erik Nordlander expressed the view that Lightmatter “is taking a differentiated approach” to photonics in AI workloads “by using silicon photonics and bringing together a deeply technical team to further its mission.”