AttoTude Secures $50M in Series B, Boosting THz Interconnect Tech for AI and Data Centers
April 3, 2025
Dave Welch, co-founder and CEO of AttoTude, emphasized that their technology represents a paradigm shift necessary to meet the demands of modern AI workloads.
This Series B funding will also facilitate the scaling of manufacturing operations, expanding customer engagements, and accelerating product development as AttoTude prepares for volume production.
AttoTude, a company specializing in terahertz (THz) interconnect technology for artificial intelligence (AI) and hyperscale data centers, has successfully raised $50 million in its Series B funding round.
Currently, AttoTude is collaborating with major hyperscale and AI infrastructure companies for the initial deployment of its technology.
Navin Chaddha, Managing Partner at Mayfield, highlighted the urgent need for innovative interconnect solutions as data center architectures evolve in response to advancements in AI.
This latest funding round brings AttoTude's total capital raised to $91 million, following a $30 million Series A and an $11 million seed round in 2024, showcasing significant investor confidence in the company's technology.
The funds will likely be allocated to further develop AttoTude's THz interconnect solutions, which aim to enhance data transfer speeds and efficiencies in advanced computing environments.
AttoTude's growth trajectory reflects a broader trend of investment in innovative technologies that support the evolving needs of data centers and AI systems.
The company's technology is positioned to meet the increasing demands of AI applications and large-scale data processing, which require high-bandwidth and low-latency interconnections.
AttoTude's THz interconnect technology significantly enhances bandwidth, reliability, and energy efficiency, addressing challenges faced by conventional interconnect methods.
The platform allows for communication over distances of up to 40 meters without relying on photonics, utilizing standard ASIC manufacturing processes instead.
Summary based on 2 sources