
Sparxity closes Series A as cumulative funding reaches $90 million
Sparxity computing racks. Photo: Sparxity.Sparxity, the Dresden chip and computing-system company formerly called SpiNNcloud, has closed a Series A financing. Its November 10 announcement puts cumulative funding secured by the team at $90 million. The amount of the new Series A alone was not disclosed, so the $90 million figure is not this round's size.
The company develops sparse computing: chips and software that skip calculations and data transfers that a task does not need. In an AI model, only a small part of the network may be useful for a particular input. Hardware designed to exploit that pattern could reduce wasted work compared with running every part of the model.
ICM HPQC Fund led the financing. Named investors include Lockheed Martin Ventures, GlobalFoundries, GS Futures, Cloudberry Ventures, TGFS and the EIC Fund. Angel investors Richard Socher and Nancy Xu also participated, with continued support from Atlantic and Silicon Catalyst.
The funding will support specialised chips for fine-grained dynamic sparsity, where the parts of an AI model that are active change with the task. Sparxity is targeting faster AI inference, the stage when a trained model answers a request. Its stated target of up to 44 times faster inference concerns ultra-sparse models and is a development goal, not a general result for all AI workloads.
Sparxity reports three completed chip designs sent for fabrication, more than 70,000 chips produced, 12 deployed racks and more than 30 customers. These are company-reported operating figures. It also reports selected-workload improvements in sparse solvers, event-based machine learning and optimisation; those comparisons do not establish a universal speed or energy advantage.
The financing follows a multi-million-dollar deployment at the Jülich Supercomputing Centre and an announced collaboration with Deutsche Telekom. These earlier developments are context for the new financing, not additional rounds or proof that the next-generation inference chip is already in volume use.
DTT summary based on Sparxity / Business Wire's report. The full article is on the publisher's site.
Additional reporting
Series A closed; new round amount undisclosed. $90 million is cumulative funding secured by the team.
ICM HPQC Fund.
Develop specialised chips for fine-grained dynamic sparsity in AI inference.
More than 70,000 chips, 12 racks and more than 30 customers, according to Sparxity.