Dresden-based AI chipmaker Sparxity has closed its Series A funding round, lifting its total capital raised to $90M to fund hardware built for dynamic sparsity.
ICM HPQC Fund led the Series A round.
Backers and Funding Partners
Additional participants in the round included Lockheed Martin Ventures, GlobalFoundries, GS Futures, Cloudberry Ventures, TGFS, and the EIC Fund. Angel investors Richard Socher, founder of Recursive Superintelligence, and Nancy Xu, whose start-up Moonhub was bought by Salesforce, also joined the round.
Sparxity did not disclose the exact amount raised specifically in the Series A.
“Sparxity has demonstrated the ability to manufacture its technology and deploy systems with major customers,” Matthew Gould of ICM HPQC Fund said.
Targeting Dynamic Sparsity
The company will channel the new capital into developing chips designed for dynamic sparsity. Many artificial intelligence models utilize only a small fraction of their network for any given task.
Sparxity’s chips skip the unused parts of the network to save time and energy. The firm aims for AI inference—the operational stage where a trained model answers requests—to run up to 44 times faster in highly sparse models.
Hector Gonzalez, Sparxity’s co-founder and chief executive, stated that the fresh funding will assist data centers in transitioning toward hardware built specifically for sparse AI workloads.
Origins at TU Dresden
The company was founded in 2021 as SpiNNcloud, operating as a spin-off from TU Dresden. Its core architecture was built on SpiNNaker2, a chip inspired by biological brain functions.
Last month, the enterprise changed its name to Sparxity to reflect its specialized focus on sparse computing. To date, Sparxity has manufactured more than 70,000 chips across three distinct designs and installed 12 system racks.
The company currently counts more than 30 customers, including prominent supercomputing centers globally. Its revenue quadrupled between 2024 and 2025.
Supercomputer Deployments and T-Systems
Existing systems handle both scientific computing and artificial intelligence workloads. On selected tasks, the company reports its systems run calculations 10 times faster while utilizing 18 times less energy for specific machine learning operations.
The funding follows two major commercial developments. In October, Sparxity installed a multi-million-dollar system at the Jülich Supercomputing Centre in Germany, home to JUPITER, which is Europe’s first exascale supercomputer.
Last week, Sparxity signed a letter of intent with T-Systems, the business arm of Deutsche Telekom, to integrate its technology into T-Systems data centers.
“We are looking forward to having Sparxity on board in our AI factory soon,” Ferri Abolhassan, chief executive of T-Systems, said.
Sparxity also secured European Union backing in 2025 through the European Innovation Council’s accelerator programme, receiving €10m in grants and equity.
