%201.webp)
Sparxity has closed a Series A funding round, bringing the cumulative funding secured by its team to USD $90 million. The round was led by ICM HPQC Fund.
The Dresden-based company will use the financing to develop specialised chips for fine-grained dynamic sparsity in artificial intelligence inference. In AI, sparsity means activating only part of a model's compute resources for a given task rather than using all parameters at once.
Investors included Lockheed Martin Ventures, GlobalFoundries, GS Futures, Cloudberry Ventures, TGFS and the EIC Fund. The round also drew angel backing from Richard Socher and Nancy Xu, while existing supporters Atlantic and Silicon Catalyst participated again.
Sparxity aims to build systems that reduce unnecessary computation and data movement in AI and high-performance computing workloads. With its next chip designs, it is targeting up to 44 times faster AI inference in ultra-sparse models.
The company already has commercial deployments. It has completed three tapeouts, produced more than 70,000 chips and deployed 12 racks. It now serves more than 30 customers worldwide, including several large supercomputing facilities.
Sparxity said revenue quadrupled between 2024 and 2025, with momentum continuing in 2026. It did not disclose revenue figures or the size of the Series A round.
Chip design
The fresh capital comes as chipmakers and AI infrastructure groups look for ways to cut the cost and energy demands of running models. Much of the AI hardware market has been built around dense computing architectures, in which all or most of a model is engaged for each query. Sparxity is positioning its technology around sparse workloads, where only the most relevant parts of a model are activated.
Its current systems are used in high-performance computing and AI settings. The company reported selected results including 10 times faster sparse solvers, 18 times lower energy use for event-based machine learning and 20 times shorter runtimes for optimisation problems on certain workloads.
Its next-generation chip technology is being designed specifically for dynamic sparsity in AI inference. Sparxity argues this could allow model developers to build finer-grained sparse architectures while giving operators greater throughput and lower infrastructure and energy costs.
"Our work with PEER demonstrated the potential of dynamic sparsity, activating only a fraction of the compute across millions of fine-grained experts," said Xu Owen He, Former Google DeepMind Researcher and Sparxity Strategic Advisor.
"Yet today's hardware limits how far we can take these approaches. Co-designing AI models, distillation pipelines and hardware around sparsity unlocks a new level of computing efficiency," He said.
Investor backing
The investor syndicate combines financial backers with strategic names from the semiconductor and industrial technology sectors. GlobalFoundries' involvement is notable given Sparxity's focus on chip development and manufacturing scale-up, while Lockheed Martin Ventures adds another strategic investor with an interest in advanced computing infrastructure.
Matthew Gould of ICM HPQC Fund pointed to the company's manufacturing and deployment record as part of the investment case.
"Sparxity has a unique complementary positioning in the infrastructure stack. Its architecture is grounded in first principles, addressing the inevitable shift toward sparse activation and the structural limitations of conventional hardware," said Matthew Gould of ICM HPQC Fund.
"Sparxity has demonstrated the ability to manufacture its technology and deploy systems with major customers. We are pleased to lead this financing as the team builds on that track record to address next-generation AI inference," Gould said.
Commercial progress
The funding follows two commercial developments. One was a multi-million-dollar deployment of Sparxity systems at the Jülich Supercomputing Centre, part of Forschungszentrum Jülich and home to Europe's first exascale supercomputer.
The other was a partnership with Deutsche Telekom to integrate Sparxity's technology into T-Systems data centre infrastructure. The agreement is intended to bring sparse-native AI computing into industrial-scale environments.
Sparxity was previously known as SpiNNcloud and recently adopted its new name as it broadened its focus on sparse computing. The rebranding reflects a shift from its earlier positioning toward a wider role in AI and high-performance computing infrastructure.
Chief Executive Officer Hector Gonzalez said the company sees sparse architectures as a response to rising demand for more efficient AI computing.
"The next era of AI will be defined not just by how much computing power we can deliver, but by how intelligently and efficiently we use it," said Hector Gonzalez, Co-Founder and Chief Executive Officer of Sparxity.
"This financing is a powerful validation of our vision for sparse-native computing and the complementary positioning of our systems. With the backing of leading strategic investors, we're accelerating the transition from dense, energy-intensive architectures to heterogeneous datacenters with dedicated silicon for ultra-sparse workloads to unlock greater AI performance and lower costs," Gonzalez said.