Sparxity, a German semiconductor and AI infrastructure company developing sparse-native computing systems, has closed its Series A financing, bringing its cumulative funding to $90 million. The investment will support the development of specialized chips designed to improve the efficiency of AI inference by concentrating computing resources on the operations that are needed. The round was led by ICM HPQC Fund, with participation from Lockheed Martin Ventures, GlobalFoundries, GS Futures, Cloudberry Ventures, TGFS, and the European Innovation Council Fund.
Individual investors included Richard Socher of Recursive Superintelligence and Nancy Xu, whose previous company Moonhub was acquired by Salesforce.
Existing investors Atlantic and Silicon Catalyst also participated.
Formerly known as SpiNNcloud, Sparxity is headquartered in Dresden, Germany, and develops full-stack computing systems built around sparse processing.
Sparse computing reduces unnecessary mathematical operations by activating only the parts of a model or computing system required for a particular task.
The approach contrasts with dense computing architectures that can perform large numbers of operations even when portions of the computational capacity are not necessary for a given workload.
Sparxity believes specialized hardware can improve the efficiency of AI models designed around sparse activation.
Its next-generation chip technology is being developed for fine-grained dynamic sparsity in AI inference.
The company is targeting performance improvements of up to 44 times faster inference for certain ultra-sparse models.
Those projections depend on the characteristics of the models and workloads involved.
Sparxity’s existing systems support high-performance computing and AI applications.
The company reports achieving tenfold acceleration for certain sparse solver workloads.
It has also reported 18 times lower energy consumption for selected event-based machine learning applications and 20 times shorter execution times for certain optimization problems.
These figures reflect specific workloads rather than universal performance improvements across all computing tasks.
Sparxity has completed three semiconductor tapeouts and manufactured more than 70,000 chips.
It has deployed 12 computing racks and serves more than 30 customers worldwide.
Its customers include major supercomputing facilities.
The company also reported that revenue quadrupled between 2024 and 2025.
The latest financing follows several commercial and technical milestones.
Sparxity announced a multimillion-dollar deployment of its computing systems at the Jülich Supercomputing Centre in Germany.
The center is associated with Forschungszentrum Jülich and operates advanced high-performance computing infrastructure, including Europe’s first exascale supercomputer.
The company also established a partnership with Deutsche Telekom to integrate sparse-native computing technology into the data center infrastructure of its T-Systems subsidiary.
The collaboration is intended to support deployment of energy-efficient AI computing at industrial scale.
Sparxity sees demand for specialized architectures increasing as AI infrastructure operators seek ways to reduce power consumption and improve utilization.
Training and inference systems frequently require substantial computing resources, creating pressure to improve the performance delivered per unit of electricity.
Sparse architectures offer one potential approach by reducing redundant operations and data movement.
The company is developing hardware and software intended to support these architectures across a range of AI workloads.
Its strategy involves combining custom semiconductor designs with systems capable of operating at rack and supercomputer scale.
The new funding will accelerate development of its next-generation inference chips while supporting additional customer deployments.
By investing in specialized computing architectures, Sparxity aims to provide an alternative to relying exclusively on conventional dense processors for increasingly complex AI workloads.
KEY QUOTES:
“Our work with PEER demonstrated the potential of dynamic sparsity, activating only a fraction of the compute across millions of fine-grained experts. Yet today’s hardware limits how far we can take these approaches. Co-designing AI models, distillation pipelines and hardware around sparsity unlock a new level of computing efficiency.”
Xu Owen He, Strategic Advisor To Sparxity And Former Google DeepMind Researcher
“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. 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.”
Hector Gonzalez, Co-Founder And CEO Of Sparxity
“Sparxity has a unique complementary positioning at the infrastructure stack. Its architecture is grounded in first principles, addressing the inevitable shift toward sparse activation and the structural limitations of conventional hardware. 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.”
Matthew Gould, ICM HPQC Fund
“We are looking forward to having Sparxity on board in our AI factory soon. The German company scores particularly well with its energy-efficient use of AI computing power.”
Ferri Abolhassan, CEO Of T-Systems
