Sparxity closed a Series A to develop its next AI chip, taking cumulative funding to $90 million, the Dresden company announced on November 10. The financing arrives ahead of a planned 2028 deployment with Deutsche Telekom’s T-Systems.
The company, formerly SpiNNcloud, builds processors, software and rack systems for research centers and AI infrastructure operators. Its designs skip calculations and data transfers that a model does not need, aiming to reduce the power and cost of running AI.
ICM HPQC Fund led the round, with Lockheed Martin Ventures and chipmaker GlobalFoundries among the investors. Sparxity did not disclose the Series A’s standalone amount.
Sparxity says it has produced more than 70,000 chips, deployed 12 racks and signed more than 30 customers. At the Jülich Supercomputing Centre, its two-rack JUBIS system forms part of a research program covering AI, brain simulations and quantum-computer emulation.
The next chip is still under development, with Sparxity targeting up to 44 times faster inference on ultra-sparse models. CEO Hector Gonzalez told Tech Funding News that the financing would also pay for manufacturing to fill existing orders.
Under a letter of intent announced on November 3, the partners will first use Sparxity’s current hardware to work out integration of the next chip generation, Deutsche Telekom says. Deployment is planned for 2028, with collaboration details to be agreed in the coming months.