Sparxity, a Dresden‑based developer of "sparse‑native" AI hardware and racks, announced it has closed a Series A that brings cumulative funding secured by the team to $90 million. The company said the round was led by ICM HPQC Fund and included strategic investors such as Lockheed Martin Ventures, GlobalFoundries and the EIC Fund. The Business Wire release that carried the announcement also pointed to a separate cooperation with Deutsche Telekom's T‑Systems to integrate and validate Sparxity hardware in the carrier's AI datacentre infrastructure.

Deal and funding

The company — formerly known as SpiNNcloud — described the Series A as accelerating development of chips and systems designed for fine‑grained dynamic sparsity, an approach that aims to reduce the compute and energy required for many AI inference workloads. Sparxity and its press material cited technical benefits from earlier deployments, including claims of 10× faster sparse solver execution and up to 18× lower energy for certain event‑based machine‑learning workloads.

Alongside the funding announcement, Sparxity said it has a Letter of Intent or partnership with T‑Systems/Deutsche Telekom to work on integrating its hardware into the operator's so‑called AI factory in Munich, with initial deployment work targeted in 2028. T‑Systems' executives were quoted as saying they looked forward to having Sparxity "on board in our AI factory soon." The cooperation is presented as part of a broader push to build energy‑efficient, sovereign AI infrastructure in Germany and Europe.

Technical claims and next steps

Sparxity has already published examples of rack‑scale deployments: the company described a multi‑million‑dollar installation at the Jülich Supercomputing Centre and said its systems are intended for HPC and sparse AI inference workloads. Sparxity's site and the Business Wire material state the firm has produced multiple tapeouts and deployed racks in Europe and the US, positioning the business as a full‑stack vendor from silicon to software.

What this means for the market

For Europe’s automotive and supplier industry the announcement underlines two linked trends. First, demand for specialised AI compute is rising as automakers and Tier‑1 suppliers trial larger, more capable ADAS, autonomy and in‑car AI applications that require low‑latency inference and significant local processing. Second, OEMs and fleets are showing interest in reducing operational energy and cost for AI workloads and in securing European compute supply chains rather than relying solely on US or Asian hyperscalers.

Concretely, a local compute supplier that claims big energy and throughput gains could help vehicle manufacturers run more of their development and validation workloads on European infrastructure, or support on‑premise model serving for production systems such as advanced perception stacks or cockpit LLMs. The T‑Systems partnership signals that telcos and systems integrators are positioning datacentre capacity and services to host more automotive AI workloads — an option that may appeal to OEMs seeking data sovereignty and lower latency for simulation, map‑building or fleet telematics.

That said, the technology Sparxity promotes — sparse‑native chips and systems — will need independent benchmarks and production‑grade reliability before major automakers shift critical workloads onto it. Automotive use cases impose long product‑lifecycle, functional‑safety and certification demands that are not identical to HPC or cloud inference workloads. Integrations with operators such as T‑Systems, and deployments at research centres like Jülich, are logical early steps but not a short‑cut to automotive qualification.

WorldTrader24 B2B vehicle marketplace for dealers →

Sources: Business Wire, TelecomTV

Partner note: EastAutoNews technology partner DaVinci AI develops AI solutions for the entire automotive industry – from market and dealer analytics to autonomous AI agents and image and video production.