Sparxity

๐Ÿ“‚ Hardware & Deep Tech๐Ÿ“‚ Artificial Intelligence๐Ÿ“ Dresden๐Ÿ—“๏ธ Founded: 2021

Dresden company, formerly SpiNNcloud, developing hardware and software for sparse AI computing to cut operations and energy use.

Europe's most interesting startup stories. Every week.

No spam. Unsubscribe anytime.

Podcast Episode

๐ŸŽ™๏ธ Deep Dive

Sumary, Sparxity, PORT โ€” $25M for AI accounting, $90M total for sparse chips and โ‚ฌ14.5M for autonomous logistics โ€” Daily News November 10, 2026

Listen to the episode

About Sparxity

Sparxity was founded in 2021 as SpiNNcloud, a TU Dresden spinout built on technology developed in the EU Human Brain Project; it is led by co-founders Hector Gonzalez (CEO) and Matthias Lohrmann (CTO), with Christian Georg Mayr. It builds computing systems designed for sparsity, processing only the active parts of models, and is developing chips for dynamic sparsity. It reports 12 racks deployed, more than 70,000 chips produced, three tapeouts and more than 30 customers, including a system at the Jรผlich Supercomputing Centre and a partnership with Deutsche Telekomโ€™s T-Systems. It has received EIC funding, including โ‚ฌ10 million from the EIC Accelerator in 2025. On 10 November 2026 it closed a Series A led by ICM HPQC Fund, with Lockheed Martin Ventures, GlobalFoundries, GS Futures, Cloudberry Ventures, TGFS and the EIC Fund, bringing total funding to $90 million; the round amount alone was not disclosed.

The Story

TU Dresden spinout founded in 2021 as SpiNNcloud, based on technology developed in the Human Brain Project.

How Sparxity works

Business Model

B2B sales and deployment of compute systems.

Products & Services

Key Products

  • Sparse-native compute systems
  • Dynamic-sparsity chips (in development)

Core Use Cases

  • AI inference
  • High-performance computing
  • Optimisation

Market & Clients

Key Customers

Supercomputing centres, data centres and enterprises; more than 30 customers reported.

Geographic Presence

DE

How Sparxity competes

Competitive Advantages

  • Hardware-software co-design
  • Systems already deployed at supercomputing centres

How Sparxity grows

Growth Strategy

Move sparse chips and systems from supercomputing centres into industrial AI infrastructure, including with T-Systems.

Distribution Model

Direct sales and infrastructure partnerships.

Regulatory Context

Export controls and European semiconductor policy.

Key Risks

  • Non-generalisable benchmarks
  • Long semiconductor cycles
  • Dependence on manufacturing and supply chain

Strategic Insights

  • The advantage depends on co-designing models and hardware.
  • Benchmarks must be compared on equivalent workloads.

Funding & Investors

Total Funding

90000000

Funding Rounds

  • 2026 - Series A: Non divulgato (totale raccolto $90M)
    Lead: ICM HPQC Fund

Key Investors

ICM HPQC FundLockheed Martin VenturesGlobalFoundriesGS FuturesCloudberry VenturesTGFSEIC Fund

Founding Team

Founders

Hector GonzalezCo-founder & CEO
Matthias LohrmannCo-founder & CTO
Christian Georg MayrCo-founder

Key Metrics

Capitale raccolto totale$90M (2026)
Rack installati12 (2026)
Chip prodottioltre 70.000 (2026)
Clientioltre 30 (2026)

Lessons from Sparxity

  • In deep tech, real deployments and tapeouts matter more than performance claims.

Similar Startups

๐Ÿ“ Munich

Helsing

Munich company founded in 2021 building autonomous systems and AI software for defence, from the HX-2 drone to the SG-1 underwater system. $1.8B Series E in 07/2026, at $18B.

Read Analysis
๐Ÿ“ Munich

RobCo

Munich company founded in 2020 out of TUM research offering manufacturers modular industrial robots with software and a monthly fee. Valued above $1B in the October 2026 transaction.

Read Analysis
๐Ÿ“ Berlin

Sereact

German startup founded in 2021 in Stuttgart. Develops AI enabling industrial robots to receive natural language instructions and adapt to new objects without reprogramming. Clients: BMW, Amazon.

Read Analysis

Sources

๐ŸŽ™๏ธ Scalable Podcast โ€” European startup stories ยท ๐Ÿ‡ฎ๐Ÿ‡น in Italian
Spotify ๐ŸŽง Apple