Sparxity's $90M Bet on Sparse AI Could Change Everything

Jan de Vries · 2026-11-10

Sparxity, the German startup formerly known as SpiNNcloud, just closed a Series A round that brings its total funding to $90 million. That's a lot of cash for a company betting on a very specific idea: that AI doesn't need to compute everything to be smart. The round was led by ICM HPQC Fund, with participation from heavy hitters like Lockheed Martin Ventures, GlobalFoundries, and the EIC Fund. Angel investors Richard Socher and Nancy Xu also joined in. So what's the big deal? Sparxity is building what it calls "sparse-native" computing. In plain English: instead of crunching every single number in a neural network, their hardware only activates the tiny fraction that actually matters for a given task. It's like turning on only the lights you need in a huge office building instead of every single one. The result? Less energy, faster processing, and lower costs. ### Why Sparsity Matters Now The timing makes sense. AI models are getting bigger and hungrier for power, and data centers are feeling the squeeze. Sparxity's CEO Hector Gonzalez puts it well: "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." That's the pitch. And investors are buying it. Sparxity's systems are already deployed in rack- and supercomputer-scale setups across Europe and the US. The company reports some impressive numbers: 10x faster sparse solvers, 18x lower energy use for event-based machine learning, and 20x shorter runtimes for optimization problems on select workloads. Plus, they've quadrupled revenue between 2024 and 2025 and have over 30 customers worldwide. ### The Bigger Picture: Europe's AI Hardware Push Sparxity isn't alone. 2026 has seen a wave of European investments in specialized AI hardware. The UK's OLIX raised $305 million for AI inference chips, Spain's Openchip secured $130 million for AI and HPC processors, and the Dutch startup EUCLYD announced a Series A exceeding $225 million for energy-efficient AI systems. Germany's QuantumDiamonds reported a $103 million financing package for semiconductor inspection tech. Add it all up, and you've got more than $870 million in disclosed financing across these deals—and that's before you count Sparxity's cumulative total. This isn't just about money. It's about building a European alternative in the AI hardware stack, which has long been dominated by US and Asian players. The EU Inc proposal and similar initiatives are making it easier for startups to scale across borders, and investors are taking notice. ### What's Next for Sparxity? The company is already working on its next-generation chip, designed specifically for dynamic sparsity in AI inference. And they've got the foundation: three tapeouts, over 70,000 chips produced, and 12 racks deployed. With this fresh capital, they plan to accelerate the transition from dense, energy-intensive architectures to heterogeneous data centers with dedicated silicon for ultra-sparse workloads. As Matthew Gould of ICM HPQC Fund said, "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." In other words, they're not just building faster chips—they're rethinking how computing should work in the age of AI. And if they're right, the next wave of AI innovation might be a lot more efficient than the last one.