Euclyd

Syncthetics B.V., operating as Euclyd, is a European technology startup developing integrated circuits and systems for artificial intelligence, particularly foundation and large language models. The company is headquartered in Eindhoven, The Netherlands, with an office in San Jose, California.

Founded by Bernardo Kastrup (CEO), Gerard Egelmeers, Ingolf Held, and Harm Peters, the company is guided by a team that includes founder and advisor Atul Sinha. The leadership team has deep experience in the high-tech industry. Kastrup, who holds PhDs in both philosophy and computer engineering, previously co-founded parallel processor company Silicon Hive, which was acquired by Intel in 2011, and worked as a technology strategist at ASML. The company is backed by notable investors and advisors from the semiconductor industry, including Peter Wennink (former CEO of ASML), Federico Faggin (inventor of the microprocessor), and Steven Schuurman (founder of Elastic).

Euclyd's business model centers on designing and developing hardware that dramatically improves the efficiency of AI silicon. The company's approach focuses on three areas: an optimized neural processing unit (NPU) architecture, a scalable system-level design that leverages data locality, and a highly integrated memory hierarchy. This strategy aims to significantly lower the energy consumption, cost, and physical footprint of AI data center infrastructure.

The company's announced product is 'CRAFTWERK,' an inference architecture designed to provide a low-power, low-cost solution for agentic AI. The architecture features 'craftwerk', described as the world's first agentic AI silicon, a system-in-package that includes a massive number of processors and ultra-high-bandwidth memory. This is designed to be integrated into solutions like the 'Craftwerk Station CWS 32,' an exascale AI system for data centers.

Keywords: AI hardware, integrated circuits, neural processing unit, large language models, silicon systems, agentic AI, AI data centers, semiconductor design, inference architecture, exascale computing, processor architecture, memory hierarchy, energy-efficient computing, AI chips, foundation models, system-on-chip, VLSI design, hardware-software interface, datacenter infrastructure, sustainable AI