Europe Takes a Major Step Toward Tech Independence with Homegrown AI Infrastructure
Espoo, Thursday, 29 October 2026.
Nokia and Semidynamics have partnered to build homegrown European AI computing, combining specialized chips and advanced networking to challenge foreign dominance over the continent’s data.
A Strategic Alliance to Challenge Semiconductor Dependencies
On October 28–29, 2026, Finnish multinational Nokia (NYSE: NOK) and Barcelona-based semiconductor designer Semidynamics formalized a Memorandum of Understanding (MoU) to co-develop sovereign AI infrastructure for Europe [1][2]. The collaboration combines Semidynamics’ proprietary memory-centric RISC-V AI accelerator chips with Nokia’s AI-optimized data center networking fabric, aiming to provide European governments, enterprises, and research institutions with locally sourced, high-performance AI computing stacks [1][2]. By integrating Semidynamics’ inference accelerators—designed specifically for low-latency, high-bandwidth memory access—with Nokia’s liquid-cooled rack systems and Data Center Fabric solutions, the partnership seeks to deliver scalable, energy-efficient infrastructure that adheres to Europe’s strict data sovereignty and privacy mandates [1][2].
Why Europe Is Betting on In-House AI Infrastructure
The urgency behind this initiative stems from Europe’s broader strategic goal of reducing reliance on non-EU chip manufacturers, particularly those based in the United States and Asia, which currently dominate the global semiconductor supply chain [GPT]. European policymakers have increasingly emphasized the need for technological autonomy in critical sectors such as artificial intelligence, citing concerns over geopolitical vulnerabilities, export controls, and data privacy risks associated with foreign-controlled hardware [GPT]. Semidynamics’ 200+ engineers in Barcelona specialize in memory-centric AI inference silicon, a niche critical for real-time AI workloads, while Nokia brings over a decade of expertise in building high-speed, low-latency data center fabrics optimized for AI workloads [1][2]. Together, they aim to create an open ecosystem of trusted technologies that can support Europe’s AI ambitions at scale [1][2]. Roger Espasa, CEO of Semidynamics, underscored this vision: ‘Europe’s AI ambitions need European technology that can be deployed at scale’ [1][2].
Technical Integration and Commercial Roadmap
Under the MoU signed in Barcelona, the two companies will embark on a multi-phase evaluation process starting immediately [1][2]. The technical roadmap includes Semidynamics assessing Nokia’s Data Center Fabric solutions to determine compatibility with its RISC-V-based AI accelerators, with potential integration into liquid-cooled rack systems for data centers [1][2]. Matthieu Bourguignon, Nokia’s Senior Vice-President for Network Infrastructure in Europe, highlighted the broader implications: ‘Strengthening Europe’s AI capabilities means contributing technology and expertise to advance a broader, open ecosystem for trusted AI infrastructure’ [1][2]. While no specific commercial launch date has been disclosed, the collaboration signals a commitment to validating and deploying joint solutions within the next 12 to 18 months [1]. Analysts note that this timeline aligns with Europe’s Digital Decade targets for AI adoption, which call for 75% of EU enterprises to use cloud or AI services by 2030 [alert! ‘no quantified source provided for timeline alignment’].
Broader Implications for the European Tech Landscape
This partnership arrives amid intensifying global competition in AI chip development, where U.S. and Asian firms currently lead in both high-performance computing and AI accelerator design [GPT]. By leveraging open-standard RISC-V architectures—which are royalty-free and customizable—the collaboration could offer European customers a competitive alternative to proprietary solutions from NVIDIA, AMD, or Huawei [GPT]. Nokia’s role as a connectivity infrastructure leader further strengthens the initiative, as its Data Center Fabric solutions are already deployed in hyperscale data centers worldwide [2]. The MoU also reflects a growing trend among European firms to consolidate around shared technological sovereignty goals, as seen in recent EU initiatives such as the Chips Act and Gaia-X data infrastructure project [GPT]. For Semidynamics, the partnership represents a validation of its memory-centric approach, which is particularly suited to inference tasks in generative AI models [1].