Build sovereign AI infrastructure for Europe
Traditional IT is hitting a wall.
Businesses are forced to choose between the performance required to stay competitive and an unsustainable reality. Energy costs are rising, European regulations are tightening, and data privacy mandates remain strict. Traditional IT architectures can no longer support this duality—but the next generation of computing can.
Businesses are currently trapped by three challenges:
Exploding AI compute demand
Soaring energy costs
Sovereignty & Compliance Gap
AI is transforming the requirements for modern data centers, where both energy constraints and performance demands have become the primary limiting factors.
With the Fujitsu MONAKA Server, European organizations of every size can run AI inference efficiently on CPU-based systems - leveraging a platform born from supercomputing, with sovereign models designed, developed, and manufactured in Japan.
Christian Leutner
Senior Vice President and Head of Fsas Technologies Europe
The Fujitsu MONAKA Servers redefine the standard for modern data centers. At its heart lies a high-performance, 2nm Arm-based CPU. The CPU is engineered specifically to handle the intensive demands of AI inference, High-Performance Computing (HPC), and large-scale data analytics.
Our 3D many-core architecture draws directly from the legendary supercomputing heritage of the K computer and Fugaku. It delivers unprecedented compute density in a remarkably space-saving design. This is performance built for the next generation of data-intensive workloads.
Designed for the Real World, Not Just the Lab:
Infrastructure-Ready: Optimized for air-cooled data centers, helping avoid the cost, downtime, and complexity associated with liquid-cooling retrofits.
Sovereign & Transparent: Our "Made in Japan" manufacturing process provides complete supply chain traceability. This ensures the hardware-level data integrity and regional control that highly regulated industries demand.
System-Level Optimization: By combining our 2nm processor technology with advanced resource-pooling and intelligent thermal management, we deliver a holistic server ecosystem that drives down your Total Cost of Ownership through significant improvements in power efficiency, rack density, and operational simplicity.
High-Performance Simulation
Accelerate engineering cycles with supercomputing heritage; slash simulation times by replacing power-heavy clusters with high-efficiency surrogate modeling.
Scalable Database Platforms
Maximize throughput for RDB/NoSQL workloads while significantly reducing your server footprint and total energy draw.
Optimized Software
Accelerate CI/CD build pipelines and real-time player analytics with a high-throughput, Arm-native development environment.
Secure Transaction Processing
Protect sensitive financial data in use with hardware-level Confidential Computing (Arm CCA) and mission-critical, "always-on" reliability.
Real-time Forecasting
Scale AI demand and inventory forecasting across your entire catalog without the cost-prohibitive constraints of GPU-heavy procurement.
Responsive Supply Chains
Deploy real-time analytics directly into existing logistics hubs with air-cooled infrastructure that lowers latency and operational costs.
Rapid Automotive CAE
Shorten time-to-market by pairing traditional structural simulations with AI-based surrogate modeling on standard, high-efficiency server hardware.
Digital Twin
Maintain high-fidelity digital replicas of physical sites by processing massive, real-time spatial and sensor data streams with ease.
Max operating frequency (GHz)
Memory transfer rate (MT/s)
AI inference throughput, half the power consumption
Air-cooling support
3D-Many Core Architecture: 2nm process for logic, 5nm for SRAM/IO. Stacking reduces data travel, latency, and power loss.
Performance Metrics: Max operating frequency 3.8 GHz, Memory transfer rate 8800 MT/s.
Security: Hardware-level Confidential Computing (Arm CCA) ensures data remains protected in-memory.
Double-Impact: Deliver 2x the AI inference throughput at half the server count and power consumption.
1U Server: Optimized for AI agents and inference; supports CXL/CDI for resource pooling; 40°C air-cooling support.
2U Multi-Node Server: 4 nodes/chassis, 2 CPUs per node. Engineered for HPC, surrogate modeling, and high-precision workloads.
What is Fujitsu MONAKA Server?
Fujitsu MONAKA Server is an Arm-based server designed for AI inference, HPC and data-intensive workloads. It is powered by the 2nm FUJITSU-MONAKA CPU.
Can AI inference run efficiently on CPUs?
Yes. CPU-based servers can support AI inference when processor architecture, memory performance and software are optimized for the workload. FUJITSU-MONAKA is designed for AI inference and combines a 3D many-core architecture with high memory transfer rates.
How can organizations reduce AI infrastructure power consumption?
Power-efficient processors, higher compute density and efficient cooling can help reduce the infrastructure required for AI workloads. FUJITSU-MONAKA is designed to increase AI inference throughput while reducing the server count and power consumption required for equivalent processing loads.
What is sovereign AI infrastructure?
Sovereign AI infrastructure focuses on greater transparency and control over the systems used to process AI workloads. Fujitsu MONAKA Server is designed, developed and manufactured in Japan, with supply chain traceability and hardware-level Confidential Computing.
Do AI servers require liquid cooling?
Not always. AI infrastructure can also use air-cooled systems where the server and workload requirements allow it. The 1U Fujitsu MONAKA Server supports air-cooled operation at temperatures up to 40°C, enabling deployment in existing air-cooled data center environments.
How does Confidential Computing protect data during AI processing?
Confidential Computing protects data and applications while they are being processed in memory. FUJITSU-MONAKA supports hardware-level Confidential Computing based on Arm Confidential Compute Architecture (CCA).
Disclaimer: This new technology applied to the FUJITSU-MONAKA is based on results obtained from a project subsidized by the New Energy and Industrial Technology Development Organization (NEDO).