Pranav Hotkar
Espoo, Finland - February 23, 2026 - Nokia and KDDI have demonstrated a quantum-safe optical transport network designed to secure AI workloads and data center connectivity against future quantum-computer threats, the companies announced Monday.
The trial focuses on protecting data moving between high-capacity compute clusters, a growing concern as hyperscale AI training and inference increasingly depend on long-distance data-center interconnect (DCI) links. The partners said the technology integrates post-quantum cryptography into optical transport, enabling encrypted communications resistant to attacks from future quantum computers.
According to Nokia, the demonstration shows how operators can maintain the confidentiality of sensitive enterprise and AI data even after quantum computing becomes capable of breaking traditional encryption methods. The solution combines secure key exchange and optical networking capabilities to safeguard high-throughput traffic across metro and long-haul networks used by cloud and AI infrastructure.
The announcement reflects a broader shift in telecom and cloud infrastructure planning. As AI models scale and data moves continuously between facilities, network transport, not just servers, is becoming part of the security perimeter. Quantum-safe encryption is therefore emerging as a prerequisite for future-proof data-center architecture.
KDDI said protecting inter-data-center traffic is critical for next-generation services such as distributed AI processing, edge computing, and real-time analytics. The operator is exploring commercial deployment scenarios in high-capacity backbone and data-center interconnect environments.
For Nokia, the project aligns with growing demand from hyperscalers and telecom operators preparing infrastructure for the “post-quantum” era. The company has increasingly positioned optical networking as foundational to AI infrastructure, alongside compute and storage.
Industry analysts note that quantum-safe networking is moving from research to implementation because data encrypted today could be stored and decrypted later once quantum computers mature, a risk particularly relevant to government, financial, and enterprise AI datasets.
The companies did not disclose commercialization timelines but indicated the technology is intended for future rollout in AI-ready networks and cloud data-center environments.
The development signals that cybersecurity strategy in data centers is expanding beyond software controls to include the physical transport layer, as operators prepare for both exponential AI traffic growth and next-generation cryptographic threats.
