Quantum-based Randomness Processing Units (RPUs) for High-Performance Computation and Data Security

Project Information

RPU

Grant agreement ID: 101145131

DOI

10.3030/101145131

Project closed

EC signature date 7 February 2024

Start date 1 March 2024

End date 28 February 2026

Funded under

• The European Innovation Council (EIC)

Total cost

€ 3 908 437,50

EU contribution

€ 2 499 999,00

Coordinated by

QUSIDE TECHNOLOGIES SL
Spain

Periodic Reporting for period 1 - RPU (Quantum-based Randomness Processing Units (RPUs) for High-Performance Computation and Data Security)

Reporting period: 2024-03-01 to 2025-02-28

Summary of the context and overall objectives of the project

Industries like finance, cybersecurity, data centers, logistics, and telecom face a growing crisis of efficiency and security. One of the most urgent challenges is the inability of current hardware, whether CPUs, GPUs, or even TPUs, to generate truly high-quality random numbers at scale. This creates weak points across encrypted communication, secure data storage, and stochastic operations such as Monte Carlo simulations. Compounding this is the lack of support for randomized workloads, which are increasingly common in high-performance computing and post-quantum cryptography. The result is a system that consumes vast amounts of resources but still leaves organizations vulnerable to attacks exploiting poor entropy sources.
Quside addresses these twin challenges with the Randomness Processing Unit (RPU), a hardware accelerator that combines quantum-based entropy generation with reprogrammable logic to deliver both security and performance. At its core, the RPU integrates a Quantum Random Number Generator (QRNG) that uses quantum physics to produce provably unpredictable randomness. This entropy is then harnessed by FPGA-based acceleration logic, which optimizes cryptographic operations like key generation and encryption. Designed to comply with standards from NIST, BSI, and ISO, the RPU offers a drop-in solution for organizations needing secure, high-throughput randomness.
In parallel, Quside is advancing its photonic technology by designing a next-generation Photonic Integrated Chip (PIC). This chip, which improves efficiency and scalability, will be central to the future evolution of the RPU product line

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

During this phase, Quside developed a PCIe-based prototype capable of delivering 1 Gbps of quantum entropy using a Gen3 x4 interface. The hardware was built on a custom PCB integrating the company’s proprietary QRNG chip alongside reprogrammable acceleration logic. This setup serves as a platform for accelerating full cryptographic workloads in real-world applications, not just the entropy generation itself. A key milestone was the decision to incorporate selected RPU functionalities into Quside’s Garnet product line to enable commercial acceleration of cryptographic processes.
Simultaneously, software development progressed to support the RPU’s integration into existing infrastructure. This includes a PCIe driver and SDK that simplify interaction between the card and host systems. Efforts also began to integrate cryptographic algorithms such as AES, SHA-3, and NTT, enabling the QRNG to function not just as a source of entropy but as a full cryptographic accelerator. All algorithmic work was benchmarked against established standards to ensure compliance and facilitate adoption.
On the PIC front, the team successfully designed a uMZI interferometer with path-length asymmetry. This configuration enables efficient quantum interference at 1 GHz repetition rates, marking a significant improvement over previous PIC designs. The result is a robust, low-loss platform that supports scalable entropy generation in future deployments.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

Quside’s RPU technology represents a substantial leap in secure computing, enabling Europe to lead in quantum-classical hardware innovation. The core achievement, a reprogrammable PCIe accelerator optimized for post-quantum cryptography and stochastic computing, was complemented by the development of a high-speed dual-laser PIC. Together, these components offer true, certifiable entropy generation at scale.
The project also delivered a production-ready software stack with Linux drivers and integrations for tools like OpenSSL and libOQS. This ensures compatibility with modern IT environments and eases adoption in data centers. The commercial impact is already visible, with some of the RPU’s core features integrated into the Garnet Plus A41 QRNG, opening new revenue opportunities and operational cost savings through more efficient, hardware-level entropy delivery.
Pilot projects with the Barcelona Supercomputing Center, Vodafone, and PQShield helped validate the technology across diverse environments. Nine patents covering cryptographic and hardware innovations were filed, solidifying Quside’s IP position and future roadmap.
The scientific implications are significant. By bridging quantum entropy with classical computing, Quside enables breakthroughs in AI, hybrid systems, and lattice-based cryptography. Economically, the solution reduces the computational overhead of secure randomness, lowering costs and energy consumption. From a societal perspective, it strengthens digital trust in sectors like healthcare and finance, aligning with the EU’s climate and sovereignty goals by offering a secure and energy-efficient European solution