Project description
Innovative processing technology accelerating random workloads
Commercial processing hardware, including CPUs, GPUs and TPUs, utilises software-based random number processing solutions for executing stochastic workloads. While efficient, this can lead to increased processing time and higher costs for stochastic workloads, causing bottlenecks and even impacting security in many applications. The Randomness Processing Unit (RPU), developed by Quside, utilises quantum random number generators and hardware-based stochastic acceleration techniques and is designed to enhance performance and efficiency in stochastic high-performance computing and post-quantum cryptography. Based on this solution, the EIC-funded RPU project aims to accelerate stochastic workloads in sectors such as data centres, finance and physics/chemistry simulations. It will target markets exceeding EUR 10 billion and is expected to generate over EUR 100 million in turnover by the end of the decade.
Objective
Quside has developed the Randomness Processing Unit (RPU) that is an innovative hardware accelerator designed for world-class performance, optimization, and efficiency in stochastic High Performance Computing (HPC) and Post-Quantum (PQ) cryptography. The purpose of the RPU is to accelerate and optimize randomized workloads, which are used in countless sectors (secure data transfer, finance, physics/chemistry simulations etc.). All the commercial processing/acceleration hardware (CPUs, GPUs, TPUs, etc.) generate pseudo-random numbers (PRNGs), which are sequences of digits with certain “random” statistical properties. In fact, PRNG are completely predictable and therefore are inefficient/ineffective for randomized workloads. In short, the existing processing hardware are not optimized for random numbers and operations. As a result, customers run highly inefficient processes that impact the quality of their results and increases their costs (both OPEX and CAPEX).
Our technology is brand-new (the TRL 7 prototype of the RPU was presented in December 2022) and protected by 14 patent families (filed or granted). The RPU is based on the high-speed, high-quality Quantum Random Number Generators (QNRG) chips manufactured by Quside. Ours is one of the few companies worldwide working on this technology. Quside is a spin-off of the Institute of Photonics Sciences in Barcelona, and our solutions have been used in the experiments awarded with the 2022 Nobel Prize for Physics. According to our financial forecasts, Quside will generate over 170 M€ in turnover and 40 M€ in net profit in the period 2024-2028, mostly from sales of RPUs to customers in the public and private sectors.
Note that this is a direct resubmission to the pitch phase, as recommended by the EIC jury.
Fields of science
- natural sciencescomputer and information sciencescomputer securitydata protection
- natural sciencescomputer and information sciencescomputer securitycryptography
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringcomputer hardwaresupercomputers
Programme(s)
- HORIZON.3.1 - The European Innovation Council (EIC) Main Programme
Funding Scheme
HORIZON-EIC-ACC-BF - HORIZON EIC Accelerator Blended FinanceCoordinator
08860 Castelldefels
Spain
The organization defined itself as SME (small and medium-sized enterprise) at the time the Grant Agreement was signed.