TEXTAROSSA · Project

Energy-Efficient Supercomputer Chips and Cooling That Slash Data Center Power Bills

digitalTestedTRL 5

Imagine your company's biggest computer running so hot it needs its own power plant just for cooling — and still can't crunch numbers fast enough. TEXTAROSSA built specialized chips and cooling systems that let supercomputers do more calculations while using far less electricity. Think of it like replacing a gas-guzzling truck engine with an electric one that's also faster. They created ready-to-use hardware designs for faster AI processing, smarter data compression, and built-in security — all tested on real European supercomputer platforms.

By the numbers

consortium partners

countries represented (DE, ES, FR, IT, PL)

FPGA-synthesizable IP blocks delivered

total project deliverables completed

industry partners in consortium

SMEs in the consortium

The business problem

What needed solving

Data centers and HPC facilities face a compounding crisis: computational demands for AI and scientific simulation are growing exponentially, but energy costs and thermal limits are hitting a ceiling. Current hardware wastes power on unnecessary precision, lacks built-in security for emerging quantum threats, and relies on cooling systems that themselves consume enormous energy. Companies need computing solutions that deliver more results per watt without compromising on speed or security.

The solution

What was built

The project delivered 10 concrete FPGA-synthesizable IP blocks: an AI accelerator with mixed-precision Posit arithmetic, data compression engines, quantum-resistant lattice cryptography modules, RISC-V based fast task schedulers, and low-latency inter-node communication links. It also developed two-phase liquid cooling technology and multilevel thermal management, all validated on European supercomputer test platforms across 36 total deliverables.

Audience

Who needs this

Data center operators seeking to reduce power consumption and cooling costsOil and gas companies running large-scale seismic and reservoir simulationsAI companies needing faster and more energy-efficient model training hardwareClimate research centers requiring exascale computing for weather and climate modelsDefense and financial institutions needing quantum-resistant encryption at hardware speed

Business applications

Who can put this to work

Data Center Operations

enterprise

Target: Large-scale data center operators and cloud service providers

If you are a data center operator dealing with skyrocketing electricity costs and thermal management headaches — this project developed two-phase liquid cooling technology and energy-efficient computing IPs that reduce power consumption while boosting processing throughput. The 10 FPGA-synthesizable hardware designs cover AI acceleration, data compression, and fast task scheduling, all tested on European supercomputer architectures.

Oil and Gas Exploration

enterprise

Target: Upstream oil and gas companies running seismic simulation workloads

If you are an oil and gas company spending millions on high-performance computing for reservoir modeling and seismic analysis — TEXTAROSSA built mixed-precision computing accelerators and data compression IPs specifically demonstrated on Oil & Gas workloads. These FPGA-based designs speed up numerical simulations while cutting the energy cost per computation, tested across 15 partner organizations in 5 countries.

AI and Machine Learning Services

mid-size

Target: Companies deploying large-scale AI inference and training pipelines

If you are an AI company struggling with the cost and speed of training and running large models — this project created an AI accelerator with mixed-precision arithmetic including Posit number formats, which squeeze more accuracy from fewer bits of data. The accelerator IP is synthesizable on FPGAs and was developed alongside 36 deliverables covering the full stack from hardware to programming tools.

Frequently asked

Quick answers

What would it cost to adopt these technologies?

The project produced FPGA-synthesizable IP designs (hardware blueprints), not turnkey commercial products. Licensing costs would depend on negotiations with the consortium partners, primarily led by ENEA in Italy. As an RIA (Research and Innovation Action), the IP was developed with public funding, which may facilitate licensing terms.

Can these solutions scale to industrial production?

The IPs were tested on Integrated Development Vehicles mirroring the European Processor Initiative ARM64-based architecture and on an OpenSequana testbed. Scaling to commercial chip production would require additional engineering, but the FPGA-proven designs across 10 demo deliverables demonstrate the core technology works at hardware level.

Who owns the intellectual property and how can I license it?

IP ownership sits with the 15 consortium partners across 5 countries. ENEA (the Italian national energy agency) coordinated the project. Licensing would need to be arranged with the specific partner that developed the relevant IP — for example, the AI accelerator, cryptography, or cooling components may have different owners.

How does this compare to existing commercial HPC solutions?

TEXTAROSSA specifically targets gaps in European exascale computing: mixed-precision Posit arithmetic for better accuracy-per-bit, lattice-based cryptography for quantum-resistant security, and RISC-V based task scheduling. These address limitations in current commercial offerings, particularly around energy efficiency and post-quantum security.

What is the timeline to deployment?

The project closed in March 2024 with 36 completed deliverables including 10 FPGA-synthesizable IP blocks. The technology has been validated on test platforms but would need further integration engineering for production deployment. Based on available project data, commercial readiness could be achieved within additional development cycles.

Is the cooling technology ready for my data center?

The two-phase liquid cooling system was developed and tested within the project's demonstrator environments. Based on available project data, it has been validated at lab and testbed scale but would require site-specific engineering for deployment in a commercial data center facility.

Consortium

Who built it

The TEXTAROSSA consortium brings together 15 partners from 5 European countries, with a research-heavy composition: 7 research organizations and 5 universities provide deep technical expertise, while 3 industry partners and 2 SMEs add commercial perspective. The 20% industry ratio is typical for a Research and Innovation Action focused on pre-competitive technology. The project is led by ENEA, Italy's national agency for energy and sustainable development, which gives it institutional credibility but means commercialization will likely depend on industry partners or spin-off licensing. The geographic spread across Germany, Spain, France, Italy, and Poland covers Europe's major HPC ecosystems.

AGENZIA NAZIONALE PER LE NUOVE TECNOLOGIE, L'ENERGIA E LO SVILUPPO ECONOMICO SOSTENIBILECoordinator · IT
INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET AUTOMATIQUEparticipant · FR
UNIVERSITA DEGLI STUDI DI TORINOthirdparty · IT
INSTYTUT CHEMII BIOORGANICZNEJ POLSKIEJ AKADEMII NAUKparticipant · PL
UNIVERSITA DI PISAthirdparty · IT
E 4 COMPUTER ENGINEERING SPAparticipant · IT
CONSIGLIO NAZIONALE DELLE RICERCHEparticipant · IT
UNIVERSITE DE BORDEAUXthirdparty · FR
BULL SASparticipant · FR
POLITECNICO DI MILANOthirdparty · IT
BARCELONA SUPERCOMPUTING CENTER CENTRO NACIONAL DE SUPERCOMPUTACIONparticipant · ES
CONSORZIO INTERUNIVERSITARIO NAZIONALE PER L'INFORMATICAparticipant · IT
ISTITUTO NAZIONALE DI FISICA NUCLEAREparticipant · IT

How to reach the team

ENEA (Agenzia Nazionale per le Nuove Tecnologie, l'Energia e lo Sviluppo Economico Sostenibile), Italy — contact through SciTransfer for introductions

Order a report on this consortium See AGENZIA NAZIONALE PER LE NUOVE TECNOLOGIE, L'ENERGIA E LO SVILUPPO ECONOMICO SOSTENIBILE profile →

Next steps

Talk to the team behind this work.

Want to explore licensing TEXTAROSSA's energy-efficient computing IPs or cooling technology for your data center? SciTransfer can connect you directly with the right consortium partner. Contact us for a tailored introduction.

Order a report on this topic More in Digital & ICT