From the Research Universities edition of October 23, 2026
Greece inaugurates Daedalus (19 Oct) (85.69 petaflops, €58.9 million, NVIDIA GH200), run by GRNET at NTUA's Lavrio site, to process sensitive research data domestically; business access via Pharos AI Factory planned for March 2027
GRNET (National Infrastructures for Research and Technology), National Technical University of Athens, Pharos AI Factory (as reported by Naftemporiki) · National research and education computing infrastructure · Greece / European Union (international; transferable to U.S. state research-computing initiatives with limits)
- Publisher
- Naftemporiki
- Original publication
- 19 October 2026 (Naftemporiki English edition; republished by China Economic Net on 23 October)
- Source retrieved
- 2026-10-24
- Event date
- 2026-10-19
What happened
Naftemporiki reports that Greece inaugurated Daedalus, its most powerful supercomputer, in a former power station at the Lavrio Technological and Cultural Park operated by the National Technical University of Athens. It has a €58.9 million budget, measured 85.69 petaflops, ranks 31st worldwide and was 23rd on the June 2026 Green500. Built on liquid-cooled NVIDIA GH200 technology, it is operated by GRNET; HPE supplied computing equipment (€41 million) and Mainsys the data-centre fit-out (€17.9 million), funded by the Recovery and Resilience Facility and EuroHPC. A key stated benefit is processing sensitive or confidential data without sending it abroad, for research institutions, healthcare providers, public bodies and businesses. Technical testing follows; Pharos AI Factory business services are expected in March 2027, with free access for start-ups and SMEs under EU state-aid limits and usage charges considered for large firms; services, terms and pricing are not final. GRNET is tendering a second, €21.57 million system at the University of Western Macedonia for AI and sensitive-data analysis, intended to run models trained on Daedalus.
Why it matters
Mirrors U.S. state initiatives (for example New Jersey's and New York's) that pair universities and governments around shared AI compute: sovereign handling of sensitive data, a separate access channel for business, and a train-here, run-there two-site design. Transferability is limited by EU funding and state-aid rules.
Evidence and measured results
Read the full Naftemporiki article (JSON-LD datePublished 19 Oct 2026); the same text was republished by China Economic Net on 23 Oct: inauguration, location and operator, €58.9 million budget, 85.69 petaflops, 31st worldwide, Green500 23rd (June 2026), GH200 and liquid cooling, HPE €41 million and Mainsys €17.9 million contracts, RRF and EuroHPC funding, sensitive-data rationale, Pharos AI Factory (€30 million, 50% EuroHPC) timeline and access terms, Hellenic AI Factory ownership, and the Kozani tender (€21.57 million).
Limitations and uncertainty
Greek business-news report; figures are as reported, not checked against GRNET's own release or the TOP500 list. Published 19 October, slightly before the last edition window; included because it was not covered by any stream, was republished internationally on 23 October, and pairs with the Cal Poly launch. System is entering a testing period, so no usage or research outcomes yet. EU funding and state-aid rules limit transferability to U.S. institutions.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-23; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Public research institutions need large-scale AI compute that can handle sensitive data without relying on foreign or commercial infrastructure.
- Stakeholders
- State research-computing consortium board
- University CIOs and research VPs
- State economic development agency
- Research network operator
- Discovery questions
- Which research datasets cannot leave your jurisdiction or institution?
- How would you separate research and business access?
- Is a second site for inference or sensitive data planned?
- Value hypothesis
- A shared sovereign system with defined tiers may let institutions run sensitive AI research locally while serving business users separately.
- Potential engagement
- An access-tier and data-governance design study for a shared state or consortium AI system.
- Unsupported claims
- Research productivity gains
- Lower cost than commercial cloud
- Direct applicability of EU funding models in the U.S.
Pre-sales engineering
Role takeaway
- Fit
- Fits consortia planning a shared training-class system with sensitive-data workloads.
- Architecture and integration
- Liquid-cooled training system plus a separate inference and sensitive-data site, linked through a research network with sandboxes and trusted runtime environments.
- Prerequisites
- Federated identity across member institutions
- Data-classification scheme
- Network connectivity between sites
- Constraints
- EU state-aid and funding rules
- System in testing; no operational data
- Reported figures not independently verified
- Security
- Trusted runtime environments and secure sandboxes for confidential data, with domestic processing as a design requirement.
- Proof of value
- Run a sensitive-data workload in a trusted environment end to end
- Move a trained model from training to serving site
- Measure allocation turnaround for research users
Delivery
Role takeaway
- Work
- Define research and business access tiers, data rules, onboarding, training and support for a shared system.
- Dependencies
- Funding and procurement completion
- Member-institution agreements
- Pricing policy decisions
- Ownership
- Research network operator, with a consortium oversight board for allocation policy.
- Skills and adoption
- Researchers and SMEs need training on agents, LLM tooling and secure environments; planned cohorts are small.
- Governance checkpoints
- Allocation policy approval before general access
- Data-protection review of trusted environments
- Pricing and state-aid compliance review
- Acceptance criteria
- Proposed: published access terms before general availability
- Proposed: sensitive workloads confined to trusted environments
- Proposed: research allocation turnaround within an agreed target
- Risks
- Business demand crowding out research
- Undefined pricing and terms
- Second-site budget and schedule slippage
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
A two-site model: a training-class national system plus a separate green data centre for inference and sensitive-data analysis, with sandboxes and trusted runtime environments offered through the AI factory.
Governance
Who approves, reviews and stays accountable for outcomes?
Access rules, pricing and service terms are still undefined at launch; research priority versus business access will need explicit allocation policy.
Security and privacy
What data, permissions and controls need testing?
Domestic processing of sensitive or confidential data is the headline rationale; trusted runtime environments and secure sandboxes are planned services.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Planned AI-agent training (45 participants, 10 November) is small relative to the system's capacity; user support will shape uptake.
Procurement
What should contracts, pricing and exit terms secure?
Separate contracts for compute (HPE) and facility fit-out, funded through EU recovery and EuroHPC programmes; the second site has a 12-month delivery period and a budget likely to rise.
Operating model
Which teams own the service once it runs?
A national research-network operator runs the system while a separate state-owned company handles business commercialisation.
Publication history
- 2026-10-23Research Universities · Issue 454 resources
Stable resource ID: greece-daedalus-supercomputer-sovereign-ai-research-20261023