Greece’s DAEDALUS Supercomputer Gives Pharos AI Factory a New Compute Backbone
Greece has inaugurated DAEDALUS , a new supercomputer in Lavrion near Athens that will serve as the computing backbone for the country’s Pharos AI Factory. The system is part of the European Union’s wider EuroHPC strategy for AI-focused computing infrastructure, with the stated aim of making advanc
Posted on Oct 23 • Originally published at scalevise.com
Greece has inaugurated DAEDALUS, a new supercomputer in Lavrion near Athens that will serve as the computing backbone for the country’s Pharos AI Factory. The system is part of the European Union’s wider EuroHPC strategy for AI-focused computing infrastructure, with the stated aim of making advanced AI resources more accessible to researchers, startups, SMEs and public institutions.
According to the EuroHPC JU and GRNET inauguration announcement, DAEDALUS delivers sustained performance of 85.69 petaflops and peak performance of 115 petaflops. Built by Hewlett Packard Enterprise using NVIDIA GH200 Grace Hopper technologies, it is hosted and operated by Greece’s National Infrastructures for Research and Technology, GRNET.
For businesses, the important development is not simply the machine’s performance ranking. DAEDALUS gives Pharos a domestic high-performance computing foundation for AI innovation. Whether that produces practical value for a specific company will depend on the access programmes, support services and project requirements that Pharos makes available.
DAEDALUS is the physical computing system. Pharos is the Greek AI Factory initiative designed to use that capacity to accelerate AI research and innovation, while connecting Greece to the wider European AI Factory network and related EU initiatives.
The distinction matters because high-performance hardware alone does not make an AI project easier to deploy. Companies need a viable use case, suitable data, technical expertise and a route for connecting any resulting model or workflow to day-to-day operations.
DAEDALUS is funded through a EUR 36 million acquisition project. The EuroHPC Joint Undertaking contributed about 35% of that cost, with the remaining funding coming from Greece’s National Recovery and Resilience Plan, Greece 2.0, under NextGenerationEU. Separately, the Pharos AI Factory has EUR 30 million in total funding, including about EUR 15 million in EU contributions.
Petaflops measure computing performance. A sustained figure indicates performance achieved over a continued workload, while peak performance refers to the system’s highest theoretical capability. These figures help show why DAEDALUS can support demanding research and AI workloads that would be difficult to run on ordinary business infrastructure.
That does not mean every AI initiative needs supercomputing resources. Many business use cases, such as connecting existing AI services to customer support, sales or internal knowledge workflows, may not require this type of infrastructure. But access to an AI Factory can matter where a project genuinely needs advanced compute capacity, specialised experimentation or collaboration with research organisations.
Pharos is also positioned within a European network rather than as an isolated Greek project. The initiative is intended to connect with National Competence Centres, European Digital Innovation Hubs, DeployAI and other EU activities. For teams exploring more ambitious AI work, those links may create routes to technical support, partnerships and infrastructure beyond a single provider.
The confirmed information establishes the infrastructure, its performance, its funding and Pharos’s access-oriented mission. It does not establish a public price list, a standard commercial access process, service-level commitments or a defined timetable for every type of user.
Businesses should therefore avoid treating the announcement as a guarantee of immediate, low-cost compute for any project. The practical questions that remain include:
The strongest near-term opportunity is for organisations with a clearly defined, compute-intensive AI problem to follow Pharos developments and assess whether its programmes fit their work. A smaller team does not need to operate a supercomputer to benefit from the initiative, but it does need to match its project to the capability and access route that ultimately becomes available.
If your team is assessing whether public AI infrastructure can support a real workflow, Scalevise can turn broad compute access into a focused plan: prioritize a use case, assess data and integration needs, and define a measurable pilot. Our practical AI implementation consultancy helps businesses avoid spending effort on experiments that do not reach operations. Request an AI consultancy conversation to identify the next feasible step.
DAEDALUS is a supercomputer inaugurated in Lavrion near Athens on October 19, 2026. It is hosted and operated by GRNET and provides the computing backbone for Greece’s Pharos AI Factory.
DAEDALUS delivers 85.69 petaflops of sustained performance and 115 petaflops of peak performance. It was built by Hewlett Packard Enterprise with NVIDIA GH200 Grace Hopper technologies.
Who is Pharos AI Factory intended to serve?
Pharos is intended to offer access to AI computing capacity for researchers, startups, SMEs and public institutions, while supporting AI innovation in Greece and links to EU initiatives.
The verified information does not provide a public price list or standard commercial pricing for DAEDALUS or Pharos access.
DAEDALUS gives Greece a significant new AI computing resource and supplies the technical foundation for the Pharos AI Factory. Its business relevance will depend on how Pharos translates that capacity into accessible programmes, support and project pathways. For organisations with demanding AI ambitions, the development is a reason to watch the Greek and European AI Factory ecosystem closely, while keeping project selection and implementation needs firmly in view.