NUMTECH

Their core work

NUMTECH is a French technology SME based near Clermont-Ferrand that develops and applies scientific computing software, contributing domain application expertise to large-scale HPC and heterogeneous computing research consortia. Their H2020 participation reveals a focus on making high-performance computation practical for industrial and societal use cases — specifically through programmable heterogeneous architectures and domain-specific language toolchains that allow engineers to run demanding analytics at extreme scale. In EVEREST, they worked on the software environment side of heterogeneous platforms, suggesting their core product or service involves numerical simulation or data-intensive computation that benefits directly from next-generation HPC infrastructure. They operate at the intersection of applied software engineering and high-performance computing, translating research-grade platform advances into industrial workflows.

What they specialise in

How they've shifted over time

NUMTECH's first H2020 project (LEXIS, 2019) left no recorded keywords, making it difficult to characterize their early focus precisely — the project title points to infrastructure-level HPC for industrial application, but their specific contribution is opaque from the data. By EVEREST (2020), their profile sharpens considerably: keywords cluster around heterogeneous architectures, high-level synthesis, domain-specific languages, and AI, indicating a shift toward the software toolchain and programmability layer that sits above raw HPC hardware. The trajectory suggests they moved from being consumers of HPC infrastructure toward contributing to the tools and environments that make heterogeneous platforms programmable for domain engineers.

NUMTECH is moving up the stack — from using HPC infrastructure to shaping the software environments and domain-specific toolchains that make heterogeneous computing accessible, a direction that aligns with growing demand for AI-accelerated scientific computation in industry.

How they like to work

NUMTECH participates exclusively as a consortium partner — they have never led an H2020 project — which positions them as a specialist contributor that brings focused technical expertise rather than project management capacity. Despite only two projects, they engaged with 24 distinct partners across 9 countries, suggesting they join large, diverse consortia rather than tight bilateral arrangements. This breadth indicates they are comfortable operating within complex multi-stakeholder research environments, though prospective partners should not expect them to drive project governance or coordination.

NUMTECH has built a surprisingly wide network for a two-project SME — 24 unique partners spanning 9 countries, all within European HPC and digital infrastructure ecosystems. Their geographic reach is solidly European with no evidence of partnerships outside the continent.

What sets them apart

NUMTECH occupies a rare niche as a French SME that bridges scientific computation applications and the heterogeneous hardware platforms that run them — most players in this space are either pure hardware vendors or academic research groups. Their combination of domain-specific language expertise and AI integration within HPC contexts makes them relevant to any consortium needing an industrial software perspective on extreme-scale computing. For consortium builders, they offer the credibility of a private-sector actor with direct exposure to production-scale computation challenges, without the overhead of a large corporate partner.

Highlights from their portfolio

• EVEREST
  Their highest-funded project (EUR 265,938) and the source of all recorded technical keywords — a long-duration IA project (2020-2024) targeting extreme-scale big data analytics on heterogeneous platforms, revealing NUMTECH's most specific and current technical identity.
• LEXIS
  NUMTECH's entry into H2020 collaboration, joining a large-scale RIA project connecting HPC, cloud, and LEXIS middleware infrastructure to industrial and societal use cases — establishing their position within the European HPC ecosystem.