SciTransfer
Organization

XILINX IRELAND UNLIMITED COMPANY

Global FPGA semiconductor company active in EU research on AI hardware acceleration and terabit optical transceivers for data centers.

Large industrial companydigitalIENo active H2020 projectsThin data (2/5)
H2020 projects
2
As coordinator
1
Total EC funding
€165K
Unique partners
9
What they do

Their core work

Xilinx Ireland is the European base of Xilinx — now part of AMD — the world's leading designer of Field-Programmable Gate Arrays (FPGAs), reconfigurable silicon chips used to accelerate compute-intensive workloads in data centers, telecommunications, and aerospace. In EU-funded research, they have contributed FPGA hardware expertise to neural network acceleration and served as an industry partner in the development of terabit-capable optical transceivers for data center applications. Their participation in EU projects reflects deliberate alignment of internal product roadmaps with emerging academic research in AI hardware and high-speed photonic interconnects. As a global semiconductor company, they bring commercial deployment capacity that most research partners cannot offer.

Core expertise

What they specialise in

FPGA-based AI and neural network accelerationprimary
1 project

Coordinated TPANN (2017–2018), a project explicitly targeting tensor processing for artificial neural networks on FPGA hardware.

High-speed optical transceivers for data communicationssecondary
1 project

Participated in Caladan (2019–2023), an Innovation Action targeting terabit/s optical transceivers for datacom using silicon photonics and micro transfer printing.

Silicon photonics and photonic assemblyemerging
1 project

Caladan keywords include Silicon Photonics, GaAs quantum dot laser, and Photonic assembly and packaging — all core to next-generation optical interconnect design.

Evolution & trajectory

How they've shifted over time

Early focus
FPGA neural network acceleration
Recent focus
Silicon photonics optical transceivers

Their earliest H2020 involvement (TPANN, 2017–2018) focused on compute acceleration — mapping neural network tensor operations onto reconfigurable FPGA silicon, which was prescient given the AI hardware boom that followed. By 2019, their focus shifted decisively toward optical interconnect technology: silicon photonics, micro transfer printing, and GaAs quantum dot lasers for terabit-scale transceivers. This trajectory mirrors a broader industry shift: as AI workloads filled data centers, bandwidth between chips and servers became the new bottleneck, and Xilinx moved to address it at the photonic layer.

They are tracking the data center infrastructure stack from compute (FPGA AI acceleration) toward connectivity (terabit optical transceivers), making them a relevant partner for any consortium working on AI infrastructure, edge computing, or next-generation data center networking.

Collaboration profile

How they like to work

Role: specialist_contributorReach: European5 countries collaborated

Xilinx has taken both a leadership and a supporting role across their two projects — coordinating TPANN as a Marie Curie fellowship host institution, and joining Caladan as a technical industry partner in a larger Innovation Action. Their small EU network (9 partners, 5 countries) points to selective, strategically targeted engagement rather than broad consortium participation. As a major semiconductor vendor, they likely enter EU projects to gain early access to research that intersects with their product roadmap, lending commercial credibility to the consortia they join.

Their EU research network spans 9 unique partners across 5 countries — compact by H2020 standards, indicating focused rather than expansive collaboration. No repeated partners are visible across the two projects, suggesting they engage with different research communities depending on the technology area.

Why partner with them

What sets them apart

Xilinx (now AMD Xilinx) is one of very few global semiconductor companies with direct H2020 participation, bringing a level of commercial scale and product deployment capability that research institutes and SMEs cannot replicate. A consortium that includes Xilinx gains industrial anchoring — their involvement signals that a technology has a plausible path to real-world chips and systems. For scientists or companies building consortia around AI hardware, optical interconnects, or data center infrastructure, Xilinx represents a direct line to the FPGA and photonics markets that dominate those industries.

Notable projects

Highlights from their portfolio

  • TPANN
    Xilinx acted as coordinator and MSCA host, placing a researcher inside their R&D environment to work on FPGA-based neural network acceleration — a topic that became commercially critical within two years of the project's completion.
  • Caladan
    A four-year Innovation Action targeting terabit/s optical transceivers using micro transfer printing and GaAs quantum dot lasers — directly addressing the data center bandwidth crisis that defines modern hyperscale infrastructure investment.
Cross-sector capabilities
AI and machine learning hardware infrastructureHigh-performance computing and data center networkingTelecommunications and optical communicationsSemiconductor research and advanced manufacturing
Analysis note: Profile is built on only 2 projects with sparse keyword data. However, Xilinx is a globally recognized semiconductor brand (acquired by AMD in 2022 for $49B), so their domain expertise in FPGAs and high-speed computing is well-established beyond what H2020 records reflect. The MSCA coordinator role represents hosting an incoming researcher, not leading a full technical consortium — this distinction matters for assessing their project leadership experience in EU research.