Collaboration

Partnerships for research, development, and technology transfer – from collaborative projects to tailored innovation solutions.

About SAL

Insights into SAL’s mission, values, research activities, and contribution to Europe’s innovation ecosystem.

ViTFOX

ViTFOX is an international collaboration between Europe and Korea that aims to advance next-generation semiconductor technology for artificial intelligence. The project focuses on developing “ferroelectric-augmented” chips capable of running Vision Transformers (ViTs) with exceptional energy efficiency—exceeding 50 trillion operations per second per watt (TOPS/W).

Project goals

The project’s goal is to build the fundamental blocks of an in-memory network accelerator by exploiting the energy efficiency of ferroelectric switching.

The project aims to contribute to the semiconductor value chain at three key levels:

  • Integration and Testing: Implementation of a neuromorphic engine at board level
  • Back-End Processing: Enabling BEOL functionalization through monolithic integration of complex oxides and heterogeneous integration of advanced functional materials
  • Materials level: Control of ferroelectric switching at the nanoscale to realize low-power, high-density memory devices

Project
facts

Title: Vision Transformers with Ferroelectric Oxides
Program: Horizon 2024
Funding agency: Chips Joint Undertaking
Project Coordinator: National Center for Scientific Research Democritus
Duration: 1.10.2024 – 30.9.2027

ViTFOX

Vision Transformers with Ferroelectric Oxides - Bridging Europe and South Korea to Build the Next Generation of AI Chips

Similar Projects

Discover more projects with similar focuses and technologies.

AURORA

In collaboration with TDK Electronics, Silicon Austria Labs (SAL) has developed next-generation 2D micromirrors for AR/VR applications. The focus was on a wide field of view, low power consumption, and a compact design. The components were realized as prototypes using SAL's 200mm PiezoMEMS technology platform....

PrintedRadar

Print­e­dRadar aims to explore the addi­tive manu­fac­turing of novel and effi­cient 3D wave­guide radar antenna systems.

EDISON

EDISON develops AI-driven algorithms to optimize the management of decentralized battery storage systems within a virtual power plant. The goal is to maximize the utilization of renewable energies, improve grid stability, and advance sustainable energy practices through reinforcement learning, federated learning, and practical validation...

Contact

Business Development
Lisa Kainz MA
Business Development Microsystems
Graz
Project Manager
Dr. Venkata Raveendra Nallagatla
Scientist Thin Film Technologies
Villach

Newsletter

Our Science & Stories newsletter provides the latest news, bietet aktuelle Neuigkeiten, Forschungshighlights und spannende Einblicke
von Silicon Austria Labs – direkt ins Postfach.