GlobalFoundries Unveils FDX Fusion Platform for Physical AI with 2028 Dresden Launch

GlobalFoundries Advances FD-SOI Technology with Physical AI Focus

GlobalFoundries has announced plans for FDX Fusion, a next-generation semiconductor platform engineered specifically for Physical AI applications that demand real-time processing beyond cloud infrastructure. The FD-SOI platform represents a significant advancement in the company's technology roadmap, targeting manufacturing launch in 2028 at its Dresden facility in Germany.1

1

The platform addresses the emerging requirements of systems that must sense, process and interact with physical environments instantaneously, marking a shift from traditional cloud-based AI processing to edge-based intelligence.

Technical Specifications Promise 7nm-Class Digital Performance

The first generation of FDX Fusion technology is expected to deliver 7nm-class digital performance while providing more than 2x density improvement over first-generation FDX technology.2

2

The platform combines advanced substrate engineering with front-end transistor innovation and sophisticated back-end interconnect capabilities. RF performance specifications reach up to 0.5 THz ft/fmax, positioning the technology for demanding wireless and communication applications.1

1

Additional features include analog and mixed-signal capabilities, low-leakage operation, in-memory computing, embedded non-volatile memory and dense logic integration, creating a comprehensive solution for Physical AI workloads.

Targeting Sense, Think, Act, Communicate Workloads

FDX Fusion is purpose-built for STAC (Sense, Think, Act, Communicate) workloads that define Physical AI systems.2

2

Ed Kaste, senior vice president of GlobalFoundries' CMOS business, explained that "Physical AI is creating a new class of semiconductor requirements as intelligence moves beyond the cloud and into systems and devices that can process data in real time and interact seamlessly with the physical environment."1

1

The platform aims to reduce system complexity, latency and power consumption while enabling comprehensive edge and Physical AI applications. This approach reflects growing industry recognition that AI-driven applications increasingly require on-device processing rather than cloud dependency.

Broad Market Applications Across Industries

The technology targets applications spanning automotive, consumer and industrial markets, with specific focus on edge inference, robotics, industrial automation, autonomous systems, sensor-fusion architectures and intelligent connected devices.1

1

GlobalFoundries' existing FD-SOI technology already powers billions of chips across automotive radars and sensors, wireless microcontrollers, display drivers and front-end modules for SATCOM and 5G devices.2

2

The new platform builds on this established foundation while addressing next-generation requirements for more sophisticated Physical AI implementations.

Industry Partnerships Shape Development Timeline

GlobalFoundries is collaborating with several customers on early development and evaluation, using real application requirements to inform product design and accelerate progress from platform definition to production readiness.2

2

Major industry players including BMW, Bosch, Infineon, NXP, Siemens and Tesla have expressed support for the technology development. Patrick Press, Executive Vice President External Wafer Manufacturing at Infineon, stated: "Together, Infineon and GlobalFoundries are strengthening their collaboration to define and explore future 7nm class FDX Fusion technologies for next generation radar and wireless communication products."1

1

Demonstrator silicon for FDX Fusion technology is targeted for early customer evaluation in early 2027, with an initial process design kit release planned for mid-2027.1

1

Design Enablement Ecosystem Accelerates Time-to-Market

To accelerate development cycles, customers can leverage GlobalFoundries' design enablement ecosystem, including advanced MIPS and ARC portfolio with workload-driven processor IP, RISC-V instruction set architecture and application-specific processor design capabilities optimized for Physical AI.2

2

These resources complement the FDX Fusion platform with scalable, standards-based solutions across the full Physical AI signal chain. The platform will include optimized process flow and cost structures designed for rapid prototyping and faster time-to-market, addressing customer demands for shortened development cycles in competitive markets.

European Chips Act Supports Dresden Manufacturing Expansion

The Dresden facility expansion receives support under the framework of the European Chips Act, with manufacturing focused on semiconductor solutions for security and defense, industrial competitiveness and supply chain resilience.1

1

GlobalFoundries intends to pursue the technology's development through the upcoming IPCEI Advanced Semiconductor Technologies (AST) program, subject to German federal government approval.3

3

This government backing positions Europe to strengthen its semiconductor manufacturing capabilities and reduce dependency on external suppliers for critical AI infrastructure components. The 2028 timeline suggests GlobalFoundries is positioning itself to capture market share as Physical AI applications mature and demand for specialized edge processing solutions accelerates across automotive and industrial sectors.