← Back to feed

GlobalFoundries marked the topping-out of its SPRINT expansion project in Dresden on October 9, 2026 — a construction milestone that marks the physical completion of the main structure — and simultaneously announced FDX Fusion, a next-generation process platform targeting Physical AI applications.

The expansion is one of the largest active semiconductor capital projects in Europe.

SPRINT Expansion: Capacity Numbers

The €1.1 billion SPRINT project will increase Dresden’s annual production capacity to more than one million 300mm wafers by the end of 2028, up from a current ceiling of approximately 950,000 wafers per year depending on product mix.

That makes Dresden the largest foundry site in Europe by output. GF has operated in Dresden since 1995 and has invested more than €10 billion in the site since 2009. Current headcount is around 3,000 people across approximately 60,000 m² of cleanroom and laboratory space.

The expansion is funded in part through the European Chips Act and supported by the German federal government and the Free State of Saxony. It qualifies as a strategic project under the EU’s broader effort to increase domestic semiconductor production from roughly 10% to 20% of global output by 2030.

FDX Fusion: What It Is

FDX Fusion is GlobalFoundries’ next iteration of its Fully Depleted Silicon-on-Insulator (FDSOI) technology, designed specifically for Physical AI — the class of applications that require real-time sensing, actuation, and control at the edge: autonomous vehicles, industrial robots, aerospace systems, and connected defense hardware.

The platform is engineered to combine four capabilities in a single process:

• Digital performance: 7nm-like transistor performance at lower power
• Analog and mixed-signal: High-precision sensing and signal conditioning without a separate process
• Embedded memory: On-chip SRAM and non-volatile memory for local model storage and inference
• High-performance RF: Wireless connectivity integrated at the die level

The integration target matters for Physical AI. Current edge AI deployments typically require separate dies for compute, connectivity, and analog sensing, with integration happening at the package level. FDX Fusion is designed to collapse those into a single process node, reducing system size, power, and latency for closed-loop control.

GF intends to pursue FDX Fusion development through the IPCEI Advanced Semiconductor Technologies program, subject to German federal government approval.

Strategic Context

GF occupies a specific position in the semiconductor industry: it does not chase leading-edge logic (it exited sub-7nm in 2018) but dominates differentiated mature nodes where power efficiency, analog integration, and long production lifetimes matter more than peak transistor density. Automotive, aerospace, industrial IoT, and defense customers are the core.

Physical AI as a workload class is well-matched to that position. Training frontier models requires leading-edge GPU clusters; deploying them in physical systems at the edge requires low-power, highly integrated silicon that can run reliably in harsh environments for a decade. GF’s FDX technology already powers automotive radar, LiDAR control chips, and secure microcontrollers in defense applications.

FDX Fusion is the bet that Physical AI demand will sustain and expand that market through the end of the decade.