GlobalFoundries has officially unveiled FDX Fusion at its facility in Dresden, introducing a brand-new generation of its fully depleted silicon-on-insulator (FD-SOI) technology explicitly tailored for the burgeoning field of physical AI. With this major strategic initiative, the foundry is making a calculated bet that the next massive wave in the semiconductor industry will diverge sharply from the heavy-duty graphical processing unit (GPU) race currently dominating data centers.
The announcement was made against the backdrop of a major celebration in Dresden: the official topping-off of SPRINT, GlobalFoundries’ local fab expansion project. Supported under the European Chips Act, the SPRINT initiative is actively expanding cleanroom, laboratory, and manufacturing capacity at the German site, which GlobalFoundries intends to leverage heavily for differentiated, highly specialized semiconductor technologies moving forward.
According to Manfred Horstmann, senior vice president and general manager of European fabs at GlobalFoundries, the FDX Fusion architecture addresses a fundamental operational split in the market. The chips required for massive data center workloads differ profoundly from the silicon needed when artificial intelligence transitions out of the server rack and moves directly into physical machinery, such as automobiles, autonomous robots, drones, and industrial equipment.
"Data center AI is extremely power hungry," Horstmann said in an exclusive interview ahead of the official public unveiling. "Physical AI is a different class of problem—and it needs more than compute. It needs sensors, actuators, communications, low power, and cost efficiency."
To capture this market, FDX Fusion focuses on a framework GlobalFoundries calls STAC—an acronym representing sense, think, act, and communicate. Rather than prioritizing maximum digital compute density at all costs, the new platform is engineered to integrate substantial processing power alongside the wide array of physical interaction functions that autonomous machines require to operate safely and efficiently in real-world environments.
From FD-SOI to FDX Fusion
At its core, FDX is GlobalFoundries’ proprietary production implementation of FD-SOI technology. In this architecture, the transistor channel sits inside a remarkably thin silicon film—measuring approximately 70 angstroms, or 7 nanometers—positioned directly above a buried oxide layer. Because this silicon layer is so exceptionally thin, the channel is inherently fully depleted during operation.
"You don’t have to waste energy depleting the channel," Horstmann explained.
This physical design grants FDX many of the advanced electrostatic performance advantages normally associated with FinFET architectures, but without requiring the manufacturer to build complex vertical fins. "A lot of the functionality comes from the substrate itself," Horstmann noted. "So we don’t need to build a fin to get a fully depleted transistor."
To bring this next-generation substrate to life, GlobalFoundries has been collaborating closely with research partners CEA-Leti and Soitec under a European cooperative project identified as "Fast and FAMES." The partnership has successfully established a specialized pilot line in France dedicated to producing the strained-silicon substrate. This advanced substrate utilizes a strained 70-angstrom silicon film designed specifically to boost electron mobility by approximately 25%. Horstmann pointed out that this ongoing collaboration builds directly on the foundation laid by the same consortium that previously partnered with GlobalFoundries on its established 22FDX technology.
This advanced substrate plays a crucial role in GlobalFoundries’ engineering roadmap to achieve 7-nanometer-class performance without relying on the complex and expensive manufacturing flows associated with advanced FinFET processes. "For 7-nm-class performance, we don’t need EUV," Horstmann said.
Instead, GlobalFoundries expects the initial FDX Fusion process to require roughly 50 lithography masks. This represents a significant reduction compared with the 70 to 80 masks typically cited for advanced 2-nm and 3-nm FinFET fabrication lines. "The substrate gives us a much simpler process," he said, explaining that fewer required masks directly decrease manufacturing complexity and minimize the potential points of error where defects can be introduced, thereby supporting both higher overall production yields and better cost structures.
Rather than turning to extreme ultraviolet (EUV) lithography, GlobalFoundries plans to lean on advanced ArF immersion lithography, tighter overlay control protocols, and higher-precision photomasks sourced from Dresden-based mask manufacturer AMTC. According to Horstmann, the newest generation of ASML immersion systems delivers substantial improvements in overlay and critical-dimension control. Simultaneously, AMTC’s multi-electron-beam-written masks provide roughly 30% better overlay and critical-dimension margins when compared against standard industry masks.
Physical AI Needs More Than Logic
This deliberate process strategy is vital because GlobalFoundries is not positioning FDX Fusion merely as a lower-cost alternative for shrinking digital transistors. Instead, the platform is intentionally designed to blend dense logic circuitry with the diverse array of analog, radio frequency (RF), memory, and power management components that physical AI systems cannot function without.
"You can use the bulk area for high-voltage devices, and the FD-SOI area for dense logic," Horstmann said. "That makes integrating these different functions much simpler."
The company projects that the initial generation of FDX Fusion will deliver 7-nm-class digital performance while achieving more than twice the density of GlobalFoundries’ first-generation FDX technology. The platform will seamlessly combine dense digital logic with high-performance RF capabilities, analog and mixed-signal functions, low-leakage operational modes, embedded nonvolatile memory, and advanced in-memory computing features.
Furthermore, Horstmann stated that the platform is targeted to reach logic densities ranging from approximately 14 million to 18 million gates per square millimeter, while strictly preserving the low-power consumption and system integration strengths that defined the original FDX architecture.
"You maintain the advantages of 22FDX—extremely low power and very good integration with sensors and actuators," he said. "Extremely low power is what you need for mobile autonomous systems and drones."
To fund this ambitious roadmap, GlobalFoundries expects to invest roughly $1 billion over the coming years into the development of FDX Fusion. This financial commitment includes about $250 million earmarked specifically for the initial capability toolset. Looking ahead at the commercial timeline, the company is targeting initial demonstrator silicon for early 2027, with the rollout of the initial process design kit planned for mid-2027.
A Different AI Opportunity for Europe
For the European semiconductor landscape, the rise of physical AI represents a strategic pivot. Europe faces a distinct structural reality in the artificial intelligence sector, as much of the traditional data center ecosystem and hyperscale infrastructure is heavily concentrated in other parts of the world.
"The data center AI train has left the station," Horstmann said. "But for physical AI, the world is still open."
His central argument is that Europe does not need to contest the specific race already won by overseas competitors. Physical AI heavily rewards a specialized combination of low-power computing, environmental sensing, robust communications, analog integration, and advanced power electronics. This precise matrix aligns much more closely with Europe’s established, long-standing industrial semiconductor strengths.
The industrial cluster centered in Dresden serves as a cornerstone for this regional strategy. The area already hosts major operations for GlobalFoundries, Infineon, and Bosch, alongside specialized research organizations, equipment suppliers, and the upcoming ESMC joint venture led by TSMC.
"The cluster is becoming self-propelling," Horstmann observed. As the regional ecosystem continues to expand and mature, specialized suppliers establish local manufacturing and inventory capacity, equipment vendors maintain dedicated service teams nearby, and design houses are naturally drawn to the geographic hub.
Even the establishment of competing facilities like ESMC serves to strengthen the broader ecosystem, he pointed out, by scaling up Dresden’s overall manufacturing capacity and making the region an even more attractive destination for global supply chain partners.
For GlobalFoundries and the broader European tech sector, that unique convergence of established manufacturing scale, localized supply chains, and differentiated process technology creates a compelling window of opportunity.
"This is a big opportunity for Europe to define a new market," Horstmann concluded, "and to be first."
Leave a Reply