GlobalFoundries has unveiled a 7nm-class FD-SOI process node, dubbed FDX Fusion, which the company says doubles transistor density compared with its established 22FDX platform. The announcement marks the foundry's most significant logic process roadmap step since it abandoned its 7nm FinFET program seven years ago.

The move signals that GlobalFoundries sees a viable commercial path forward on planar fully depleted SOI technology rather than returning to FinFET or adopting gate-all-around architectures for its mainstream customers. The company has long positioned 22FDX as a cost- and power-efficient alternative to bulk FinFET nodes for mobile, IoT and edge applications, and FDX Fusion extends that pitch into territory the industry normally associates with 7nm-class FinFET silicon.

What does FDX Fusion change?

The headline claim is density: GlobalFoundries says the new node doubles transistor density relative to 22FDX, putting a 7nm-class label on an FD-SOI flow. That matters commercially because it lets the foundry target designs that previously required migrating to FinFET-based competitors — TSMC's N7, Samsung's 7LPP — without asking customers to leave the FD-SOI design ecosystem built around 22FDX.

The timing is equally notable. GlobalFoundries canceled its 7nm FinFET program in 2018, citing the prohibitive cost of EUV-class extreme ultraviolet tooling and the multi-billion-dollar investment required to compete at the leading edge. Seven years later, the company is re-entering 7nm-class territory on its own planar terms, a bet that FD-SOI scaling plus advanced packaging and design-technology co-optimization can serve the bulk of demand in its addressable markets.

Why the FD-SOI bet matters now

GlobalFoundries' strategy has consistently targeted customers who value power efficiency, integrated RF and cost predictability over raw density leadership. The 22FDX family has served RF front-ends, IoT controllers and automotive applications — segments where bulk FinFET's advantages are smaller and wafer costs are higher. A 7nm-class FD-SOI node extends that value proposition to more compute-intensive edge designs, including automotive processors and connected devices that need denser logic but not leading-edge performance.

The competitive context also shifted during those seven years. Leading-edge capacity at TSMC and Samsung tightened sharply after the pandemic-era demand surge, and geopolitical pressure has pushed governments to fund geographically diverse, non-leading-edge capacity — including GlobalFoundries' own fab investments in Malta, New York, and Dresden, Germany. A denser FD-SOI node gives the foundry a differentiated offering in exactly the capacity tiers those programs target.

Open questions

The announcement leaves key commercial details unanswered for now. GlobalFoundries has not yet published a full datasheet of confirmed figures — the doubling of 22FDX transistor density is the company's own claim for the roadmap node, and typical disclosures of this kind distinguish between demonstrated silicon and target specifications. Key variables for customers include:

• Whether FDX Fusion requires EUV lithography layers or relies on immersion DUV multipatterning, which would shape wafer cost
• Qualification timelines for automotive and industrial customers, which typically lag volume production by 18–24 months
• Which RF, embedded memory and packaging options carry over from the 22FDX ecosystem

A measured return to advanced scaling

For GlobalFoundries, FDX Fusion is a carefully hedged move: it restores a 7nm-class badge to the company's roadmap without reopening the FinFET capex war it walked away from in 2018. The node's success will hinge on whether designers of edge, automotive and RF-heavy chips see doubled 22FDX density as sufficient to stay on FD-SOI — or whether competitive pressure at mainstream FinFET nodes pulls those designs away. With the disclosure now public, expect customer engagement details and first silicon milestones to define the story over the coming quarters.