Designing PAs at the Speed of AI: Falcomm Introduces GaNdalph.ai using GF RFGaN1 at IMS 2026

June 5, 2026 -- As the RF and microwave community gathers in Boston this month for IMS 2026, Falcomm and GlobalFoundries (GF) are bringing something new to the conversation: an AI-native EDA software tool that can do in minutes what has historically taken months. As the world’s premier RF conference opens its doors at the Thomas M. Menino Convention Center, Falcomm is unveiling GaNdalph.ai, an AI-native software tool for RF/mmWave chip design, debuting with first-class support for GF’s RFGaN1 (130RFGaN) process and a roadmap that extends across the broader GF portfolio and beyond.

A Foundation Model for RF Design

GaNdalph.ai is not a general-purpose AI tool adapted for RF use. It is a purpose-built AI-native software developed using proprietary foundation models for power amplifier design, developed by Falcomm with the help of GF and engineered to capture electromagnetic physics of the metal stack, transistor nonlinearities, manufacturability rules, and design tradeoffs that define modern gallium nitride (GaN) circuits.

The model’s first supported process is GF’s RFGaN1 technology, validated through a deep technical collaboration with GF. That partnership matters. By training on rich, process-aware data developed alongside GF, GaNdalph.ai learns the actual behavior of real GaN transistors as they operate under heavy drive, which is the regime that governs real-world power amplifier performance and the regime that makes GaN design so demanding. Measured device physics, rather than generalized RF theory, is what enables the model to produce designs that hold up in silicon.

But the architecture is not bound to a single process. GaNdalph.ai is designed as a multi-node platform from day one, extensible to additional GF technologies including 45RF-SOI, NSX, and FDX™ FD-SOI, with a methodology that can be applied to GaN and CMOS processes across the industry.

Single-stage and Two-stage RF PAs, Fully Optimized

GaNdalph.ai supports both single-stage and two-stage RF power amplifier design, delivering optimized results automatically in minutes. This is where the technology becomes most compelling. Two-stage PA design involves an extraordinarily complex trade space: interstage matching networks, bias conditions, device sizing, output loading, and efficiency all interact simultaneously across dozens of coupled constraints. Navigating this design space requires deep expertise that takes years to accumulate.

GaNdalph.ai performs full optimization across this entire trade space, producing designs comparable to or exceeding those of expert RF engineers with years of GaN experience. The model does not simplify the problem or trade off one parameter for another in a crude way. It finds genuinely optimal solutions within the full complexity of the design challenge, and it does so in a reduced timeframe that a designer would typically require.

For companies needing high-performance RF chips, this changes the economics of GaN design entirely. Organizations that previously lacked the in-house expertise to fully leverage a GaN process node can now access competitive results without years of ramp up time. Experienced RF designers, in turn, gain a powerful tool for exploring a broader solution space far more quickly than was previously possible.

A Partnership That Sets the Standard

The quality of GaNdalph.ai’s initial release is a direct product of the collaboration behind it. GF contributed modeling expertise and process insight that helped shape how the model learns device behavior. Falcomm brought the architectural vision and AI training methodology needed to turn that knowledge into a working foundation model. The result is a tool that is genuinely native to advanced GaN design, with GF’s RFGaN1 process as the foundation.

This kind of partnership represents a new model for design enablement in the semiconductor industry, where deep process knowledge and machine learning combine to bring intelligent automation to the design cycle. The methodology generalizes. As GaNdalph.ai expands to additional processes and circuit types, the same AI-native tool development approach can be applied wherever the goal is faster, optimized, more accessible RF design.

What Comes Next

GaNdalph.ai is designed to grow. The platform’s architecture is extensible to additional GF process nodes including 45RF-SOI, NSX, and FDX, and the same approach is being applied to other RF and analog design challenges beyond power amplifiers, including a library of high-performance RF passives, low noise amplifiers, mixers, multiplexers, and more. The broader ambition is a suite of AI-native tools that cover the full scope of RF and mixed signal design across multiple semiconductor foundries and process nodes. Each new node and circuit type added to the platform extends what is possible and brings RF design within reach of Falcomm’s AI-native suite.

See It at IMS 2026

IMS 2026 runs June 7–12 in Boston, bringing together more than 8,000 RF and microwave professionals from over 65 countries. GF will be exhibiting at Booth #21018, while Falcomm will be showcasing its latest innovations at Booth #22082. Visit us on the show floor to see GaNdalph.ai in action and learn how AI-native design automation is transforming the future of RF GaN product development. We look forward to meeting you in Boston. To schedule a meeting or demo ahead of the show, contact us at info@falcomm.com.

Learn more about GaNdalph.ai at falcomm.com and explore GF’s RF GaN technologies at https://gf.com/technologies/rf/rf-gan

Explore Amplifier IP:

• 28GHz High-Efficiency CMOS Power Amplifier for 5G mmWave Applications
• 39GHz High-Efficiency CMOS Power Amplifier for 5G mmWave Applications