industry news
Subscribe Now

Curtiss-Wright Introduces New PacStar 431 Edge AI Module Accelerated by NVIDIA IGX Thor

PacStar 400-Series-based small-form-factor solution delivers real-time AI inferencing and sensor fusion at the edge
ASHBURN, VA. – Oct. 28, 2025 – Curtiss-Wright today announced the release of PacStar® 431, a next-generation edge AI system engineered to meet the growing demand for high-performance real-time sensor data processing at the edge. Designed to deliver advanced AI/ML capabilities in a compact, low-SWaP footprint, PacStar 431 is powered by NVIDIA IGX Thor™ technology, enabling high-speed sensor ingest, AI reasoning, inference, and autonomous decision-making.
“PacStar 431 delivers the real-time processing performance our customers need to unlock AI capabilities in challenging edge environments, enabling faster, more informed decisions,” said Roark McDonald, General Manager, PacStar. “By integrating NVIDIA IGX Thor into a low-SWaP, ruggedized module, we’re accelerating the deployment of advanced processing solutions that support critical national infrastructure and mission-essential applications, ensuring our technology stays ahead of rapidly evolving global demands.”
Built with a powerful NVIDIA Blackwell GPU, Arm® CPU with integrated high-speed DDR5X memory, Gen 5 NVMe storage, and 25 GbE networking, PacStar 431 delivers up to 2070 TFLOPS of AI performance – 8x more than the previous-generation IGX Orin 500 – within a compact 5.3″ x 7.1″ x 1.7″ PacStar 400-Series enclosure. It supports real-time inferencing, sensor fusion, and autonomous processing, enabling new AI-driven capabilities previously unattainable at the edge.
Key Features of PacStar 431 include:
  • Accelerated by NVIDIA IGX T5000 SOM with 2070 TFLOPS of AI compute performance
  • 14-core Arm Neoverse V3AE CPU optimized for real-time AI and machine learning workloads
  • Highest speed memory and storage ever in a PacStar 400-Series module (128GB DDR5X, Gen 5 NVMe)
  • Multiple 25 GbE network interfaces for high-speed sensor ingest and connectivity
  • Seamless integration with PacStar 400-Series ecosystem, enabling modular, scalable solutions
Curtiss-Wright delivers rugged, long-lifecycle access to the latest AI infrastructure while enabling customers to leverage the full breadth of the NVIDIA AI Enterprise ecosystem. This collaboration helps customers accelerate deployment, streamline software integration, and ensure long-term product continuity aligned with NVIDIA’s technology roadmap.
PacStar 431 further expands Curtiss-Wright’s portfolio of NVIDIA-enabled processing solutions, which span multiple architectures and form factors to support a wide range of edge AI applications. With these offerings, Curtiss-Wright enables customers to bring advanced AI capabilities closer to the point of need, improving responsiveness and effectiveness across complex, data-driven environments.
For more information, visit http://www.curtisswrightds.com and LinkedIn.
About Curtiss-Wright
Curtiss-Wright is a global integrated business that provides highly engineered products, solutions and services mainly to Aerospace & Defense markets, as well as critical technologies in demanding Commercial Nuclear Power, Process and Industrial markets. We leverage a workforce of approximately 9,000 highly skilled employees who develop, design and build what we believe are the best engineered solutions to the markets we serve. Building on the heritage of Glenn Curtiss and the Wright brothers, Curtiss-Wright has a long tradition of providing innovative solutions through trusted customer relationships. For more information, visit www.curtisswright.com.

Leave a Reply

featured blogs
Mar 20, 2026
From machines that see and think, to systems that act, and the humans that nudge them along....

featured video

Cadence Chiplets Solutions | Helping you realize your chiplet ambitions

Sponsored by Cadence Design Systems

In this webinar, David Glasco, VP of Compute Solutions at Cadence, discusses how Cadence enables customers to transition from traditional monolithic SoC architectures to modular, scalable chiplet-based solutions, essential for meeting the growing demands of physical AI applications and high-performance computing.

Read eBook: Helping You Realize Your Chiplet Ambitions

featured chalk talk

Global Coverage With NTN
In this episode of Chalk Talk, Paul Fadlovich from TE Connectivity and Martin Lesund from Nordic Semiconductor and Amelia Dalton explore the what, why and how of NTN technology. They also explore the role that antennas play in satellite communication systems, and how Nordic Semiconductor’s nRF9151 System-in-Package and TE Connectivity’s broad range of antenna solutions can jump start your next global IoT design.
Feb 19, 2026
26,892 views