Xilinx Strengthens Adaptive Computing with Acquisition of Falcon Computing Solutions

Xilinx’s recent acquisition of Falcon Computing Solutions expands software programmability and adaptive computing for developers, leveraging Falcon’s compiler to generate custom, high‑performance FPGA and SoC‑accelerated applications.

Falcon Computing Solutions specializes in high‑level synthesis (HLS) compiler optimizations that translate software into efficient hardware acceleration. Xilinx will integrate this capability into the Vitis Unified Software Platform, offering automated, hardware‑aware optimizations that lower the barrier for developers.

By embedding Falcon’s compiler within Vitis, developers can accelerate C++ programs with little hardware knowledge. The source‑code transformations eliminate the need for manual code rewrites or architecture‑specific directives, delivering substantial acceleration automatically.

“The growing demand for adaptive computing is driving a new era of FPGA adoption in the data center and embedded applications,” said Salil Raje, EVP and GM of Xilinx’s Data Center Group. “Falcon Computing’s innovative compiler and expert team will advance software programmability and bring the benefits of adaptive computing to a wider developer base.”

Jason Cong, co‑founder and chairman of Falcon Computing, added, “Our compiler enables developers to achieve an order‑of‑magnitude speed‑up over CPUs with minimal FPGA expertise, thanks to automation of off‑chip data movement, on‑chip reuse, memory partitioning, and parallel/pipelined acceleration. The Open‑MP‑like single‑source style is highly approachable for C/C++ developers in HPC and embedded systems.”

Falcon’s flagship product, the Merlin Compiler, targets parallel workloads across heterogeneous hardware—including multi‑core CPUs, GPUs, and FPGAs. With a growing pool of software developers and data scientists, Merlin’s automated C/C++ to FPGA flow empowers them to harness specialized hardware without deep hardware expertise, bridging the gap between modern applications and heterogeneous platforms.

This marks the third Xilinx acquisition led by Jason Cong. Cong, UCLA’s Volgenau Chair for Engineering Excellence, director of the Center for Domain‑Specific Computing, ACM and IEEE fellow, and National Academy of Engineering member, co‑founded Falcon Computing in 2014. With deep academic roots, Cong has previously sold AutoESL (now Vitis HLS) to Xilinx in 2010 and Neptune Design Automation (now part of Vivado) in 2013. Falcon Computing, based in Los Angeles, serves enterprise clients and academic institutions throughout the U.S. and China.