News › ai-infrastructure
By Zayden R., June 29, 2026
ZLUDA v6 has improved PhysX performance on AMD GPUs, but faces funding cuts from its latest backer. This open-source project continues to adapt, now focusing on multi-GPU CUDA for AI applications.
ZLUDA version 6 has been released, bringing notable improvements to PhysX performance on AMD GPUs. This open-source initiative, which started as a CUDA support project for Intel graphics, later expanded to AMD GPUs with funding assistance. However, recent developments have seen its commercial funding withdrawn once again. This is not the first time ZLUDA has faced financial hurdles. Initially funded by AMD, the project transitioned to open-source after losing that backing. A new, unnamed sponsor stepped in, allowing ZLUDA to pivot towards multi-GPU CUDA support with a focus on AI infrastructure. Unfortunately, this support has also been pulled. Despite these setbacks, ZLUDA's recent release demonstrates resilience and adaptability. The project's ability to enhance PhysX on AMD hardware is a significant technical achievement. Engineers who rely on AMD GPUs for AI and machine learning tasks will appreciate these improvements, even as the project's financial future remains uncertain. ZLUDA's journey reflects the broader challenges of sustaining open-source innovations in the competitive tech landscape. While its funding situation is precarious, its technical progress continues to make a mark in the AI and GPU communities.
The Linux Camp teaches these topics as hands-on labs on real virtual machines, verified as you type.