Amazon has launched a $110 million initiative aimed at providing university researchers with access to up to 40,000 of its Trainium AI accelerators. This effort, known as ‘Build on Trainium,’ seeks to foster innovation in AI by enabling academics to develop new algorithms and enhance accelerator performance. The project offers cloud-based compute hours to facilitate large-scale AI experiments. Participants are expected to open-source their advancements, benefiting the broader machine learning community. While Trainium technology is primarily designed for Amazon’s efficient internal operations, this program underscores the company’s investment in cultivating external expertise. By providing technical support and educational resources, Amazon, in collaboration with the Neuron Data Science community, aims to equip researchers with the skills needed to leverage their technology effectively. This initiative aligns with Amazon’s strategy to crowdsource software development and enhance the utility of its custom silicon.