The leading cloud providers are vying to capture the rapidly growing AI sector, each developing specialized AI chips to support the increasing demand for artificial intelligence. Google’s Trillium and Microsoft’s Maia are two such advancements, while Amazon Web Services (AWS) has its own lineup, including the Trainium, Inferentia, and Graviton chips. To promote Trainium in particular, AWS has launched a grant program designed to support AI research with its technology.
Called Build on Trainium, the program will distribute $110 million in Trainium credits and grants to AI researchers, institutions, and students. Specifically, AWS plans to allocate up to $11 million in credits to select universities with which it partners, while individual grants for AI projects in the wider research community may go up to $500,000. Additionally, AWS is establishing a research cluster of 40,000 Trainium chips that research teams can access independently.
Gadi Hutt, senior director at AWS’ Annapurna Labs, said this initiative aims to provide AI researchers with the necessary hardware resources for their work. In addition to funding, grant recipients will receive access to Trainium-related educational resources.
The scarcity of computing resources in academia contrasts with the vast infrastructure available to tech giants. For instance, while Meta uses over 100,000 AI chips for its models, Stanford University’s Natural Language Processing Group has only 68 GPUs available. This limited access can slow the pace of academic research in AI, and AWS hopes to bridge some of that gap.
However, the program has sparked mixed reactions. Some researchers, such as Ph.D. candidate Os Keyes from the University of Washington, express concerns that corporate-backed grants might influence the direction of academic research. The selection process for Build on Trainium remains vague, with AWS stating that funds will be allocated based on research merit and project needs. An AWS spokesperson clarified that a committee of AI experts would review proposals to identify promising research projects with the potential to advance machine learning.
Corporate funding of AI research has often favored projects with commercial applications. Studies show that major tech companies focus less on examining AI’s ethical implications and more on research that aligns with their business goals. Some researchers advocate for safeguards to ensure unbiased exploration of AI, free from concerns about corporate influence.
While Build on Trainium promotes Trainium technology, AWS has confirmed that grant recipients are not required to remain exclusive to its platform. AWS asks only that research outcomes be published and shared with the broader community.
Efforts to address the funding gap between academia and industry have been made, including a $140 million investment by the National Science Foundation to create university-led AI research institutes focused on fields like climate change and education. There is also a proposed $2.6 billion U.S. National AI Research Resource to provide researchers with computational resources and datasets. Yet, these public initiatives still pale in comparison to the scale of private sector funding.
With significant portions of Ph.D. graduates in AI moving into the private sector for better access to computing resources and data, most of the largest AI models now originate from industry. The academic community is seeing a growing number of AI research projects co-authored by industry professionals, underscoring the influence of corporate funding in shaping AI research today.
