Amazon Research Awards Fall 2026

The Amazon Research Awards (ARA) Fall 2026 Call for Proposals is an international research funding initiative organised by Amazon Science.

The programme supports academic research and open-source contributions addressing important challenges in artificial intelligence, information security, computing and related technologies.

Amazon invites eligible faculty members and permanent researchers at qualifying organisations to propose innovative projects that could advance scientific understanding and offer practical applications.

Successful research teams may receive institutional funding and AWS Promotional Credits. Each selected project is also assigned an Amazon research contact for technical consultation and collaboration.

The Fall 2026 round covers nine research areas.

Available Research Areas

• Agentic AI
• AI for Information Security
• Amazon Security
• Automated Reasoning
• AWS Cryptography
• Build on Trainium: Agentic Recursive Improvement of Kernels and Frameworks
• Multilingual and Contextual AI Evaluation
• On-Device Multimodal Perception
• Synthetic Data Generation for Multimodal Perception

Each research area has its own detailed call, proposed topics, funding limits and submission requirements.

Applicants should identify the call that best matches their research expertise before preparing a proposal.

Nature of the Award

The programme provides research support through eligible academic institutions and organisations.

Awards are one-time unrestricted gifts, meaning the funding is administered by the recipient's institution.

Amazon does not automatically acquire intellectual property rights to the research outcomes through these grants.

Researchers are encouraged to publish their findings and make related code available through open-source repositories.

• Be a full-time faculty member at an accredited academic institution or a permanent researcher at a legally recognised non-governmental organisation.
• Be at or above the legal age of majority in their country or jurisdiction of residence.
• Submit no more than one proposal as principal investigator during the same call period.
• Not be an Amazon employee.
• Not have worked as an Amazon director, officer, employee, intern or contractor during the preceding 12 months.
• Not be an immediate family or household member of an ineligible recent Amazon worker.
• Not be a prohibited government employee under the programme rules.
• Not be subject to applicable US export controls or sanctions.
• Not have participated in or held decision-making authority over relevant cloud infrastructure procurement involving Amazon.
• Submit a proposal aligned with one of the eligible Fall 2026 research calls.
• Comply with any additional conditions specified for the selected research area.

International Eligibility

Qualified researchers from around the world may apply, subject to applicable legal restrictions.

The programme excludes applications from Cuba, Iran, Syria, North Korea and the specified restricted regions of Ukraine, along with other locations where participation is prohibited by law.

Co-Principal Investigators

One eligible co-principal investigator may be included where permitted. The co-PI must also satisfy the programme's professional eligibility requirements.

• Unrestricted institutional research funding.
• AWS Promotional Credits.
• An assigned Amazon research contact.
• Technical consultation and research collaboration opportunities.
• AWS tutorials and training resources.
• Opportunities to participate in Amazon research events.
• Hands-on sessions with Amazon scientists and engineers.
• Support for eligible graduate researchers or postdoctoral research costs through the institution's approved project budget.
• Opportunities to publish findings and contribute to open-source research.

Funding Examples

Funding ceilings vary across research calls.

Examples from the Fall 2026 round include:

Automated Reasoning

• Up to US$80,000 in unrestricted funds.
• Up to US$40,000 in AWS Promotional Credits.

AWS Cryptography

• Up to US$80,000 in unrestricted funds.
• Up to US$40,000 in AWS Promotional Credits.

Agentic AI

• Unrestricted funds averaging no more than US$70,000.
• AWS Promotional Credits averaging no more than US$50,000.

Multilingual and Contextual AI Evaluation

• Unrestricted funds averaging no more than US$50,000.
• AWS Promotional Credits averaging no more than US$50,000.

• Research proposal prepared according to the relevant call.
• Project title and research summary.
• Principal investigator's institutional and professional details.
• Research objectives and methodology.
• Proposed project outcomes and expected scientific contribution.
• Project budget in US dollars.
• Budget justification.
• Requested AWS services and Promotional Credits, where applicable.
• Relevant supporting information specified in the chosen research call.

Proposal Requirements

• Use the Amazon Research Awards proposal template where required.
• Individual research calls generally encourage proposals of no more than four pages, excluding appendices.
• Explain how the work matches the selected research topic.
• Describe the proposed approach and expected impact.
• Include any call-specific technical questions or requested supporting information.

• Quality of scientific content.
• Creativity and originality.
• Potential impact at scale.
• Relevance to the selected research topic.
• Scientific contribution to the research community.
• Practical relevance to Amazon-related research challenges.
• Feasibility and technical credibility, where specified by the particular call.

• Collaborate with researchers working on advanced technologies.
• Develop research with potential real-world applications.
• Access cloud computing resources for eligible projects.
• Advance research in AI, security and computing.
• Strengthen institutional research partnerships.
• Support graduate students or postdoctoral researchers through research projects.
• Contribute to publications and open-source development.
• Build connections between academic and industry research.