Amazon Web Services (AWS) has opened applications for the third cohort of the AWS Generative AI Accelerator (GAIA), an eight-week global program designed to scale early-stage startups building foundational generative AI technologies.

Forty promising startups from around the world will be selected for this year’s program, which kicks off on 13 October 2025, at Amazon HQ1 in Seattle.

Applications are open until July 10, and the selected cohort will be announced on 24 September 2025. Startups can apply here.

With generative AI tools now widely available, AWS is focusing efforts in 2025 to support those startups who are developing generative AI technologies, including building models, infrastructure, fine-tuning tools, and agentic workflows that will drive the next wave of innovation.

The program aims to accelerate these teams with up to $1-million in AWS credits, technical guidance and mentorship, go-to-market support, and access to AWS’s generative AI tech stack.

“We are now at a stage where virtually all startups will be applying generative AI to their business in one shape or form. That’s why for this year’s accelerator, we are honing our focus to support those startups developing the foundational technologies that will define what’s possible with AI,” says Jon Jones, vice-president and global head of startups at AWS.

“This year’s program is part of our continued commitment to accelerate generative AI innovation around the world by providing ground-breaking startups with the credits, mentorship, and visibility they need to scale with confidence.”

The AWS Generative AI Accelerator is an eight-week hybrid program, offering virtual sessions between the in-person launch at the Amazon HQ in Seattle and conclusion of the program at AWS re:Invent 2025.

Startups building large language models, infrastructure tooling, fine-tuning platforms, or foundational agents are especially encouraged to apply.

Companies should have a functioning Minimum Viable Product (MVP), some customer traction, and a strong technical team. Prior AWS experience is not required, but those already building on AWS will benefit most from hands-on architecture guidance and marketplace enablement.