Several internal AI projects are winding down as Amazon shifts resources. The company is prioritizing key models and new strategies. Teams are being reorganized to align with evolving customer needs.
Amazon is making significant changes to its artificial intelligence operations, winding down certain in-house AI models and reallocating engineering resources to new priorities. This move is designed to better align with customer requirements and streamline the company’s AI strategy, according to an Amazon spokesperson. The company confirmed that its AI portfolio is regularly updated to reflect which models are most valuable to clients, and that it continues to provide migration support as technologies evolve.
While some models are being phased out, Amazon is maintaining its investment in the Nova family of models, which remain central to many customer workflows. Nova Forge, a platform that enables businesses to build custom models using Nova as a foundation, will continue to receive support. The company stated that models in active production typically receive long-term backing, even as research and development efforts shift toward next-generation AI capabilities.
Business Insider reported that Amazon’s recent restructuring is aimed at concentrating engineering talent and limited computing resources on the company’s highest-priority AI initiatives, rather than spreading efforts across a broad range of text, image, and video models. As part of this shift, Amazon recently closed its AGI Lab, which was established in December 2024 to advance frontier AI models. The lab had been staffed by several dozen employees from the startup Adept, including its co-founder and CEO David Luan, who left Amazon in February. Over a dozen former Adept team members have also departed in recent months.
These changes coincide with layoffs in Amazon’s AGI unit, as first reported by Reuters. Leadership overseeing the AGI strategy and the lab’s operations have exited the company, signaling a step back from the race to develop frontier AI models. Industry observers note that large language models still face significant challenges before reaching artificial general intelligence.
Amazon’s evolving approach to AI comes as other publishers and platforms are also rethinking their strategies to engage audiences and diversify revenue. For example, The Hill recently introduced a premium subscription offering exclusive content and interactive features for its most dedicated readers, as detailed in this report on new membership initiatives.
