Amazon has announced a significant transition within its artificial intelligence leadership team, signaling a pivotal moment in the company’s long-term pursuit of artificial general intelligence (AGI). Rohit Prasad, who has served as the head of AGI initiatives, is set to depart from the company next year. His exit coincides with what Chief Executive Officer Andy Jassy has described as an “inflection point” in Amazon’s technological evolution—a juncture at which the company’s research, development, and product strategies in AI are poised to advance into a more mature and strategically unified phase. Jassy shared this development publicly on Wednesday, framing it as both a natural consequence of Amazon’s rapid progress and a deliberate restructuring to strengthen the company’s innovation pipeline.
As part of this organizational shift, Peter DeSantis, currently a senior vice president at Amazon Web Services (AWS) and a veteran of the company’s infrastructure and systems architecture teams, will assume leadership of a newly established division. This expanded group will not only oversee the advancement of Amazon’s most sophisticated artificial intelligence models but also integrate initiatives in custom chip design and quantum computing—two domains increasingly critical to achieving scalable and efficient next-generation AI performance. By consolidating these efforts under DeSantis’s direction, Amazon aims to foster deeper cross-disciplinary collaboration between its hardware, cloud, and AI divisions, ensuring that foundational technologies evolve in tandem rather than in isolation.
Rohit Prasad’s tenure at Amazon has been marked by a decade of transformative contributions. Having joined the company in 2013, Prasad was instrumental in shaping the design, personality, and core intelligence of Alexa, Amazon’s flagship voice assistant. Under his stewardship, Alexa transitioned from a novel consumer product into a central component of the company’s broader AI ecosystem. Later, Prasad guided the launch of the Amazon Nova AI model series, which represented an important step toward more general and adaptive machine learning capabilities across Amazon’s product lines. His promotion in 2023 to lead the AGI team reflected both his technical acumen and his influence in driving the company’s strategic vision for artificial intelligence.
In public discussions, Prasad has expressed a cautious and critical stance toward the AI industry’s growing fixation on benchmark performance. In a recent interview with journalist Alex Heath—founder of Sources and an alum of *The Verge*—he stressed that many of the common evaluation tests used to compare AI models fail to capture true generalization and reasoning power. Describing current benchmark evaluations as increasingly unreliable, he remarked that these metrics have become noisy and disconnected from the real intellectual capabilities that artificial intelligence systems are beginning to demonstrate. This commentary underscored a philosophical divergence between Amazon’s long-term approach and the more metrics-driven narratives promoted by rivals such as OpenAI, Google, and Meta.
Despite Amazon’s ambitions, the company’s Nova 2 AI model family, launched earlier this month, has not yet reached the top of several industry leaderboards. Observers have noted that over recent years, Amazon’s pace of progress in generative AI research appeared slower than that of its closest competitors—companies like Microsoft and Google that have aggressively integrated advanced AI into their enterprise and consumer ecosystems. Furthermore, the highly anticipated transformation of Alexa into a fully AI-powered conversational system has encountered multiple delays, and according to user feedback and internal reporting, the assistant’s new features have yet to meet the heightened expectations Amazon originally set.
To reinvigorate the company’s momentum, Amazon is entrusting Peter DeSantis with the task of overseeing a significantly expanded AI organization, one designed to unify high-level research and engineering under a coherent strategic direction. DeSantis, who has devoted approximately three decades to Amazon—spanning roles in cloud infrastructure, large-scale systems optimization, and advanced computing—will lead the push toward developing new frontiers in machine intelligence. Complementing this leadership adjustment, renowned AI researcher Pieter Abbeel will take charge of the frontier model research team, a group dedicated to exploring the outermost boundaries of learning algorithms and general intelligence systems. By pairing DeSantis’s operational experience with Abbeel’s deep research expertise, Amazon appears to be constructing a leadership structure capable of accelerating both scientific discovery and practical application.
CEO Andy Jassy concluded his announcement with an optimistic reflection on Amazon’s trajectory. He emphasized that the organization now stands on a strong foundation built through years of investment in machine learning infrastructure and research, and that the company’s growing traction across various AI initiatives signals a moment ripe with opportunity. Jassy expressed confidence that under Peter DeSantis’s stewardship, the AGI and advanced computing teams will concentrate their efforts with renewed unity and focus, enabling Amazon to deliver meaningful and transformative AI capabilities that enhance experiences for both enterprise clients and everyday users. As Jassy succinctly put it, the company’s course forward is not just about catching up with industry leaders, but about shaping the next era of artificial intelligence development through depth, precision, and sustained innovation.
Sourse: https://www.theverge.com/news/846721/amazon-new-agi-leader-peter-desantis-rohit-prasad
