Amazon’s Artificial General Intelligence group has eliminated an undisclosed number of roles, even as the company projects $200 billion in capital expenditure for 2026 and describes AI as “one of the most important things we’re working on.”
The layoffs hit teams working on model customization and post-training, overseen by VPs Adeeb Shanaa (AGI Data Services) and Vishal Sharma (AGI Information), according to Reuters and employee reports. Affected teams described cuts of approximately 10% on some teams.
In a statement, an Amazon spokesperson said the company is “sharpening our focus on the initiatives that matter most for customers, so we can move faster on what counts”, a standard restructuring rationale that sits uncomfortably alongside Amazon’s simultaneous proclamation that AI is its biggest bet.
The cuts are part of a broader workforce reduction that has seen Amazon shed more than 30,000 employees since October 2025, including 16,000 in January 2026 alone. Yet the company’s AI investment trajectory moves in the opposite direction. Amazon’s custom silicon business has become a $20 billion revenue stream, AWS is on track to double its data center capacity by the end of 2027, and the company continues developing its Nova family of foundation models.
The AGI organization now reports to Peter DeSantis, a longtime Amazon executive who took over in December 2025. The restructuring signals a shift in priorities within the AGI group: foundational model development and infrastructure investment continue at pace, but the post-training and customization layer, the work of adapting general-purpose models for specific enterprise needs, is being treated as lower priority.
The contradiction is not unique to Amazon. Across the technology industry, companies are pouring hundreds of billions into AI infrastructure while simultaneously reducing headcount in the teams building and maintaining those systems. Oracle cut 21,000 jobs earlier this month as it pivots toward AI. Microsoft gutted its commercial and Xbox divisions in the name of AI acceleration. The pattern suggests that the industry’s AI strategy treats infrastructure and people as substitutes rather than complements.
Sources: Amazon’s AGI department finds humans are optional (The Register, July 23, 2026)