Important

AI & Technology

Amazon cuts AGI research jobs while committing $200B to AI infrastructure

Last Updated: July 24, 2026

Antoine

July 23, 2026

5 min

Summarize with AI

CEO at Nova Analytics

Antoine founded Nova Analytics to empower Amazon sellers with enterprise-grade analytics. He specializes in data architecture and building scalable solutions for e-commerce businesses.

Quick Summary

•Amazon confirms layoffs in the AGI research team, refocusing on Nova, Rufus, Alexa+, Seller Assistant
•Reported $200B AI infrastructure commitment continues in parallel
•Seller Central AI feature cadence expected to accelerate, not slow
•Action: turn on every free AI surface Amazon ships, keep an independent data + AI layer via MCP for measurement

Nova surfaces every Amazon fee, refund, and margin shift in your live P&L, across 21 marketplaces. Open the live P&L

The 3-line summary

Amazon confirmed layoffs in its Artificial General Intelligence group this week, cutting an undisclosed slice of the team while continuing to pour a reported $200B into AI infrastructure over the coming years (TechTimes, July 23, 2026).

Amazon framed the reduction as a refocus of the AGI team onto its highest-conviction bets: Nova foundation models, agentic shopping (Rufus, Alexa+, Seller Assistant) and internal AI infrastructure (Storyboard18, July 23, 2026). The Deep Dive corroborated the layoffs from a second source the same day (The Deep Dive, July 23, 2026).

What sellers should read from it

The cut is not a retreat from AI, it is a concentration bet. The initiatives that survive are the seller-visible ones: Rufus in the shopper journey, Alexa+ in agentic checkout, Seller Assistant in the seller console, and Nova models powering all three. That means the pace of AI-driven Seller Central changes is about to accelerate, not slow down.

The parallel $200B capex signals that Amazon believes the durable moat is compute plus proprietary retail data, not headcount on the research side. Sellers should plan for another 18 months where Amazon ships an AI feature in Seller Central roughly every quarter.

The 30-day plan

Turn on every free AI surface Amazon ships this year. Seller Assistant, the new Canvas experience, Rufus-facing catalog fields. Free distribution matters more than any single feature quality.
Own the data layer outside Seller Central. As Amazon adds AI-generated summaries and recommendations, the sellers who win are the ones with an independent view of unit economics and daily performance.
Set a quarterly review cadence to test each new AI feature against a control cohort of ASINs, so you have your own data on what actually moves conversion for your catalog.

How Nova helps

Connect Claude to your Amazon data - keep a seller-owned AI layer running against your Seller Central data via the Nova MCP connection, independent of whatever Amazon ships next.
Day-to-Day Analytics - measure every new AI feature against a control cohort with daily conversion, sessions and buy-box data.

Frequently Asked Questions

Common questions about this topic

No. The AGI layoffs sit alongside a reported $200B AI infrastructure commitment over the coming years. The cuts concentrate the team on Amazon's highest-conviction bets: Nova foundation models, Rufus, Alexa+, Seller Assistant and internal AI infrastructure. Fewer research bets, more product execution.

Expect an accelerated Seller Central AI release cadence: roughly one AI feature per quarter over the next 18 months. The advantage goes to sellers who turn on every free AI surface Amazon ships and keep an independent view of unit economics, so they can measure whether each feature actually moves the needle for their catalog.

Both. Amazon's tools are free distribution and worth using. Owning a seller-controlled data and AI layer, e.g. by connecting Claude or another agent to your Seller Central data via an MCP connection, protects against sudden feature changes and gives you a consistent analytical view that Amazon's console does not.

Verified Sources

All information verified from official Amazon sources and trusted industry analysts as of publication date.