RESOLVING THE LEDGER
This is one outlet's own report from Business Insider — the article as it was filed. Other outlets are covering the same event; open the full story to compare every source side by side.
See the full story · 1 sourcesAmazon CEO Andy Jassy Bloomberg/Getty Images
Amazon is upgrading its Indiana data center campus to support its AGI team's frontier model effort.
Amazon continues to invest in its own frontier models despite last month's AGI job cuts.
Amazon increased its 2026 AI capex to $220 billion, reflecting rising computing demand and prices.
Amazon is redesigning part of a massive AI data center campus in rural Indiana into a sprawling cluster of powerful computers to build its next frontier AI models, Business Insider has exclusively learned.
Internal planning documents reviewed by Business Insider describe an effort to consolidate multiple data centers and deploy thousands of Trainium -powered AI servers.
According to people familiar with the matter, the effort is part of a broader initiative called "AGI Pivot" supporting the company's AGI organization and its future in-house AI models. These people asked not to be identified discussing private plans.
The documents suggest Amazon's frontier-model ambitions remain intact despite recent job cuts in its AGI organization and the wind-down of the previous Nova model series. The documents describe faster deployment schedules intended to provide enough computing capacity for the AGI organization to train its next big model before the end of this year, underscoring the urgency amid industrywide capacity constraints.
Maximizing returns
The redesign also reflects Amazon's broader effort to maximize returns on its record AI infrastructure investments. By linking existing data centers into a larger, more efficient AI cluster, the company is making better use of infrastructure it has already paid for and operates.
The approach aligns with CEO Andy Jassy's comment last week that AI infrastructure should generate attractive long-term returns because data centers remain productive for decades while servers and networking equipment can be refreshed over time.
Working beside Anthropic
The Indiana campus includes facilities that are part of Project Rainier, the Trainium-powered AI supercomputer Amazon built primarily for Anthropic. The new AGI initiative is run out of the same huge data center complex, however, it will not affect the existing Project Rainier servers.
"We're always designing, upgrading, and improving our data center infrastructure to serve the diverse needs of our customers and teams," an Amazon spokesperson told Business Insider. "That work improves speed, cost, and sustainability so we can meet growing demand for compute to train and serve models on AWS."
The AGI SuperCluster
The planning documents suggest Amazon is racing to expand computing capacity for its next frontier AI model push.
One update describes an "emergent request" to deploy more than 6,000 Trainium-powered AI servers, speeding up launch schedules by several weeks. The accelerated timeline was intended to prepare Amazon's next AI model for this year's re:Invent conference, typically held in early December, according to this document.
The documents also describe a broader effort to consolidate computing across the Indiana campus into what Amazon calls an AGI SuperCluster, creating a "larger, more efficient" AI system to maximize performance.
To do that, Amazon is redesigning networking, storage, and fiber-optic infrastructure so multiple data centers function as one.
The project also converts some existing buildings into what Amazon calls "annexes," connecting them to neighboring data centers so they can share core networking equipment instead of operating independently. Some facilities are planning to replace older Trainium 2 systems with newer Trainium 3 servers.
Doubling down on frontier AI
Amazon is significantly expanding its AI investment. The company recently raised its projected 2026 capital expenditures to $220 billion from $200 billion as demand for AI computing continues to outstrip available capacity and prices for some components rise.
The latest infrastructure effort fits a broader pattern inside Amazon. Business Insider previously reported on the company's push to upgrade its AI infrastructure through projects such as Houdini and Titus , while reorganizing its AGI division around a new frontier-model effort.
The initiative underscores the growing importance of Trainium inside Amazon. Last week, the company said its custom chip business, including Trainium AI chips and Graviton processors, is on pace to generate more than $25 billion in annual revenue, up from last quarter's $20 billion projection.
The planning documents also offer a glimpse of what Amazon's next frontier-model effort will require. They call for expanding data storage to support multimodal AI training, which requires processing large numbers of high-resolution images and repeatedly saving the model's progress.
One person familiar with the effort said the push has "not slowed down," despite last month's layoffs in the AGI organization.
Have a tip? Contact this reporter via email at [email protected] or Signal, Telegram, or WhatsApp at 650-942-3061. Use a personal email address, a nonwork WiFi network, and a nonwork device; here's our guide to sharing information securely .
Read the original article on Business Insider
AIPROPX is an independent multi-source news index — we track, compare, and connect coverage from across the web into one place you won't find anywhere else.