Amazon is transforming a rural Indiana data center into a powerful AI hub
Amazon is transforming a large data center complex in rural Indiana into a powerful computing hub. This will help train its next frontier AI models. The redesign includes adding thousands of AI servers. These use Amazon's custom Trainium chips. Business Insider reports this exclusive information. The move supports Amazon's strategy to lead in AI development.
Internal planning documents were reviewed by Business Insider. They describe an initiative called AGI Pivot. This aims to combine multiple data centers into a single, efficient cluster. The goal is to support Amazon's AGI organization and in-house AI development. Despite job cuts in the AGI division, the company's frontier AI goals remain strong. The Nova model series has been discontinued, but Amazon continues its AI ambitions.
People close to the plans say the project is urgent. Amazon is working to expand computing capacity quickly. It needs to train a next major AI model before the end of the year. Demand for AI hardware is rising. The project will streamline operations. It aims to get the most from the company's big investments in AI infrastructure.
The Indiana data center is part of Project Rainier. This is a powerful AI supercomputer built using Trainium chips. It supports partner company Anthropic. The new AGI initiative will operate from the same site. It will not interfere with Rainier systems. Amazon says the site is key for both external and internal AI work.
Amazon is upgrading networking, storage, and fiber-optic systems. These upgrades will connect multiple data centers. They will work together as a unified computing system. Some buildings are being turned into annexes. They will share equipment with neighboring data centers. This will cut costs and improve reliability.
Amazon is increasing its AI spending. It now plans to spend $220 billion in 2026. That's up from a previous $200 billion target. Rising demand for AI hardware and component prices are driving this. Amazon is upgrading Trainium 2 systems with Trainium 3 servers. This keeps its infrastructure modern.
Business Insider has covered Amazon's AI overhauls before. Initiatives like Houdini and Titus are part of these efforts. They show how Amazon is reorganizing its AGI division. The new focus is on frontier AI models. Trainium chips are now central to Amazon's AI plans. The company projects Trainium and Graviton chips will earn over $25 billion annually. That's up from a previous estimate of $20 billion.
Planning documents mention technical needs for Amazon's next AI models. They call for more data storage. This is needed for multimodal AI training. The process involves handling many high-resolution images. It also needs frequent saves during training. This shows the complexity of modern AI development.
Amazon's AI work is moving forward despite recent AGI layoffs. The company is focused on speed, cost efficiency, and sustainability. These goals apply to the Indiana data center complex. Amazon expects good long-term returns on its AI investments. CEO Andy Jassy has commented on data centers' long-lasting value.
Amazon is significantly expanding its AI investment.
Have a tip?