Amazon.com Inc.’s audacious plan to build a data center complex of artificial intelligence (AI) infrastructure facilities on 1,200 acres of Indiana cropland isn’t just massive in size but outsized in ambition: to create an AI model as smart as the human brain.

As first reported by the New York Times, the e-tail behemoth so far has constructed seven data centers on site, with around 30 slated to be built, as part of startup Anthropic’s project to build a Claude AI model that is as powerful, complex, and intelligent as the human brain. The multibillion-dollar Project Rainier is so big that Amazon has selected four construction firms to get the complex finished as soon as possible.

“I don’t know if they’re competing for cash or steak dinner or what, but it’s crazy how much they’re getting up,” Bill Schalliol, a local economic development official, told the New York Times. “Steel starts to go up here, the next day steel’s going up over there.”

The gigantic undertaking in Indiana is the most breathtaking of a series of AI data center projects from Amazon, which plans another complex in Mississippi, as it jousts with rivals OpenAI and Meta Platforms Inc. Both those companies are working on facilities to develop more-advanced AI models.

A timeline on completion of Amazon’s project in Indiana remains uncertain, however, for a couple nettlesome reasons.

Local residents have protested the idea of a wetland being bulldozed over for data centers, which require vast amounts of energy and water. To help mitigate resources, Amazon is using less-advanced Trainium2 chips in the hopes it won’t need to deploy complex cooling equipment to lower the temperature in its facilities. Instead, Amazon plans to use large air handlers and municipal water to cool facilities.

Nonetheless, the Indiana complex is expected to chew up 2.2 gigawatts of electricity, enough to light up millions of residences, and will divert immense quantities of water for cooling.

Demand for data centers worldwide could grow by 240GW (some 240 billion gigawatts) between 2023 and 2030, with the U.S. accounting for 12% of all power generated, according to management consultant McKinsey.

Most data centers used to train AI models currently under construction are in the 100MW and 250MW (100 million to 250 million watts) range, and are much much larger than anything built before, clean energy entrepreneur Michael Liebreich recently shared in a BloombergNEF piece.

AI data centers projected to consume up to 9% of U.S. electricity by 2030, according to Electric Power Research Institute.