The centerpiece was Sao Luiz do Tapajos, a 7,608-meter-long dam that would have stood 53 meters above the Tapajos with a reservoir of 729 square kilometers. Thirty-eight Kaplan turbines would have generated a guaranteed 4,012 megawatts. Upstream, the Jatoba Dam would have added 2,338 megawatts more, with its own reservoir of 646 square kilometers. Three additional plants were planned on the Jamanxim River. The planners added two more to the wish list: Chacorao, a 3,336-megawatt project on the upper Tapajos, and Jardim do Ouro, a smaller 227-megawatt project on the Jamanxim. The National Waterways Plan listed the Chacorao locks as a priority, because the reservoir would eliminate the Chacorao rapids and let soybean barges run above the dam. A continuous chain of dams with locks would have wiped out every waterfall and rapid along hundreds of kilometers of river.

Planners knew the political problem from the start. Traditional hydroelectric construction builds access roads to remote dam sites, and those roads attract loggers, ranchers, and squatters who do far more damage to the forest than the dam itself. The proposed solution came out of sessions in 2004 and 2005 with Minister of the Environment Carlos Minc: the platform model. Build the dams the way offshore oil rigs are built. Move every worker, every bag of cement, every piece of machinery by barge or helicopter. Cut trees only at the dam sites, and replant after construction. The surrounding 200,000 square kilometers of conservation units, including Amazonia National Park and the Itaituba I, Itaituba II, and Jamanxim national forests, would stay intact. It sounded elegant on paper. It did not address the one thing the model could not solve, which was that the reservoirs would still flood indigenous land.

Around thirteen thousand Munduruku live along the Tapajos and its tributaries. The dams, if built, would have drowned parts of the Sawre Muybu, Munduruku, and Sai-Cinza indigenous territories. Government-funded studies recommended relocating people, a suggestion the Munduruku fiercely rejected. FUNAI, the federal indigenous affairs agency, stated that forced relocation would be unconstitutional under the 1988 Constitution, which protects indigenous land rights. Dam supporters argued the protection did not apply because Sawre Muybu had not yet been formally demarcated. The Munduruku pointed to the Tucurui Dam on the Tocantins, where people displaced thirty years earlier were still waiting for their promised compensation. The issue was never abstract for them. They had watched other Amazonian peoples be moved out of their homelands, and they knew what happened next.

In November 2012, federal police shot and killed Adenilson Munduruku in the region. The following April, 24 researchers entered the area under military and police protection to conduct surveys for the dam. The Federal Public Ministry filed to have Operation Tapajos suspended, arguing that the armed presence violated human rights and the Indigenous and Tribal Peoples Convention of 1989, which Brazil had signed. On April 16, 2013, a federal court in Brasilia ordered the government to pull back. Two months later, in June 2013, Munduruku opposing the surveys took three biologists working for Eletrobras hostage for two days. The researchers were released only after the federal government promised to suspend the studies and consult the affected communities. The work halted, then restarted, then halted again through 2013 and 2014.

On April 17, 2016, the Brazilian Chamber of Deputies voted to impeach President Dilma Rousseff. Two days later, on April 19, FUNAI published the identification study recognizing the Sawre Muybu Indigenous Territory. On the same day, IBAMA suspended environmental licensing for the Sao Luiz do Tapajos Dam. The timing could not have been coincidental no matter what the agencies said. With Sawre Muybu now formally recognized, the dam was unconstitutional, because the Constitution expressly forbids removing indigenous people from demarcated land. On July 28, 2016, the Federal Public Ministry formally recommended that IBAMA cancel the license. IBAMA had ten days. In August 2016, it announced the official cancellation of the Sao Luiz do Tapajos environmental license, and the largest hydroelectric project proposed for the Amazon in decades was dead.

Two independent studies had already undermined the economic case. A 2014 analysis concluded that electricity demand forecasts had been overstated and that improvements in energy efficiency and falling costs of solar and wind power were not reflected in the plan. Once transmission costs were included, the project was not cost-effective even before adding losses from fishing, tourism, and water quality degradation. A December 2015 study of the reservoir emissions found that the Cachoeira do Cai and Cachoeira dos Patos plants would probably generate methane at levels comparable to a natural gas power plant, and that Cachoeira do Cai might emit more carbon equivalent than a coal plant. Hydroelectric power in tropical flooded forests is not the clean energy source it was once assumed to be. The Munduruku had stopped the dams on constitutional grounds. The economics were never going to save them either.