Microsoft has launched two new data halls in São Paulo, part of the company's $2.7bn commitment to artificial intelligence (AI) and cloud infrastructure in Brazil.

Revealed during the Microsoft AI Tour in Brazil by country president Priscyla Laham, the first two data halls are "in operation," while construction of additional units is progressing according to schedule.

It is unclear whether this means new buildings or additional rooms within existing facilities.

Laham further noted that all external areas of the site have been completed.

The exact location of the data centers, or their size and capacity, has not been disclosed.

During her talk, Laham said: "Microsoft believes that artificial intelligence can be a great asset in making Brazil more competitive, increasing productivity and innovation, and placing the country in a prominent global role."

Laham has held the leadership role since the start of 2025, just a few months after Microsoft CEO Satya Nadella said the company would spend 14.7 billion Reais ($2.7 billion) in cloud and AI infrastructure by the end of 2027.

During a question-and-answer session at the AI Tour, Laham said that the recently introduced Bill 278, which would create a Special Tax Regime for Data Center Services, known as Redata, was not "changing what we agreed with Brazil."

The bill aims to encourage data centers and their associated production chain to enter the country. It will also offer exemptions from PIS/Pasep, Cofins, and IPI on equipment purchases for AI data centers.

Microsoft already has two Azure regions in the country: Brazil South, based in São Paulo state and launched in 2014, and Brazil Southeast, based in Rio de Janeiro.

Microsoft launched the São Paulo Azure cloud region in Campinas and expanded to three availability zones in 2021. It was later revealed the company was working on additional facilities in the Hortolândia and Sumaré areas of São Paulo state.

The Rio region is now flagged as having restricted access "to support specific customer scenarios, such as disaster recovery within a specific geographic area."