Brazil will invest about 2.3 billion reais ($444.2 million) to bolster its artificial intelligence ecosystem, splitting major projects between United States and Chinese technology firms in a strategic balancing act between the world's two AI superpowers.
The plan, announced by President Luiz Inacio Lula da Silva's government on Thursday, was reported by Reuters and underscores Brazil's effort to maintain working ties with both Washington and Beijing while building sovereign AI infrastructure. It is the latest example of national AI strategy reshaping global infrastructure spending, a trend we track in our daily AI policy coverage.
A two-track strategy
The investment is divided into two flagship projects. Just over half of the total — 1.3 billion reais ($251 million) — will fund a supercomputing infrastructure project in Rio de Janeiro developed in partnership with China's Huawei Technologies and iFlytek, the government said. That infrastructure will be used primarily to develop large language models for general and sector-specific applications.
Separately, about 1 billion reais ($193.1 million) will be allocated through a tender for a supercomputer that Brazil expects to rank among the world's ten most powerful AI processing machines. The system is to be installed in the northeastern state of Rio Grande do Norte, a region chosen for its energy potential. Lula attended an announcement ceremony in the state on Thursday.
Unnamed government officials told Reuters they expect US chipmaker Nvidia to win the supercomputer tender. Science and Technology Minister Luciana Santos told the Folha de S.Paulo newspaper last week that she also anticipated Nvidia would be the supplier.
'Not to depend on a single company, technology or country'
The deliberate split between Chinese and American suppliers is the core of the strategy, according to the government itself.
"The strategy is not to depend on a single company, technology or country," Lula's administration said in a statement, adding that the investments are aimed at strengthening national sovereignty over data.
The geopolitical context makes the calculation clear. China, a leading player in AI, has expanded its role as Brazil's largest trading partner. The United States, meanwhile, remains the biggest source of foreign direct investment in Latin America's largest economy, even as it has lost market share in trade and recently imposed additional tariffs on Brazilian goods.
By awarding the Rio de Janeiro LLM infrastructure project to Huawei and iFlytek while steering the flagship supercomputer tender toward Nvidia, Brazil is hedging its bets — securing access to both ecosystems without formally aligning with either government's technology bloc.
What each project is for
The two efforts have distinct missions. The Rio de Janeiro infrastructure built with Huawei and iFlytek is oriented toward model development: the government says it will be used primarily to train large language models for general purposes and for specific sectors of the Brazilian economy. For a country whose businesses have long relied on models built primarily on English-language data, domestic training infrastructure for Portuguese-language and domain-specific models is the central prize.
The Rio Grande do Norte supercomputer, by contrast, is a bid for raw ranking power. Brazil expects the machine to place among the ten most powerful AI processing systems in the world — an ambition that would put Latin America's largest economy into the same conversation as the elite AI clusters of the United States and China. The state's selection for the installation reflects a hard practical constraint of frontier AI: energy. Data centers of this class consume enormous amounts of electricity, and siting them where power is abundant and affordable is increasingly the deciding factor in where AI infrastructure gets built.
Timeline and funding
The investments will be funded by Brazil's National Fund for Scientific and Technological Development (FNDCT) through phased disbursements — a funding structure that spreads the fiscal load over time and ties spending to project milestones.
The government expects the supercomputer to begin operating by the end of next year, while the cooperation agreement with the Chinese companies is scheduled to start in July 2027. That means Brazil's sovereign AI capability will arrive in stages: the Nvidia-supplied machine first, with the Chinese-partnered LLM infrastructure following later.
A pattern among middle powers
The two-track approach mirrors moves by other middle powers seeking AI capability without exclusive dependence on US or Chinese suppliers. Gulf states have courted investments from both American labs and Chinese hardware vendors; Southeast Asian nations have signed overlapping AI agreements with Washington and Beijing. Brazil's bet is among the largest such commitments in Latin America to date, and it is being watched as a test case for whether non-aligned AI infrastructure strategies can work in practice.
Why it matters
For Brazil, the stakes are both economic and strategic. A top-ten AI supercomputer would give Brazilian researchers and companies domestic training and inference capacity, reducing reliance on foreign cloud providers. The Huawei partnership, meanwhile, gives the country a second pole of AI capability at a moment when US export controls increasingly shape who can buy which chips.
The project also tests whether a balancing strategy can survive intensifying pressure from both sides. The tender outcome — particularly whether Nvidia indeed wins and delivers under US export rules — will be watched closely as an indicator of how far middle powers can go in playing both camps.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
