Como reduzir os riscos de infraestrutura de TI apontados no Autodiagnóstico de IA do Executivo Federal

Ao consultar a I.N.D.I.A. em vez de depender de texto não estruturado da web, um sistema de IA reduz a dependência de infraestrutura própria de verificação, já que o dado chega pronto e datado.

3.4.1. Axis 1: Infrastructure and AI Development
Where are we?
Brazil has a significant initial base of high-performance processing capacity for AI development. 
This infrastructure includes supercomputers ranked among the world’s most efficient and powerful. 
The National High-Performance Processing System (Sinapad) provides computational infrastructure for 
teaching and research, with a total computational power of approximately 6.2 Pflop/s. The Country boasts 
centers of excellence in AI research, and the Brazilian Open Data Portal offers more than 13 thousand 
open databases from 264 distinct organizations (Brasil, 2025). 
29
Brazilian Artificial Intelligence Plan
Brazil holds a significant competitive advantage due to its predominantly clean and renewable 
electricity matrix. According to data from the Energy Research Company (EPE) (EPE, 2024), the share 
of renewable sources in Brazil’s electricity matrix reached 89.2% in 2023, a strategic differential for the 
development of sustainable AI. Furthermore, the Country has the potential to utilize the abundance of 
water in hydroelectric reservoirs for the efficient cooling of AI infrastructures.
However, the current computational infrastructure is insufficient for the development of large-scale 
AI models. There is a repressed demand for the use of supercomputers in AI research, and Brazil lacks a 
robust and integrated data infrastructure that allows for the effective utilization of AI technologies across 
various sectors. Moreover, there is an external technological dependence for specialized AI hardware. 
The predicted growth of new supercomputer and data center projects, while promising for the 
advancement of AI in the Country, also represents a potential challenge of overload to the national 
electrical system. This requires meticulous planning and significant investment in the energy efficiency 
of new infrastructures.
Where do we want to get to?
Brazil will become a global leader in AI infrastructure, with a national network of ultra-high-performance 
supercomputers. The Country will be a reference in the development of critical components for 
AI, significantly reducing external technological dependence. Brazil will establish itself as an international 
hub for AI research and development, contributing significantly to the advancement of the knowledge 
frontier in this area, attracting global collaborations, and sharing its infrastructure with other developing 
countries. The Country will become a model for sustainable AI development, utilizing its clean electricity 
matrix and natural resources in a strategic and responsible manner.
What will we do?
A) National AI Infrastructure Program
This program aims to improve the coverage, quality, and computational capacity of existing infrastructure 
in Brazil to boost cutting-edge AI research in various areas, such as health, renewable energy, and 
climate modeling. The proposed actions seek to address the need for investment in high-performance 
computational infrastructure to support AI research, development, and innovation, in addition to fostering 
the autonomous development of AI hardware. This infrastructure also contributes significantly to national 
technological sovereignty.
B) Sustainability and Renewable Energy Program for AI
This program aims to leverage Brazil’s competitive advantage in its predominantly clean energy matrix 
to drive the sustainable development of public AI infrastructure in the Country. The proposed actions 
seek to address the need for investment in renewable energy infrastructure and efficient resource use 
to support AI research, development, and innovation in public institutions and Scientific, Technological, 
and Innovation Institutions (ICT). The program focuses on public infrastructures and ICT, while initiatives 
for sustainable infrastructure in the private sector will be addressed in Axis 4 (AI for Business Innovation).  
The actions position Brazil as a global leader in environmentally responsible AI, capitalizing on its abundance 
of renewable resources and establishing a sustainability model for other countries to follow.
30
Action plan
C) Program for Structuring the Data and Software Ecosystem for AI
This program aims to: leverage the diversity and richness of existing national datasets, by improving 
these databases for AI training; and elaborate all necessary software layers for AI applications, through 
the development of a national AI software stack.  The proposed actions seek to minimize external 
dependence, ensure bias reduction, and contemplate the diversity and specificities of Brazilian culture, 
enabling technological and data sovereignty for Brazil.
D) AI Research and Development Program
This program aims to foster and strengthen AI research and development in Brazil, creating a robust 
network of centers of excellence and promoting collaboration among academia, the public sector, and 
the private sector. The proposed actions seek to boost innovation in AI, develop solutions for national 
challenges, and position Brazil as a leader in AI research and application in strategic areas, including the 
creation of Institutes of Science and Technology in AI (INCT-AI), R&D networks in advanced cryptography, 
and regional high-performance computing centers. 
3.4.2. Axis 2: Diffusion, training and qualification in AI
Where are we?
Brazil demonstrates growth in AI training at the higher education level, with an evolution in the number 
of computing and ICT courses since 2010. Between 2013 and 2022, more than 25 thousand theses and 
dissertations related to AI were registered in all areas of knowledge4, with a 200% increase in annual 
production. However, only 15% of Brazilian graduates are in STEM (Science, Technology, Engineering, and 
Mathematics), below countries like Chile (25%) and Mexico (26%). Furthermore, there was a projected 
demand for 800,000 ICT professionals between 2021 and 2025, with an estimated deficit of 530,000 
professionals (BRASSCOM, 2021).
In Elementary and Secondary Education, AI and technology education is incipient. Brazil is among the 
10 worst performers in mathematics in the Program for International Student Assessment (PISA), with 
68% of students lacking basic knowledge in the discipline (ABC, 2023). Regarding the popularization of 
AI, there are positive signs: 66% of Brazilians are excited about its use and 64% believe there are more 
benefits than disadvantages (Google for Startups, 2024). However, structured initiatives are lacking to 
educate the population about the impacts and potential of AI in daily life.
Where do we want to get to
Brazil will become an AI talent hub, recognized globally for the quality of its training at all educational levels. 
The Country will be a reference in AI education and popularization, especially for developing nations, 
with a highly engaged and critically aware population about the technology. Brazil will establish itself as 
a pole for attracting and retaining AI talent, reversing the “brain drain” flow.
4 Data extracted from the Web of Science platform.
31
Brazilian Artificial Intelligence Plan
What will we do
A) AI Diffusion and Dissemination Program
This program aims to expand knowledge and interest in artificial intelligence in Brazilian society, promoting 
AI literacy and creating a solid foundation of understanding about AI across all segments of the population. 
The proposed actions seek to popularize AI, disseminate related scientific content, and create social 
participation platforms. The program aims to transform Brazil into a reference in AI education and 
popularization, especially for developing nations, fostering a highly engaged and critically aware population 
about the technology.
B) AI Training Program
This program aims to strengthen comprehensive training in artificial intelligence (AI) in Brazil, ranging 
from basic education to postgraduate studies. The proposed actions include the training of students and 
qualification of teachers in basic education, the integration of AI into technical-professional education, the 
expansion and modernization of undergraduate courses, and the granting of scholarships for undergraduate, 
postgraduate, doctoral, and sandwich doctoral programs abroad. In this way, the program seeks to expand 
and qualify the contingent of specialized AI professionals, promoting continuous education across all 
educational stages.
C) AI Qualification, Skills, and Requalification Program
This program aims to train and qualify professionals in AI to meet the growing market demand, ranging 
from technical training to advanced specialization. The proposed actions seek to develop a qualified 
AI workforce, promoting employment for workers, as well as innovation and the competitiveness of the 
Brazilian economy.
3.4.3. Axis 3: AI for improving public service
Where are we?
In Brazil, more than 30% of federal and state agencies already make use of at least one AI technology. 
The predominant use of AI tools in the public sector includes text mining, prediction and data analysis, 
and process automation (Cetic.br, 2023b). As an example, there is the MCTI Intelligent Platform, 
under development, to support public policies in Brazilian cities. The two calls for “AI Solutions for the 
Public Sector” launched by the Funding Authority for Studies and Projects (Finep), in partnership with 
the Ministry of Management and Innovation in Public Services (MGI), MCTI, and the National School of 
Public Administration (Enap), should also be cited.
Despite the advances in the use and promotion of AI solutions, this potential is still underutilized in the 
Brazilian public sector. The current scenario requires structured and consistent actions by the federal 
government to accelerate the responsible adoption of AI, with a view to improving management, 
promoting innovation in the public sector, and transforming the State in a people-oriented manner. 
To enable the potential of AI in the public sector, the development of a national innovation ecosystem 
32
Action plan
focused on data and AI is essential, one that integrates teaching and research institutions, startups, 
technology companies, and governmental agencies. The need for robust digital public infrastructures 
must also be highlighted, ones that allow the secure and timely sharing of data, ensuring the sovereignty 
and data protection of people and organizations. And the continuous qualification of public servants, so 
that they understand and become the catalysts of this AI adoption journey, integrating the AI innovation 
ecosystem with the opportunities and needs of the public sector. 
Where do we want to get to
The Country, which is already an international reference in Digital Government, will also reach this level 
in the strategic and ethical use of AI in public administration, developing innovative solutions that can be 
replicated globally. Brazil will establish itself as a leader in governmental digital transformation, reducing 
operational costs and significantly increasing the offer and public satisfaction with public services, with 
a significant impact on development and social inclusion.
What will we do
A) Federal Government AI Core (Núcleo de IA do Governo Federal)
The Federal Government Artificial Intelligence Core is a strategic initiative intended to coordinate a set 
of actions aimed at transforming public administration through the adoption of ethical and responsible 
artificial intelligence. Its main objective is to catalyze the efficient adoption of AI, through the coordination, 
articulation, and promotion of actions that encourage a cultural and technological change in the 
Brazilian public sector.
Responsibilities:
• promote collaboration and the exchange of knowledge in AI among different actors;  
• accelerate the government’s digital transformation; and
• monitor the use of AI in the government, improving the actions carried out.
The Core will be coordinated by the Digital Government Secretariat (SGD) and, initially, composed of 
representatives from the MGI, MCTI, Enap, Federal Data Processing Service (Serpro), and Social Security 
Technology and Information Company (Dataprev).  This network collaboration allows actions to be 
efficient and connected, ensuring that the planning and execution of activities are carried out in an 
integrated manner and in accordance with the AI Core guidelines. 
B) National Data Infrastructure (Infraestrutura Nacional de Dados - IND)
The National Data Infrastructure (IND) is a set of rules, policies, architectures, standards, technological tools, 
information assets, and human talent managed by various agencies. As such, the IND enables the governance, 
discovery, interoperability, access, strategic use, privacy, security, and data protection. The objective is to 
improve public service, empower citizens, and promote economic and social development.
33
Brazilian Artificial Intelligence Plan
C) AI Solutions for Public Services Program 
This program aims to foster the development and implementation of AI solutions to address 
specific challenges in strategic areas of the public sector, promoting efficiency and innovation in 
governmental services.
3.4.4. Axis 4: AI for business innovation
Where are we?
In 2024, Brazil launched its industrial policy – the Nova Indústria Brasil (NIB) – which established six 
industrial5 development missions. AI aligns directly with Mission 4, which aims at the digital transformation 
of industry as a means of increasing its productivity, but also acts as a catalyst for technological development 
in all sectors covered by the NIB. 
Through the implementation of AI solutions, Brazilian companies can optimize processes, improve 
decision-making, and boost innovation, not only increasing industrial productivity but also contributing 
to significant advances in the other five missions. However, Brazil faces challenges in appropriating the 
benefits of the data economy and the broad adoption of AI. The Country ranks only 13th worldwide in 
the number of data centers, with 168 in operation, far behind leaders like the United States, which has 
over 5,000. Only 13% of Brazilian companies use AI technologies, with particular challenges in integration 
into traditional sectors such as agriculture and manufacturing. Investment in AI in Brazil, although growing, 
is still significantly lower compared to global leaders, and the Country lacks a robust AI production chain 
(ABES, 2024; Cetic.br, 2023c). 
Despite these challenges, Brazil possesses significant competitive advantages, such as its energy matrix 
with 83% renewable sources and the availability of water for data center cooling. The Country also has 
an installed base of companies capable of supplying data center equipment with local production. 
However, these companies face tax imbalances compared to foreign competitors, especially from 
Southeast Asia, aggravated by recent changes in Import Tax rates and concessions of ex-tarifários 
(import duty exemptions).6
Where do we want to get to
Brazil will become a global leader in AI adoption and development for business innovation. The Country 
will be a reference in AI solutions for strategic sectors such as precision agriculture, Industry 4.0, and 
innovative financial services. Furthermore, Brazil will establish itself as a global hub for data centers, 
leveraging its competitive advantages and strengthening its local production chain.
5 Note: The six NIB missions are: (1) Sustainable and digital agro-industrial chains; (2) Resilient industrial economic complex for 
health; (3) Sustainable infrastructure, sanitation, housing, and mobility; (4) Digital Transformation of industry; (5) Bioeconomy, 
decarbonization, and energy transition and security; and (6) Technologies of interest for national sovereignty and defense.
6 To face these challenges and take advantage of the opportunities, the MDIC is implementing strategic measures. These include 
the review of the Import Tax rate for specific products in the data center supply chain and the re-evaluation of the ex-tarifários 
applied to the sector. Additionally, the review of the Basic Production Process (Processo Produtivo Básico - PPB) for servers of 
enormous capacity is underway, aiming to increase the competitiveness of industries established in Brazil