As Brazil increases its investment in artificial intelligence infrastructure, Marcelo Borgheti, Managing Director at Vistra Brazil, considers why converting data centres into a broader AI economy requires governments to focus on operating environments as much as megawatts.
With the global race for technology leadership gathering pace, Brazil is stepping up its investment in the infrastructure needed to compete in AI. Its government recently announced approximately 2.3 billion reais (US$444.2 million) in AI-related infrastructure investment, including 1.3 billion reais for supercomputing infrastructure in Rio de Janeiro and a separate public procurement, valued at around 1 billion reais, for a high-performance AI supercomputer in Rio Grande do Norte. Both projects sit within the wider Brazilian Artificial Intelligence Plan, which places advanced computing infrastructure at the heart of the country’s AI ambitions.
The choice of Rio Grande do Norte partly reflects the state’s energy potential, a fundamental advantage available to several markets across Latin America. IRENA’s 2025 South America outlook describes the region as a global leader in clean energy, supported by substantial hydropower, wind and solar resources. Combined with expanding digital demand, it gives Latin America a strong position in the global race for AI infrastructure.
That potential is only useful if projects can actually get off the ground. Data centres need reliable power, grid connections, water, suitable land and the right permits, and problems in any one of those areas can quickly delay development.
Even when that infrastructure is successfully delivered, the economic opportunity goes much further than the facility itself. A new data centre or supercomputer may add computing capacity, but its wider impact depends on what grows around it. Technology companies need to be able to set up and expand, specialist suppliers need access to the market, and skilled workers need a reason to stay. That is what turns infrastructure investment into a broader AI economy.
The Operational Realities of AI Infrastructure
Reliable, green power and digital connectivity are essential foundations for data-centre projects. However, international companies assess these physical assets as part of a much broader operating environment. Foreign investors evaluating Latin American markets analyse the speed with which capital can be deployed, licenses secured, and operational units activated. Administrative overhead and regulatory ambiguity often convert physical time-to-market advantages into prolonged structural delays.
The companies building and supporting AI infrastructure face challenges that go well beyond simply entering a new market. High-density computing relies on specialized supply chains, including hardware vendors importing advanced GPUs under complex customs rules, engineering firms handling specialized cooling infrastructure, cloud providers deploying multi-region architectures, and technology companies hiring cross-border machine learning talent.
Equipment suppliers often need local entities and bank accounts before they can contract, while engineering teams may need immigration, payroll, and employment support. Service providers also need accounting, tax registrations and reporting processes that can keep pace with growth. A major data-centre investment can attract dozens of related businesses, but the economic multiplier depends on how easily those firms can move from interest to operation.
For these businesses, moving from an investment decision to an operational reality requires establishing local entities efficiently, structuring compliant multi-jurisdictional payroll, navigating digital service tax regulations, and managing ongoing reporting without administrative uncertainty. When administrative processes are slow or unpredictable, the momentum of an entire tech ecosystem can stall. Delays in entity formation push back recruitment and local contracting, while tax and regulatory complexity adds ongoing management burden.
Navigating a Fragmented Regional Market
Converting physical infrastructure into a broader AI economy is particularly complex in Latin America because companies operate across a collection of distinct jurisdictions rather than a single uniform market. Brazil, Chile, Mexico, Colombia, and other regional markets each maintain different legal frameworks, employment laws, tax regimes, and regulatory requirements.
The same AI supply chain can therefore face a different operating model each time it crosses a border. A company expanding from Brazil into Chile or Colombia may need to revisit entity structures, employment arrangements, tax registrations, statutory reporting and governance processes rather than simply replicating its existing setup. Regional growth can become administratively intensive long before a business reaches significant scale, especially when different local advisers, systems and reporting cycles are involved.
For international technology providers and their supply chain partners, managing this fragmented landscape through separate, country-by-country administrative providers creates operational friction and compliance risk.
Expanding businesses need an integrated operational infrastructure, particularly one that pairs on-the-ground local expertise in each jurisdiction with global governance and compliance standards across the whole operation.
Whether establishing an initial local presence, managing ongoing tax and cross-border payroll, or coordinating compliance across multiple regional entities, companies require consistency, visibility, and accountability throughout their lifecycle. Reducing these administrative hurdles allows technology firms to focus on scaling their core operations rather than navigating local bureaucracy.
A Broader Focus for Governments
The true economic benefit of the AI boom emerges when capital commitments evolve into functioning, expanding operations to employ local talent, engage domestic suppliers, and drive secondary investment.
For governments across Latin America, the focus must therefore extend beyond securing power commitments and breaking ground on physical assets. This reflects a wider challenge identified by ECLAC, whose 2026 foreign direct investment report calls for investment attraction to be more closely aligned with productive development. For AI, this means paying equal attention to the day-to-day operating environment
Clear incorporation processes, predictable tax administration, workable employment rules, efficient immigration procedures and consistent reporting requirements all influence how quickly an international company can begin operating. These may attract less attention than a new data centre or supercomputer, but they help determine whether the companies around those assets can hire, sign contracts and continue investing. Markets that reduce unnecessary uncertainty are better placed to capture more of the economic activity created around major infrastructure projects, rather than competing primarily on subsidies or energy costs.
Latin America has a credible opportunity to position itself as a global hub for AI infrastructure, backed by undeniable advantages in energy and market demand. But global capital remains selective.
The markets that ultimately succeed will be those that combine their energy and location advantages with regulatory predictability, clear tax and compliance frameworks, and an efficient operating environment.
Attracting investment for a data centre is an important first step, but a thriving AI economy will only take root if international businesses and their surrounding ecosystems can establish, hire, comply, and scale with confidence.