Back Research Notes Brazil and the Green Compute Arbitrage of the AI Decade Published on February 2, 2026 By Jordi Visser Executive Thesis Brazil is no longer best understood as a cyclical commodity exporter leveraged to global growth. That framing materially understates the structural shift now underway. Brazil is emerging as one of the most strategically advantaged geographies of the AI physical era, a country where energy abundance, mineral leverage, institutional depth, and a rare monetary setup converge to create a durable, long-horizon alpha opportunity. As the United States and Europe increasingly collide with power-grid bottlenecks, multi-year interconnection queues, rising marginal electricity costs, and political friction around new infrastructure, Brazil is quietly moving in the opposite direction. Real rates are falling from restrictive levels, renewable generation capacity is scaling well beyond near-term domestic demand growth, critical minerals essential to AI hardware are being unlocked, and multi-gigawatt data-center campuses are being authorized at a scale that is increasingly unfeasible in OECD power markets. The result is what can be described as Green Compute arbitrage: the conversion of surplus renewable energy and mineral endowment into globally tradable AI workloads. In this framework, Brazil is not merely a supplier to the AI buildout. It becomes a host economy for AI itself, capturing value not only from inputs, but from compute, infrastructure, and downstream digital activity. What follows is grounded in physical constraints, capital structure, and macro dynamics, but also informed by firsthand experience living and working in Brazil, experience that shaped my conviction that the country’s long-term strengths are often underestimated by outside investors. Whenever I reflect on my career, there are many moments of luck along the path to who I am today. But no moment mattered more than moving to Brazil and running an office there less than five years out of college. I love the people, the country, the intensity, and the ambition that permeates Brazilian life. Those three years permanently altered how I think about emerging markets, institutions, and long-term potential. I begin this paper there because it is rare to have the opportunity to write about a country you know personally at the precise moment when its structural advantages begin to align. I wrote a brief paper on Brazil in June 2025. Since then, the global physical AI upgrade cycle has accelerated materially. As compute, power, and capital constraints move from abstract concepts to binding limits, it is time to go deeper into why Brazil may be one of the most important and misunderstood beneficiaries of the AI era. The Monetary Pivot: Real Rates, Inflation Psychology, and AI Deflation Brazil has historically been “first in, first out” of global inflation cycles, and the current regime is no exception. After front-loading monetary credibility through aggressive tightening that drove policy rates to restrictive levels, Brazil now sits at the threshold of one of the most meaningful real-rate declines in the investable universe. Consensus expectations point to policy rates falling materially over the next 12–24 months, while inflation expectations remain relatively well anchored. This places Brazil among a small and shrinking group of countries offering macro stability, positive real yields, and a clear easing trajectory. This monetary setup matters disproportionately for AI and energy infrastructure. Data centers, transmission lines, rare-earth processing facilities, grid-scale storage, and power generation are long-duration, capital-intensive assets whose valuations are acutely sensitive to real discount rates. Even modest declines in real rates can materially lower the cost of capital, expand feasible project pipelines, and unlock equity re-rating potential across infrastructure-linked assets. An underappreciated layer of this setup is the interaction between inflation psychology and the deflationary impulse of artificial intelligence. I believe AI is the most powerful deflationary force the global economy has ever encountered. Outside of the physical commodities required for power generation, transmission, and compute infrastructure, AI is a deeply democratizing force, compressing costs, lowering barriers to entry, accelerating efficiency, and relentlessly reducing the marginal cost of production across services, software, logistics, and knowledge work. This deflationary impulse stands in sharp contrast to the inflation psychology that still dominates policymaking and investor behavior in Brazil. The country’s collective memory of hyperinflation remains deeply embedded in institutional decision-making, public discourse, and the risk premia demanded by investors. As a result, Brazilian rates continue to embed a meaningful inflation premium, one that reflects historical trauma as much as forward-looking fundamentals. In an AI-accelerating world, this asymmetry matters. If AI-driven productivity gains suppress broad-based inflation over time while commodity-linked inflation remains localized to energy, metals, and infrastructure, the countries that have already priced inflation fear into their discount rates stand to benefit disproportionately. Brazil enters this transition with real yields that already compensate investors for inflation risks that may prove structurally overstated in an AI-driven economy. This creates a powerful and unusual setup: declining real rates layered on top of an inflation premium that was built for a different era. For long-duration AI, energy, and infrastructure assets, this dynamic amplifies the convexity of Brazil’s monetary pivot. The cost of capital falls not because inflation fears disappear, but because realized inflation increasingly diverges from the psychology embedded in rates. In that sense, Brazil’s inflation history, long viewed as a structural handicap, may quietly become a competitive advantage. Breaking the Rare-Earth Monopoly: Brazil as a Strategic Mineral Hedge Artificial intelligence is often framed as a software revolution. In reality, it is a mineral-intensive industrial transformation. From GPUs and advanced cooling systems to humanoid robotics, precision motors, and automated logistics, the AI stack is fundamentally built on magnet metals, neodymium, praseodymium, dysprosium, and terbium. Today, these inputs remain overwhelmingly concentrated in China, creating a structural vulnerability in the global AI supply chain. Brazil holds the world’s second-largest rare-earth reserves and, after decades of underdevelopment, is now transitioning from geological potential to industrial relevance. Projects such as Serra Verde and Viridis Mining’s Colossus deposit in Minas Gerais mark a shift from optionality to execution. Colossus, in particular, has demonstrated world-leading ionic recoveries of heavy magnet rare earths and cleared key regulatory milestones, positioning it as one of the most significant non-Chinese resources identified to date. Execution risk remains, particularly around downstream processing, environmental oversight, and time-to-scale, but the strategic importance is clear. Every humanoid robot, industrial actuator, high-efficiency pump, and cooling system inside an AI data center is ultimately a magnet story. Brazil’s emergence as a non-Chinese supplier offers hyperscalers, OEMs, and sovereign buyers a geographically diversified hedge against supply-chain concentration risk. Equally important, Brazil is no longer content to export raw material. With several billion dollars of rare-earth-linked infrastructure now being unlocked, the country is moving downstream into alloys, magnet processing, and advanced materialsvcapturing a greater share of AI’s industrial surplus. Power-Shoring AI: Brazil as a Global Green Battery The most compelling pillar of the Brazil thesis is energy. AI training is energy-inelastic.