News

I recently collaborated with Rodrigo Ardissom on a text that revisits dependency theory through the lens of contemporary rare earth extraction in Brazil. Our piece examines how the classical debates between developmentalists and dependency theorists in the 1960s and 70s remain stubbornly relevant today, as Brazil finds itself trapped in a familiar pattern: vast geological reserves, minimal industrial capacity, and foreign control of strategic resources.

We argue that Brazil’s position in the global rare earth economy reveals a form of dependence that goes beyond what even the original dependency theorists could fully anticipate. It’s not just about mineral extraction anymore—it’s about the intersection of physical infrastructure (data centers, semiconductor manufacturing, submarine cables) and digital colonialism, where Latin America provides both raw materials and behavioral data, only to have both processed elsewhere and sold back as finished products and AI services.

You can find the original Portuguese here.

The full English translation follows below:

The Imaginaries of Dependence: Brazil in the Geopolitics of Rare Earths

Rodrigo Ardissom de Souza[*]

Douglas Carvalho Ribeiro[*]

At the „A Fair Future for All“ session of the G20 Summit in Johannesburg on November 22-23, 2025, President Lula articulated a strong defense of Brazilian sovereignty over critical minerals: „Countries with large concentrations of mineral reserves cannot be seen as mere suppliers,“ he declared, „while remaining on the margins of technological innovation.“ He continued: „What is at stake is not only who holds these resources, but who controls the knowledge and added value that derive from them. To talk about critical minerals is to talk about sovereignty.“¹ The speech touched on the right issues: control of knowledge, added value, technological autonomy. But between the rhetoric of sovereignty and the reality of Brazil’s insertion in the global rare earth economy, there is an abyss that contemporary developmentalism seems incapable of crossing. Brazil has 21 million metric tons of reserves, 23% of the global total and the second largest position in the world, after China. However, it produced only 20 metric tons in 2024, while China produced 270,000 tons.² The distance between geological potential and industrial capacity is the exact measure of dependence.

Invoking sovereignty seems, in this context, somewhat trivial: as a nation-state, one asserts control over a mineral resource present within its territory. When the government establishes the National Council for Mineral Policy, when it announces that Brazil will be a „partner in the global value chain,“ it is operating within a specific imaginary: that of a country that finally assumes control of its strategic resources. The problem is that this imaginary, as it materializes in concrete policies, does not break with the logic of dependence. It updates it. There is an echo here of old debates. In the 1960s and 1970s, developmentalists and dependency theorists disputed the terms of the Brazilian problem. For the former, the goal was to accelerate industrialization, complete import substitution, and strengthen the state as an inducer of growth. For the latter—Ruy Mauro Marini, Theotônio dos Santos, Vânia Bambirra—the question was structural: dependence would not be resolved by more development, because peripheral capitalist development itself reproduced subordination. Underdevelopment was not a stage; it was a function.

Theotônio dos Santos devoted special attention to the place of science and technology in this structure. In The Structure of Dependence, published in the American Economic Review in 1970, he articulated the canonical definition: „By dependence we mean a situation in which the economy of certain countries is conditioned by the development and expansion of another economy to which the former is subjected.“³ His typology identifies three historical forms: colonial dependence, financial-industrial dependence, and the „new technological-industrial dependence“ based on multinational corporations. Dependence was not merely financial or commercial. It was, above all, technological. And it had its own materiality: it was expressed in truncated infrastructures, incomplete production chains, knowledge that systematically accumulated outside national borders. Dependency theory taught us to identify structural mechanisms of value transfer between peripheral and central economies. The superexploitation of labor, formulated by Marini in Dialéctica de la dependencia, was one of these mechanisms.⁴ But there is another, equally operative: transfer through capital ownership. Even when extraction occurs on national territory, even when it generates local jobs and taxes, value migrates to foreign shareholders, to patent holders, to those who control the stages of technological transformation. Formal ownership of the subsoil does not imply real appropriation of the value generated.

Fernando Henrique Cardoso’s solution—gestated in critical dialogue with dependency theory but radicalized in governmental practice—was to assume subordinate integration as a possible horizon: modernize through openness, attract foreign capital, accept the productive specialization dictated by the global market.⁵ Curiously, contemporary neo-developmentalism, which presents itself as a rupture with the neoliberalism of the 1990s, operates largely on the same tracks. The discourse changed, the structure remained. Yet another dimension of dependence exists that the classic debates could not fully anticipate: digital infrastructure. Contemporary capitalism is increasingly mediated by platforms that control not the production of commodities in the classic sense, but access to digital territories where economic and social life takes place. Amazon, Google, Meta, and Microsoft are not just companies. They are infrastructures on which other companies, governments, and citizens depend to operate.

What some theorists call „technofeudalism“⁶ or „platform capitalism“⁷ designates this mutation: rent extraction through the control of mandatory passage points, rather than profit through the production of goods. Nick Couldry and Ulises Mejias defined data colonialism as „an emerging order for the appropriation of human life so that data can be continuously extracted from it for profit,“ arguing that historical appropriations of land, bodies, and natural resources find a mirror in the contemporary era of pervasive datafication.⁸ The relevant point for rare earths debates is that this digital economy, often described as immaterial, „the cloud,“ rests on a profoundly physical base. Data centers consumed 415 TWh of electricity in 2024, about 1.5% of global consumption, with projections to reach 945 TWh by 2030, equivalent to Japan’s total consumption. A ChatGPT query consumed, in 2023 estimates, approximately ten times more electricity than a Google search. Submarine cables carry more than 99% of international data traffic and are increasingly owned by the big tech companies themselves. Advanced semiconductors depend on extreme ultraviolet lithography monopolized by a single Dutch company, ASML, which holds 100% of the EUV equipment market with machines that cost between 170 and 380 million euros each and weigh approximately 180 tons.

At the heart of all this are critical minerals: neodymium and dysprosium for permanent magnets in servers, terbium for optical isolators, yttrium for semiconductor compounds, cerium for wafer polishing, europium, an element with no known substitute, for red phosphors in screens. China not only controls about 69% of global rare earth mining; it controls approximately 90% of refining and holds an effective monopoly, virtually 100%, over the processing of heavy rare earths. When Beijing imposed export controls on twelve of the seventeen rare earth elements in 2025, it was not just protecting natural resources: it was demonstrating the capacity to condition the functioning of the global digital economy. This is not just commercial advantage. It is infrastructural power.

Latin America experiences, in this context, a condition of double extraction. It provides both mineral raw materials—lithium from Chile and Argentina, copper from Chile and Peru, rare earths from Brazil—as well as behavioral data that feed the platforms. Brazil is the third-largest user base of ChatGPT in the world by absolute traffic, while the region hosts only 4.8% of global data centers. The structures are parallel: extract minerals and process them abroad to sell finished products back; extract data and process it on servers in the Global North to sell artificial intelligence services back. Dependence is updated, not overcome.

It is in this context that the case of the Serra Verde mine in Goiás becomes revealing. Since January 2024, it has made Brazil the only producer at commercial scale of all four magnetic rare earth elements—neodymium, praseodymium, dysprosium, and terbium—outside Asia. But majority ownership is foreign. Denham Capital has acted as a founding investor since 2010. Energy and Minerals Group and Vision Blue Resources, the latter a British company, jointly invested 150 million dollars in January 2023. In August 2025, the U.S. International Development Finance Corporation approved a loan of up to 465 million dollars for the project, one of its largest operations in critical minerals that year. The DFC is not private capital but rather the development finance arm of the U.S. government, explicitly created to advance American strategic interests. The deal is explicit: the Serra Verde mine will primarily supply rare earth elements to American and allied companies. U.S. strategic interests, not Brazilian industrial sovereignty, structure the financing of the mine. And Brazilian rare earths will be exported for processing elsewhere, since Brazil lacks domestic refining and metal production capacity.

The Brazilian state’s response is instructive. In October 2025, the government announced a partnership between SENAI and the German Fraunhofer Institute to develop supermagnets in Brazil, with an investment of R$35 million over five years, approximately US$6 million.⁹ That same month, it was reported that Brazilian demand for rare earths could grow sixfold by 2034, driven mainly by the demand for permanent magnets.¹⁰ Consider the numbers: the U.S. government invests US$465 million to secure the extraction of Brazilian rare earths and their subsequent export to American companies. The Brazilian government responds with US$6 million over five years for research. The asymmetry is not accidental; it is structural. This gap is not merely financial—it is epistemic. The capacity to dominate the rare earth value chain requires accumulated knowledge about separation, refining, alloying, magnet engineering. This knowledge is not bought ready-made on the market. It is constructed through research institutions, trained technical cadres, integrated industrial infrastructure. Brazil does not lack geological reserves. It lacks technological mastery.

This is not a new problem. In the 1960s and 1970s, Latin American thinkers like Amílcar Herrera, Jorge Sabato, and Oscar Varsavsky were already articulating a critical diagnosis of technological dependence.¹¹¹²¹³ Sabato and Botana proposed the famous „Sabato Triangle“: the articulation between the state, productive structure, and scientific-technological infrastructure as a condition for autonomous development.¹⁴ Their diagnosis was correct. Their political defeat was absolute. What was constructed in Brazil were islands of excellence—UNICAMP, USP, CBPF, the nuclear program—surrounded by a productive structure that never organically integrated domestic technological capacity. Science was produced. It was not systematically translated into industrial sovereignty.

The history of rare earths in Brazil illustrates this pattern with disturbing precision. Between 1945 and 1960, Brazil was a global reference in processing monazite sand, a primary source of rare earths. Orville Derby, geologist of the São Paulo Geological and Geographic Commission, documented placer deposits in Minas Gerais as early as 1893.¹⁵ Brazilian production from 1945 to 1960 reached 6,030 tons, representing 13.8% of global production.¹⁶ The Rare Earths Association of Brazil, founded in 1950, operated until 1960. Companies like Sociedade Anônima Comércio e Indústria de Terras Raras, Orquima, Nuclebras Monazita e Associados, Terras Raras do Brasil, and Indústria e Comércio de Minérios constituted a significant industrial network. In January 1953, Celso Pinto de Melo published tests on the chemical treatment of monazite for the recovery of cerium with 98% purity, rivaling the best results obtained in American laboratories at the time.

Brazil exported high-purity rare earth oxides for nuclear applications in the United States, contributing to the development of post-war atomic technology. The world’s first nuclear submarine, the USS Nautilus, was commissioned in September 1954. It was not just about extraction, but technical mastery over processing, precisely the link in the chain that now escapes us. This knowledge was systematically dismantled: by external geopolitical pressures interested in the radioactive minerals associated with rare earths, by a nationalization that reoriented the complex exclusively toward the nuclear program from 1960 onward, by the state’s chronic inability to protect the strategic value of this knowledge.

Rare earths are not rare in nature; they are rare in mastery of processing. China built its monopoly not because it has more ore—Brazil has the second-largest reserves in the world—but because it invested decades in research, training cadres, vertical integration of the production chain, from mine to permanent magnet. In January 1992, during the Southern Tour, Deng Xiaoping pronounced the phrase that remains engraved in stone at the entrance of the Baotou Inner Mongolia Pioneering Rare Earth Hi-Tech Zone: „The Middle East has oil, China has rare earths.“¹⁷ As Julie Klinger observes, it was a warning about how Middle Eastern resources provoked Western imperial intervention, not a declaration of a strategic plan for domination.¹⁸ The imaginary of sovereignty is potent, and precisely for this reason it can be perverse. When mobilized to legitimize mineral exploitation without corresponding industrialization, when it serves as an alibi for subordinate insertion in global value chains, it does not liberate; it imprisons. It becomes a justification for the reproduction of dependence under new discursive garb. The challenge is to recognize that sovereignty over natural resources, unaccompanied by technological and industrial autonomy, is a façade sovereignty, and that developmentalism that does not confront this contradiction is condemned to repeat, under new slogans, the old story of subordinate insertion. The pattern repeats with historical regularity: technical capacity built, opportunity lost, return to supplier status. Dependency theorists posed this question half a century ago. It remains, stubbornly, unanswered.

---

Footnotes

[*]: Professor and researcher at the Centro de Investigación y Docencia Económicas (CIDE), trained in Law and Social Sciences. Member of the National System of Researchers (SNI) – CONACyT. PhD in Anthropology specializing in Legal Anthropology, Digital Rights, and Internet Governance.

[*]: Teaching Assistant and Postdoctoral Researcher at the Chair for Science and Innovation Studies, University of Hamburg (Prof. Dr. Sabine Maasen). Fellow at the Max Planck Institute for Social Anthropology (Project „Complexity as a Topic of Law“). PhD in Law from the University of Hamburg.

Footnotes

1. BRASIL. Agência Brasil. G20: Lula defende governança soberana em minerais críticos e IA. Agência Brasil, Brasília, 23 nov. 2025. Disponível aqui. Acesso em: 11 jan. 2026. ↩
2. UNITED STATES GEOLOGICAL SURVEY. Mineral Commodity Summaries 2024: Rare Earths. Reston, VA: U.S. Geological Survey, 2024. Disponível aqui. Acesso em: 11 jan. 2026. ↩
3. DOS SANTOS, Theotônio. The Structure of Dependence. The American Economic Review, v. 60, n. 2, p. 231-236, maio 1970. ↩
4. MARINI, Ruy Mauro. Dialéctica de la dependencia. México: Ediciones Era, 1973. ↩
5. FILGUEIRAS, Luiz. O neoliberalismo no Brasil: estrutura, dinâmica e ajuste do modelo econômico. In: BASUALDO, Eduardo M.; ARCEO, Enrique (Org.). Neoliberalismo y sectores dominantes: tendencias globales y experiencias nacionales. Buenos Aires: CLACSO, 2006. p. 179-206. ↩
6. VAROUFAKIS, Yanis. Technofeudalism: What Killed Capitalism. London: Bodley Head, 2023. ↩
7. SRNICEK, Nick. Platform capitalism. Cambridge: Polity Press, 2016. ↩
8. COULDRY, Nick; MEJIAS, Ulises A. The costs of connection: how data is colonizing human life and appropriating it for capitalism. Stanford: Stanford University Press, 2019. ↩
9. SENAI. Brasil investe em pesquisa para dominar o ciclo produtivo das terras-raras e dos superímãs. Revista Pesquisa FAPESP, out. 2025. Disponível aqui. Acesso em: 11 jan. 2026. EMPRESA DE PESQUISA ENERGÉTICA. Terras raras: demanda brasileira deve crescer seis vezes até 2034. Poços Já, 28 jul. 2025. Disponível aqui. Acesso em: 11 jan. 2026. ↩
10. SENAI. Brasil investe em pesquisa para dominar o ciclo produtivo das terras-raras e dos superímãs. Revista Pesquisa FAPESP, out. 2025. Disponível aqui. Acesso em: 11 jan. 2026. ↩
11. HERRERA, Amílcar O. Ciencia y política en América Latina. México: Siglo XXI, 1971. ↩
12. SABATO, Jorge A. (Comp.). El pensamiento latinoamericano en la problemática ciencia-tecnología-desarrollo-dependencia. Buenos Aires: Paidós, 1975. ↩
13. VARSAVSKY, Oscar. Ciencia, política y cientificismo. Buenos Aires: Centro Editor de América Latina, 1969. ↩
14. SABATO, Jorge A.; BOTANA, Natalio. La ciencia y la tecnología en el desarrollo futuro de América Latina. Revista de la Integración, n. 3, p. 15-36, nov. 1968. ↩
15. MELO, Celso Pinto de. Terras raras e soberania: o elo invisível entre ciência, indústria e poder. A Terra é Redonda, 5 dez. 2025. Disponível aqui. Acesso em: 11 jan. 2026. ↩
16. SOUSA FILHO, Paulo C. de; SERRA, Osvaldo A. Antonio. Terras raras no Brasil: histórico, produção e perspectivas. Química Nova, v. 37, n. 4, p. 753-760, 2014. DOI: 10.5935/0100-4042.20140121. ↩
17. SAUNDERS, Harry. How a Deng Xiaoping quote became China watchers‘ favorite anecdote. Domino Theory, 12 nov. 2025. Disponível aqui. Acesso em: 11 jan. 2026. ↩
18. KLINGER, Julie Michelle. Rare earth frontiers: from terrestrial subsoils to lunar landscapes. Ithaca: Cornell University Press, 2017. ↩