Anthropology of development

From Truman to Pax Silica

How development returned as technological security

14 July 2026·Martín González Senosiain

Illustration of an Andean salt flat connected through an industrial chain to batteries, microchips, data centres and a digital cloud, while a person and several llamas remain beside the water.

Illustration created for AIthropology Lab.

Factual information reviewed up to 11 July 2026.

Note on sources

Pax Silica is a rapidly expanding initiative. The composition of the coalition, the projects announced and their legal frameworks may change within a matter of weeks. This essay distinguishes between three levels of evidence.

The first consists of official documents, such as the founding declaration of December 2025. The second brings together verified journalistic reporting on accessions, negotiations and projects. The third is interpretative, connecting those materials with the anthropology of development, political ecology and studies of technological infrastructures. At this final level, the essay applies discourse analysis in the critical tradition of development studies. It does not treat official vocabulary as ornament, but as a practice that classifies territories, defines deficiencies, allocates positions and legitimises particular forms of intervention.

Whenever a claim is an inference rather than a demonstrated fact, the text states this explicitly. Price figures are identified by product, market and date because there is no single universal quotation for ‘lithium’. Where possible, the comparison of public revenue per tonne standardises quantities as tonnes of lithium carbonate equivalent and distinguishes between royalties, contractual rents and the gross revenue of state-owned companies. These magnitudes are not legally identical, nor do they by themselves represent total fiscal revenue.

In June 2026, Argentina and Chile joined Pax Silica alongside other countries from Latin America, Europe and Central Asia. The initiative, promoted by the United States Department of State, thus grew to 24 participants around an architecture encompassing critical minerals, energy, refining, advanced manufacturing, semiconductors, data centres, foundation models and artificial-intelligence platforms. The expansion took place at a new summit in Washington, where the European Union, Germany, Greece, the Netherlands, Costa Rica, El Salvador, Kazakhstan and Panama also joined (Financial Times, 2026).

Pax Silica does not present itself as a military alliance or a conventional trade agreement. Its founding declaration defines it as a partnership intended to build secure, innovative and prosperous technology supply chains among economies deemed trustworthy. The document proposes co-ordination of investment, infrastructure, incentives, protection of sensitive technologies and access to a stack that begins with energy and mineral processing and culminates in software, frontier models and applications (Australian Department of Industry, Science and Resources, 2025).

The name already contains a theory of world order. Pax evokes the formulas through which imperial powers have described as peace the stability guaranteed by their own superiority. Silica relocates that genealogy to the material substrate of computation. If the Pax Americana was associated with military power, the dollar, oil, financial institutions and control of trade routes, Pax Silica announces an order organised around computing, chips, energy, minerals and data.

Yet Pax Silica does not emerge in a historical vacuum. Its promise of prosperity, mutual security and technical collaboration reactivates a much older political grammar: the grammar of development that the United States turned into a global project after the Second World War.

This does not mean that Pax Silica is a literal repetition of Harry Truman’s Point Four, a new World Bank or an agricultural-extension programme disguised as artificial intelligence. The institutions, technologies and immediate objectives are different. The similarity lies in a deeper operation.

First, a deficiency is identified that can threaten both those said to suffer from it and the international order. Technical knowledge, investment and productive reorganisation are then presented as necessary remedies. Finally, different functions are allocated within an architecture whose centre retains an unequal capacity to set the rules.

In 1949, that deficiency was called underdevelopment; in 2026, it is called technological supply-chain insecurity.

Peace begins by naming a deficiency

On 20 January 1949, Harry Truman delivered the inaugural address that Arturo Escobar would later place at the beginning of the age of development. In the section known as Point Four, Truman grouped much of humanity under the label ‘underdeveloped areas’. He described their economies as primitive and stagnant and presented their poverty not only as a tragedy for their populations but also as a danger to prosperous regions.

The United States, he argued, possessed the technical and scientific knowledge needed to relieve that suffering. Greater production would be the key to peace and prosperity. A more vigorous application of modern science would transform living conditions and bring peoples regarded as backward closer to the model of industrial societies (Truman, 1949).

The speech did more than describe an economic difference. It constructed a political relationship between those who possessed knowledge and those represented as lacking it. The United States was not merely a society positioned higher on a shared scale. It was also the actor entitled to identify the problem, formulate its solution and determine the direction others should follow.

Escobar notes that this formulation turned industrialisation, urbanisation, agricultural modernisation, increased material production and the adoption of modern values into components of a universal trajectory. Capital, science and technology were expected to reproduce throughout the rest of the world the characteristics of societies considered advanced (Escobar, 2011).

The category of underdevelopment therefore achieved something remarkable. It transformed highly diverse histories, territories and ways of life into variations of the same deficiency. Peasant communities, newly independent countries, societies devastated by colonial exploitation and economies with institutions of their own could all be gathered into a single temporal position. They appeared to lag behind those who had already reached the future.

Escobar describes this process as a colonisation of reality. Development ceased to be one proposal among others and became the framework through which problems could be described, knowledge about them produced and interventions justified. Even those who challenged its dominant versions tended to demand another development — participatory, socialist, human or sustainable. The need to develop itself remained beyond dispute.

Pax Silica does not use the term underdevelopment. Its vocabulary is one of resilience, traceability, capacity, economic security, infrastructure and trusted partners. The shift matters, yet the operation retains a recognisable resemblance.

Countries are no longer ranked only by their level of industrialisation. They are also classified according to their position within the technological ecosystem. Some possess minerals. Others provide energy, finance, manufacturing capacity, logistics routes or markets. A much smaller group controls chip design, the most sophisticated manufacturing equipment, cloud platforms, foundation models, patents and the mechanisms capable of restricting their circulation.

The contemporary deficiency is no longer simply a failure to develop. It consists of dependence on suppliers defined as adversarial, a lack of infrastructure for AI, or exclusion from a chain designated as secure.

Nor is the mission still called development. It is presented as de-risking, alignment and trusted interdependence.

Two Cold Wars and the same technical promise

Truman’s speech did not merely follow a devastating war. It was delivered at the beginning of the Cold War, when the United States needed to offer countries emerging from colonialism a vision of the future capable of competing with the Soviet project.

Wolfgang Sachs argues that development supplied a reassuring image of a world order in which the United States naturally occupied first place. For decades it also functioned as a weapon in the competition between political systems. The promise of modernisation, investment and technical assistance was linked to the struggle for the allegiance of new states (Sachs, 2010).

Pax Silica emerges in another setting of confrontation. There is no exact reproduction of the bipolar rivalry between the United States and the Soviet Union, but there is an increasingly structured technological competition centred on China. Semiconductor production, mineral processing, computing capacity and access to advanced models are no longer treated merely as commercial matters. They have become components of national security.

As a conceptual mirror, some analyses use the expression Pax Sinica to describe a sphere of influence organised around Beijing. The term requires caution. It is not the official name of a Chinese coalition equivalent to Pax Silica. Its most precise material counterpart is the Digital Silk Road, an extension of the Belt and Road Initiative linking telecommunications networks, fibre-optic cables, data centres, cloud services and technical standards. Pax Silica and the Digital Silk Road are not symmetrical structures, but they show how digital infrastructure has become a terrain on which geopolitical blocs are formed (Hillman, 2021).

The Pax Silica declaration speaks of protecting sensitive technologies and critical infrastructure from access, influence or control deemed improper. It also calls for co-ordination against non-market practices, overcapacity and unfair competition. Although China is not named in every line, the architecture is being built in relation to the US objective of reducing dependence on Chinese industrial and technological supply chains (Australian Department of Industry, Science and Resources, 2025).

The historical analogy should not lead us to claim that Truman and Jacob Helberg are delivering the same speech. Truman offered a universal path to modernisation. Pax Silica organises a selective network among economies already occupying very different positions.

In both moments, however, technology appears as an instrument of peace and as the foundation of a geopolitical hierarchy.

Truman promised that science, capital and production would integrate underdeveloped areas into a world of abundance. Pax Silica maintains that technological specialisation and secure chains will preserve prosperity in a fragmented world.

Truman offered development to secure allies during the Cold War; Pax Silica offers economic security to organise them during a technological cold war.

From the Pax Americana to secured globalisation

A decisive transformation took place between those two moments. Post-war development coexisted with the consolidation of the Pax Americana, an order sustained by international financial institutions, military alliances, trade liberalisation and a profound US capacity to define the rules of the world economy. By the end of the twentieth century, that order was presented less as a succession of national economies moving through modernisation than as a global production network.

Global value chains distributed different stages of production according to the costs, capacities and advantages of each territory. Manufacturing could move, raw materials could come from several continents and companies could co-ordinate extensive production processes without owning every link directly. Hierarchy did not disappear; it was reorganised. Design, finance, intellectual property and brands often remained concentrated, while extraction and manufacturing shifted towards territories where costs were lower or regulation less demanding (Gereffi, 2018).

Pax Silica does not completely break with that model. It preserves the international division of functions but changes the criterion governing it. Cost is no longer sufficient. Origin, strategic alignment and security now determine which dependencies are acceptable.

Globalisation does not end. It becomes geo-fenced.

The aim is not to produce everything in the United States. It is to ensure that different components circulate within a trusted space, under co-ordinated mechanisms of investment, export control, traceability and infrastructure protection.

This is one of the most important differences from classical development. Pax Silica does not promise that every participant will build the entire technology stack. Its declaration argues precisely the opposite. The complexity of the AI economy requires the mobilisation of complementary capacities held by different companies and territories.

The political issue is what complementarity means when some capacities can be replaced easily and others generate accumulating dependencies.

When risk replaced wealth

Sachs proposed calling the period opened by Truman the ‘age of development’. In his account, development was not merely an economic policy. It operated as a perception that shaped reality, a myth that oriented states and societies, and a promise capable of justifying effort and sacrifice.

Over time, its limits became visible. Industrial societies could hardly continue to present themselves as a universal destination when their consumption of energy and materials could not be extended across the entire planet. Nor did the distance between those leading the race and those seeking to catch up narrow as promised.

Yet the exhaustion of the development promise did not mean the disappearance of its institutions or its way of representing the world. Sachs identified an especially revealing mutation: prevention replaced progress and the distribution of risk displaced the distribution of wealth.

Development apparatuses no longer promised with the same conviction that everyone would attain industrial abundance. They increasingly concentrated on containing migration, wars, disasters and threats. They continued to identify deficiencies and design interventions, but the horizon had changed.

Pax Silica can be read as an advanced expression of that transformation. It does not promise to turn every participant into a replica of the United States, nor that each will master the entire digital economy equally. It proposes a network of differentiated functions within a shared security regime, in which each territory participates through what it possesses or can provide — minerals, energy, land, ports, finance, manufacturing, talent, markets or regulatory capacity. Equality is replaced by complementarity, autonomy by selected interdependence and universal development by differentiated integration into a protected stack.

From the official perspective, this division allows all parties to benefit. Resource-rich or industrial economies would gain investment and technology. Centres controlling software, design and finance would secure more reliable supplies. Diversification would reduce vulnerability to disruption or political pressure.

Some of those benefits may materialise. An anthropological critique does not need to deny them.

What matters is what is actually distributed when functions are distributed. Are knowledge, intellectual property, decision-making capacity and the power to change the rules distributed as well? Or are production responsibilities and material risks the principal things shared?

Participating territories are no longer necessarily promised arrival at the centre. They are offered protection against the risks of remaining outside its architecture.

The chain that erases politics

James Ferguson showed that development functions as an interpretative framework capable of making very different situations intelligible through the same deficiency. In his classic study of Lesotho, international institutions constructed the country as a particular object of knowledge and designed their interventions around that representation.

The issue was not only whether the diagnosis was true or false. Even a flawed description could produce real political effects, including the expansion of state bureaucratic capacity and the transformation of territorial relations.

Ferguson called development an ‘anti-politics machine’ because of its capacity to translate historical conflicts and power relations into the apparently neutral language of planning. Political decisions were presented as technical consequences required to solve problems (Ferguson, 1994).

The concept of the supply chain performs a similar operation on a planetary scale. Viewed as a diagram, the artificial-intelligence economy appears as an orderly and antiseptic sequence. Energy and minerals stand at the beginning. Refining, advanced manufacturing, semiconductors, data centres and models follow. Platforms and applications reaching public administrations, businesses and users sit at the end.

The Pax Silica chain is therefore a technology of ordering. By translating geopolitical control over a deposit into ‘resilience’, the concentration of intellectual property into ‘specialisation’, regulatory subordination into ‘interoperability’ and alignment against China into ‘trust’, it does more than change the vocabulary. It defines which conflicts may appear as political and which are displaced into the category of externality, operational risk or logistical necessity. Land dispossession, water crises and the unequal distribution of value are encoded in protocols, indicators and contracts.

The supply chain does not eliminate politics. It reorganises politics spatially, inscribes it in technical procedures and then presents it as logistics.

The chain that stabilises hierarchy

The metaphor does not merely conceal hierarchy. It can also stabilise it.

Not every link has the same degree of replaceability. A territory supplying lithium, copper, nickel or energy competes with other territories able to provide similar resources. A new mine, a chemical innovation or a change in prices can reduce its importance.

Those controlling a platform, model, cloud architecture or technical standard occupy a different position. Every new institution adopting that system can increase its centrality. Data are adapted to its formats, workforces are trained in its tools and production processes are reorganised around its operation. Leaving it entails financial, organisational and political costs.

Whereas a mineral resource can be replaced by another supplier, digital infrastructure tends to generate cumulative dependency.

That dependency may be doubled when platforms used to map deposits, optimise pumping, co-ordinate transport or administer contracts store a territory’s operational data in proprietary clouds. The supplying country no longer provides only matter and energy. It may also provide telemetry, geological models, logistical patterns and knowledge of how its own infrastructures operate. The provider’s interfaces, formats and workflows then turn that local knowledge into an architecture that is difficult to migrate. The Pax Silica declaration does not require members to contract a particular cloud or platform, so this is an analytical risk rather than a demonstrated effect. The concern is sufficiently real for the European Data Act to impose obligations on portability, interoperability and switching between data-processing services, with the explicit aim of reducing vendor lock-in (European Union, 2023).

This difference breaks the appearance of a chain composed of equivalent links. Replaceability declines as one moves from matter towards intellectual property, design and standards.

The Pax Silica declaration promises trusted partners access to the full stack. But access to a technology does not amount to control over it. Using a model does not mean taking part in its governance. Buying chips does not confer the power to decide who may purchase the next generation. Hosting a data centre does not grant sovereignty over the software running inside it.

Complementarity can produce genuine co-operation. It can also turn a hierarchical division of labour into an architecture that is harder to leave.

The positive-sum consensus

David Mosse helps explain how such policies maintain legitimacy without requiring us to imagine a perfectly co-ordinated conspiracy.

According to Mosse, development must continually renew its moral language. When terms such as progress, modernisation or aid lose credibility, new paradigms emerge around participation, collaboration, citizenship, democracy or good governance. The field retains an insistently positive orientation towards the future. Every failure can be interpreted as the result of incomplete design, deficient implementation or the need for a better framework (Mosse, 2004).

Policies do not simply direct practice. Participating institutions and people also work to maintain coherent representations of what a project is and what benefits it produces. That coherence allows divergent interests to be assembled, funding to be secured and alliances to endure even when reality departs from the original promises.

The language of Pax Silica performs this articulating function. ‘Positive-sum’ avoids the appearance of a competition in which one side gains what another loses; ‘trusted partners’ converts a geopolitical boundary into a moral quality; and ‘mutual security’ gathers within a single expression the interests of a technological power, a mining country, a manufacturing economy, a sovereign wealth fund and a state seeking to attract data centres.

Participants do not need to understand Pax Silica in the same way. For the United States, it may be an instrument of technological competition with China; for the Netherlands, a means of protecting the strategic role of its semiconductor industry; for the Philippines, an opportunity to attract advanced manufacturing, infrastructure and employment; for Argentina and Chile, a route to investment, foreign currency, market access and greater bargaining power within the new technological economy; and for Gulf sovereign wealth funds, a secure framework for converting energy and capital into computing capacity.

These motivations are not equivalent, but they can coexist within a sufficiently broad declaration. The initiative brings participants together because it allows them to recognise themselves in a shared story of prosperity, innovation and security.

Wanting to join is also a form of agency

Critical analysis can easily represent countries in the Global South as passive, deceived or simply subordinated. Such a reading would be inadequate. An ethnography of the state requires us to see that Argentina, Chile and the Philippines do not act as monolithic blocs. They act through fragmented bureaucracies, public companies, subnational governments, technocratic elites, export sectors and territorial movements competing to define the national interest.

Akhil Gupta showed that ‘the state’ is experienced through offices, files, mediation and concrete bureaucratic practices, while James C. Scott explained how administrations make complex territories legible through simplifying categories. From this perspective, accession to Pax Silica does not express a single national will. It is the provisional outcome of internal coalitions capable of translating the alliance’s priorities into the language of investment, employment, productivity and economic security (Gupta, 2012; Scott, 1998).

People holding technocratic positions or directing export sectors may operate as intermediaries between global capital and territory. For them, joining an unequal architecture may be a rational calculation. Negotiating from within promises foreign currency, infrastructure and market access in a world where technological standards will be set with or without their presence. Immediate benefit does not remove the possibility of subordination, but it helps explain why the initiative can generate genuine adherence rather than mere obedience.

Nor is bargaining capacity distributed evenly. Countries offering a large market, finance, technology or a bottleneck that is difficult to replace can demand more favourable terms than those competing against one another to attract extractive investment with fewer resources of power.

In Argentina, the national government and producing provinces promote lithium projects through concessions and incentives such as the Large Investment Incentive Regime, or RIGI, while Indigenous and environmental organisations challenge water use, cumulative impacts and the authorisation of extraction in salt-flat ecosystems. In Chile, the state strategy led by Codelco and its partnership with SQM has advanced alongside parliamentary resistance, corporate litigation, regulatory assessments and consultations with Indigenous communities in the Salar de Atacama. Accession to an international architecture may be driven by administrative elites, state-owned companies and export sectors without thereby representing a homogeneous consensus within the country (Reuters, 10 July 2024; Reuters, 24 April 2025).

The chain attempts to represent the state as a unitary node, but politics reappears through its internal fractures. Economic ministries, provinces, public companies, environmental authorities, communities and territorial movements do not value investment, water, employment or the conservation of salt flats in the same way. Pax Silica’s power also lies in organising a field of possibilities in which integration offers concrete benefits to some sectors while many of its costs are displaced towards those who rarely take part in Washington summits.

Contemporary domination does not always force a choice. It often organises the available options.

From the philosophical system to the economic system

Jacob Helberg’s career introduces another dimension. Before becoming Under Secretary of State for Economic Affairs and the principal public architect of Pax Silica, Helberg served as a senior adviser to Alex Karp, Palantir’s chief executive. He also co-founded the Hill and Valley Forum, a space connecting Washington policymakers, venture capital, technology companies and actors linked to defence (Barron’s, 2025; Axios, 2026).

Helberg’s political trajectory accompanies that institutional shift. Barron’s traces his movement from Democratic donations to financial support for Republican organisations, Trump’s campaign and, ultimately, his appointment. The article also confirms that in 2021 Helberg explicitly rejected Trump’s false claim that he had won the 2020 election. That earlier disagreement helps measure the scale of his political turn, but it should not be confused with the cause he later gave for his realignment: opposition to particular Democratic policies and what he described as an anti-Israel turn and a rise in antisemitism among sections of the left after 7 October 2023 (Le Monde, 2026). His path is not identical to Karp’s, but both form part of a broader realignment of sections of Silicon Valley towards technological nationalism and power politics.

This continuity does not prove that Palantir designed Pax Silica behind closed doors, or that Karp directs US foreign policy. The more significant relationship is the circulation of people, concepts and understandings of power between Silicon Valley, the security apparatus and industrial policy.

Karp’s case shows how an ethical narrative can be constructed around a controversial infrastructure.

During Palantir’s early years, his background in social theory served as an important instrument of legitimation. He presented data-integration software as a ‘philosophical system’ and described privacy controls built into the product as a Hegelian remedy capable of reconciling public security with civil liberties.

The contradiction did not disappear. It was transferred into the design.

A technology built to cross-reference vast quantities of information could be presented not as a threat to privacy but as the instrument best able to protect it. Safeguards written into code became a moral argument for expanding the system.

Over time, Karp’s public discourse moved towards explicit technological nationalism. He criticised companies that refused to work with the Pentagon, defended Western superiority in terms of the capacity to organise violence and argued that hard power in the twenty-first century would be built through software (elDiario.es, 2026).

There is a formal resemblance between Karp’s Hegelian remedy and Pax Silica’s positive-sum promise. Palantir claimed that it could expand surveillance and protect freedom at the same time, while Pax Silica maintains that it can organise asymmetric dependencies and turn them into shared sovereignty. In both cases, a political contradiction is presented as a superior technical solution.

The ethical varnish does not disappear altogether, but the armour becomes visible. The Hegelian remedy legitimised surveillance by embedding ethics in software. Technological nationalism now legitimises hegemony by embedding geopolitics in infrastructure.

That shift prepares the move from the internal ethics of a product to the moral classification of an ecosystem. ‘Trusted’ no longer refers only to the security or privacy of a tool. It begins to determine which companies, suppliers, networks and states may remain within the Western technological perimeter.

Trusted partners and authoritarian technologies

The word ‘trusted’ occupies a central place in Pax Silica, but its meaning is not self-evident.

It does not appear to designate only liberal democracies with transparent institutions. The coalition includes different political systems. Trustworthiness refers above all to strategic alignment, the protection of sensitive technologies, investment security and a willingness to integrate into an ecosystem defined in opposition to geopolitical rivals.

Empirically, this trust is translated into investment-security practices, co-operation in applying national controls, protection of technologies and infrastructures against access or influence deemed improper, commitments against overcapacity and dumping, supplier selection, and deployment of networks, cables and data centres classified as trusted. It also requires legal predictability and interoperability between national policies. The founding declaration does not, however, publish an objective scoring rubric for partners, an independent audit, an explicit suspension procedure or a requirement to adopt specified ISO, IEC or NIST standards. For now, ‘trust’ functions more as a political-administrative category negotiated through national controls, contracts and project agreements than as a universally verifiable technical certification (Pax Silica Declaration, 2025).

This absence does not mean that technical standards are unimportant. It means there is not yet a single public passport for membership. Interoperability, equipment provenance, compliance with export controls and the capacity to verify a supply chain materialise trust case by case, while the final decision remains dependent on authorities and political negotiations.

That definition produces an important tension. The political and corporate field from which Helberg emerged is connected to technologies deployed in migration surveillance, military intelligence, policing analysis and target selection. Palantir has built much of its legitimacy by arguing that such tools protect Western democracies, even as they expand state powers that are difficult to scrutinise publicly (elDiario.es, 2026).

Pax Silica extends this tension to the international scale. Infrastructure is presented as trusted because it falls under the influence of Western allies and companies. Yet trust does not in itself guarantee transparency, labour rights, environmental protection or democratic accountability.

The category only becomes meaningful once we specify for whom an infrastructure is trusted, against which threat, under what procedures and with what opportunity to challenge decisions.

A chain may be trusted by a corporation because it guarantees supply and legal security. It may be less trusted by a community whose access to water depends on decisions made elsewhere. It may be trusted by a defence agency and opaque to people unable to know how their data are being used.

Trust is a political relationship. Pax Silica tends to present it as a technical property.

Europe inside and outside its own sovereignty

The European Union’s accession shows that the distinction between centre and periphery cannot be drawn simply.

Europe possesses indispensable capacities. Through ASML, the Netherlands holds a critical position in the production of lithography equipment. Germany retains a powerful industrial base. The EU has regulatory capacity, a large market and projects of its own concerning technological sovereignty. The 2023 European Chips Act was designed precisely to strengthen the resilience and capacities of Europe’s semiconductor ecosystem (European Union, 2023).

At the same time, Europe depends heavily on US platforms, models and providers. Membership of Pax Silica may strengthen supply-chain resilience and give the EU a role in defining the new order. It may also increase its subordination to Washington’s security and export priorities. The European Commission formally joined in June 2026, while the Netherlands entered amid disagreements with the United States over restrictions on exports of chipmaking equipment to China (Reuters, 25 June 2026).

The Dutch trade minister expressed this ambivalence clearly. The government shared the objective of preventing sensitive technologies from reaching destinations regarded as dangerous, but rejected the prospect of co-operation becoming a US imposition on national-security decisions or on the activities of Dutch companies (Reuters, 23 June 2026).

The Dutch position demonstrates that a territory may control an indispensable bottleneck and still face external pressure over how it is used.

It is therefore too simple to say that Europe has definitively capitulated. What emerges is a relationship of cross-dependencies governed unequally.

Pax Silica offers the EU security and co-ordination. The EU contributes markets, technology and international legitimacy. The dispute concerns who will retain the final word when US security, European autonomy and commercial interests no longer coincide.

Complementarity works while priorities remain aligned. Once they diverge, the hierarchy concealed by the term reappears.

A third path of incomplete sovereignty

The polarisation between Pax Silica and the Digital Silk Road does not exhaust every possibility. A local-first infrastructure can combine public computing centres, open models, local hosting, interoperable standards, data portability and adaptation to national languages or needs. It does not amount to an autarky capable of reproducing the full technology stack independently. Rather, it means retaining operational control over the layers where an institution can decide where a model runs, who accesses the data, how the system is audited and whether changing provider is possible.

Latin America already has a relevant experiment. In February 2026, Chile presented Latam-GPT, an open model co-ordinated by the National Centre for Artificial Intelligence, CENIA. CENIA’s launch announcement attributes its construction to more than one hundred professionals and over sixty institutions in fifteen countries. The corpus contained more than 300 billion tokens, equivalent according to the institution to roughly 230 billion words; Associated Press described the collection as more than eight terabytes of data. The figures use different units but converge in showing a regional operation of unusual scale (CENIA, 10 February 2026; Associated Press, 10 February 2026).

The project was conceived as foundational infrastructure for educational, health, administrative and productive applications, rather than as a direct commercial competitor to ChatGPT or Gemini. It also represents an attempt to build technical capacity and incorporate linguistic varieties, memories and regional contexts at the training stage instead of treating them as later corrections. That ambition makes it a concrete alternative to the role of simple technological consumer.

The case also reveals the limits of that sovereignty. Version 1.0 used a 70-billion-parameter Llama 3.1 architecture and Amazon Web Services for training. CENIA reported that AWS reduced the process from twenty-five days to nine. Associated Press stated that later versions would be trained at the University of Tarapacá’s supercomputing centre, a Chilean infrastructure conceived precisely to reduce dependence on private clouds. Regional autonomy is therefore being built through intertwined domestic and external layers: regional data, curation and governance, but chips and software frameworks drawn from global supply chains (CENIA, 2026; Associated Press, 2026).

Argentina forms part of the Latin American corpus — Wired attributed around 210,000 documents to the country — but the public sources reviewed do not yet reveal a national public-computing programme or open model of comparable scale. Chile, by contrast, occupies a position of regional co-ordination and is developing local supercomputing capacity. The asymmetry within the alternative itself is significant: a country may contribute documents, cultural knowledge and researchers without exercising equal control over the infrastructure that turns those materials into a model (Wired, 1 September 2025).

Recent initiatives in Ukraine and Portugal provide external comparisons concerning local control of models for public services and national languages, but they do not replace the Latin American problem (Reuters, 7 July 2026; Reuters, 1 July 2026). Open weights do not guarantee sovereignty when computing, energy, deployment tools or maintenance depend on external providers. The most realistic alternative is to build autonomy layer by layer through public infrastructure, auditable code, portable formats, local expertise and an effective capacity to leave a supplier. Pax Silica narrows that path when it presents specialisation within the bloc as the only rational form of autonomy.

The knowledge that declares others ignorant

Mark Hobart called ‘the growth of ignorance’ the process by which the expansion of expert knowledge also increases what institutions define as unknown, backward or incapable.

In classical development projects, local knowledge was not simply incomplete information that could be added to a plan. It was often subordinated in advance because the problem and its solution had already been defined through the categories of those intervening. Economics, technology and management offered an apparently neutral language that concealed the relationship between knowledge and power. To develop a population, it first had to be constituted as underdeveloped and ignorant (Hobart, 1993).

In the AI economy, deficiency takes on new categories: a lack of talent, institutional weakness, absence of traceability, insufficient regulatory maturity, infrastructure deficits or low innovative capacity. Those limitations may be real, but the methods used to measure them and the institutions authorised to correct them are never neutral.

A territory may possess complex knowledge of its water cycles, soils, biodiversity and forms of collective ownership, yet appear on the map of a technology chain chiefly as a mineral reserve, an energy site or land available for an industrial zone. A community may understand the ecological limits of a watershed with extraordinary precision and still be represented as an obstacle to an investment deemed strategic.

Hobart helps us understand that such disqualification does not merely ignore knowledge that was already present. Expert intervention actively produces the boundary between knowledge and ignorance: it turns some observations into technical evidence and others into belief, resistance or lack of maturity. Declaring insufficiency shapes which forms of knowledge enter the file, which authorities are allowed to speak and which realities remain outside the map.

Marisol de la Cadena provides a more precise Andean vocabulary for understanding what is lost in that operation. In Earth Beings, based on a decade of conversations with Mariano and Nazario Turpo, two runakuna from Cuzco, she shows that relations with Ausangate exceed the modern separation between nature and culture. When those relations enter administrative files, courts or state negotiations, they must be translated into recognisable categories — land, environment, heritage or property. Translation makes political intervention possible, but it also leaves out what those categories cannot contain: earth-beings, reciprocal obligations and the relations that make a territory something other than a deposit of resources. This is not merely a disagreement over the knowledge available, but over which entities may appear as part of political reality (De la Cadena, 2015).

Local knowledge need not be literally destroyed. It is enough to render it politically inaudible and admit as relevant only those forms of knowledge that can be translated into availability, productivity, legal certainty, traceability or return. Pax Silica does not declare populations primitive. It declares them insufficiently prepared for an economy defined by others.

Silicon has a territory

The same language of synthesis that Karp applied to software in order to reconcile surveillance and freedom must now be subjected to a material test. Pax Silica’s ‘shared sovereignty’ and ‘positive-sum’ promise acquire a different meaning once we calculate what each tonne is worth, how much remains in the territory and who bears the costs of extracting it.

De la Cadena’s ethnography cannot simply be transferred from Peru’s Ausangate to a Kolla community in Argentina. It does, however, offer a methodological warning: before turning a salt flat into a reserve, it is worth observing which relations have been left untranslated. That warning takes on human scale in Tusaquillas, in northern Argentina, where an Associated Press account depicts Irene Leonor Flores de Callata, a 68-year-old Kolla woman, guiding llamas and sheep along the dry bed of a river. In one of the driest places on the planet, fresh water sustains grazing, food and the community’s continuity, while beneath the nearby salt flats lies the lithium required for the energy transition. The scene does not illustrate a theory from outside. It shows the exact point at which the economic category of ‘reserve’ begins to leave vital relationships outside the map (Associated Press, 13 March 2024).

The contrast does not justify attributing a single cosmology to every Kolla community, but it does show that the dispute goes beyond calculating how many cubic metres of water an operation may consume. For the logistical architecture, the salt flat is a reserve of critical mineral to be mobilised into the value chain. For the ways of life described in the report, water, animals, salt flats and community continuity are not separable items in an inventory. The conflict concerns what the territory is and the relations that make it inhabitable, not only the distribution of future rent.

Corporate images of artificial intelligence tend to show clouds, networks and luminous flows, yet AI does not exist without an enormous mobilisation of matter. It requires minerals, water, energy, land, cables, ports, factories, cooling systems and data centres. It also produces waste, heat, territorial change and conflict over resources.

The International Energy Agency summarises this materiality directly. There is no artificial intelligence without electricity for data centres, and the expansion of computing raises questions of security, affordability, emissions and energy infrastructure (IEA, 2025).

Gian Carlo Delgado Ramos proposes examining these processes through the concept of socio-economic metabolism. Every economy appropriates energy and materials, transforms, distributes and consumes them, and generates waste. Technological sophistication does not abolish this metabolism. It can intensify it and displace its effects towards new extractive frontiers (Delgado Ramos, 2013).

Political ecology requires attention to who controls the appropriation of resources, what territorial transformations it produces and how benefits and costs are distributed. Delgado Ramos analyses land appropriation as the acquisition of packages of natural assets. Territory is reduced to economic value while consultation processes are weakened and other relationships with the land are excluded.

Argentina and Chile do not join Pax Silica merely as names in a diplomatic declaration. Together with Bolivia, they form the three points of the so-called lithium triangle. In 2025, Argentina was the world’s fourth-largest supplier and approved a $2.5 billion Rio Tinto project under its RIGI incentive regime; Chile retained the largest known exploitable reserves and was the second-largest producer. Chile is also the world’s leading copper producer, supplying a metal essential to electricity networks, data centres, vehicles and military equipment. Both countries also possess exceptional potential for solar, wind and renewable-hydrogen energy, particularly in the Atacama and Patagonia (Reuters, 20 May 2025; Reuters, 7 April 2025; Reuters, 8 July 2025; Pfennig et al., 2022).

Both countries can use this position to attract capital, expand refining, manufacture active materials, produce batteries or develop technology of their own. They may also become exposed to competition between territories seeking to export resources with little local processing. Participation does not predetermine their reduction to raw-material suppliers, but neither does it guarantee movement into the highest-value layers.

Price comparisons help reveal where value accumulates. There is no single ‘lithium price’. It varies according to whether the product is ore, concentrate, carbonate or hydroxide, its degree of purity and the market involved. As a recent public reference, the most-traded lithium-carbonate contract on the Guangzhou Futures Exchange closed on 3 March 2026 at 150,860 yuan per tonne — roughly $21,000 at an indicative exchange rate — after falling almost 13 per cent in one session. On 1 July, Reuters reported that carbonate prices were still close to three times their mid-2025 level, illustrating a volatility borne directly by extractive territories (Reuters, 3 March 2026; Reuters, 1 July 2026).

That exchange price does not necessarily equal the revenue actually received by producers. Much lithium is sold under long-term contracts, index-linked formulas, quality discounts, logistics costs and delivery conditions that can separate realised prices from the benchmark quotation. The gap between the two is particularly important when royalties or public rents are calculated.

Three regimes for capturing lithium rent

Nor is there a single, directly comparable fiscal amount per tonne. Argentina relies principally on provincial royalties levied on mine-mouth value. Chile combines price-linked lease payments, taxation and state corporate participation. Bolivia operates through a state-owned company that receives sales revenue while also bearing production costs. Despite those differences, public data make it possible to establish orders of magnitude.

Argentina. Law 24.196 caps ordinary provincial royalties at 3 per cent of mine-mouth value. For projects that had not begun construction before the 2024 reform, provinces may raise the maximum to 5 per cent. If these percentages were applied as a purely arithmetic exercise to the benchmark price of $21,000 per tonne of lithium carbonate, the ceiling would be about $630 per tonne at a 3 per cent royalty and $1,050 at 5 per cent. Actual receipts are usually lower because the base is not the gross international price but mine-mouth value after deductions for transport, treatment, marketing, administration, smelting and refining. These are direct payments to the provinces. The national state also collects taxes, but the burden depends on profits, deductions and the regime applying to each project and cannot be expressed as a fixed amount per tonne (Argentina, Law 24.196, Articles 22 and 22 bis).

Chile. The state does not apply a uniform lithium royalty per tonne. CORFO leases rights in the Salar de Atacama to SQM and Albemarle through contracts whose rents vary with the price, alongside taxes, territorial contributions and, following the partnership with Codelco, state participation in profits. As a historical indication, CORFO received 5.4 trillion Chilean pesos in additional income from those contracts during 2022 and 2023. The USGS estimated Chilean production in those years at 38,000 and 44,000 tonnes of contained lithium, together equivalent to about 436,000 tonnes of lithium carbonate equivalent. Dividing one by the other gives a rough average of 12.4 million Chilean pesos, around $14,400, per equivalent tonne. This is neither a current rate nor an exact measure of total revenue. It is a retrospective estimate corresponding to the extraordinary price peak of 2022 and 2023, combines two accounting years and captures contractual rents rather than a fixed royalty. Revenue per tonne falls when prices decline. New contracts continue to link payments to lithium prices and add Codelco participation, meaning that no exact present-day figure can be calculated without the realised price and accounts of each operation (El País, 2025; USGS, 2024; Reuters, 22 September 2025).

Bolivia. The contrast is institutional. Yacimientos de Litio Bolivianos is state-owned, so rather than charging a private company a royalty equivalent to Argentina’s, it invoices production and bears its costs. Reuters reported that in 2024 YLB earned $15.6 million from production of roughly 2,000 tonnes, equivalent to about $7,800 in gross revenue per reported tonne. This is neither net profit nor a transfer to the Treasury. Operating costs, investment, depreciation, financing and possible losses must be deducted from it. State ownership of gross revenue does not mean that the state captures the full rent. In addition, the Bolivian government announced a new lithium-specific law in January 2026, indicating that the fiscal and contractual regime remained open to revision (Reuters, 21 October 2025; Reuters, 19 January 2026).

At the time this essay was completed, Bolivia was not among Pax Silica’s 24 participants. Its model offers a different route, based on state ownership, low production volumes and an ongoing review of the conditions for attracting capital and technology. Bolivia’s absence shows that membership of the coalition is neither automatic nor universal. It also exposes the limits of a strategy that, despite formally retaining the resource within the state, has not yet transformed its enormous resources into a large-scale industry (Financial Times, 2026).

The comparison reveals three distinct relationships between territory and rent. Argentina offers a visible but limited royalty calculated on a reduced base. Chile captured a much larger share during periods of high prices through progressive contracts and state participation, albeit with great volatility and less transparency per tonne. Bolivia formally retains production within the state, but its low scale and costs prevent gross revenue from being treated as publicly available wealth. The legal percentage of ownership matters, but so do productivity, technological capacity, sales contracts and the ability to transform the resource before export.

At the manufacturing end, BloombergNEF’s survey placed the 2025 global average price of completed lithium-ion battery packs at $108 per kWh. A 60 kWh pack therefore represented about $6,480 before integration into a vehicle, while stationary-storage packs averaged $70 per kWh and electric-vehicle packs $99. This is not a direct physical equivalence: one tonne of carbonate feeds multiple batteries, and each pack incorporates other minerals, electronic components, engineering and labour. That is precisely why the comparison is revealing. Value does not concentrate at the moment the mineral is extracted. It accumulates through chemical conversion, cathode materials, cells, packs, management systems, industrial integration, software, finance and branding (Maisch, 2025).

The different fiscal regimes — Argentine royalties, Chilean rents linked to price and Bolivian state ownership — reveal not only different ways of capturing rent. They also express different expectations about the kind of integration being pursued, from an economy content to extract to one attempting to process, manufacture and retain technological capacity within the territory.

That choice determines whether participation is limited to exporting resources or also includes refining, manufacturing, scientific knowledge, a share in intellectual property and the capacity to govern systems. It also determines who bears water, energy and environmental costs. A battery, a chip and an AI model belong to the same material economy, but their benefits and harms are not distributed across the same territory.

Silicon has a body. It also has a geography.

New Clark City and sovereignty negotiated by contract

New Clark City, in the Philippine province of Tarlac, shows what happens when the supply chain acquires a concrete territorial form.

In April 2026, the United States and the Philippines announced an industrial hub of roughly 4,000 acres, or about 1,620 hectares, within the Luzon Economic Corridor. The project was intended to serve as a platform for allied manufacturing and to strengthen supply chains for critical minerals, semiconductors, electronics and technological infrastructure. At that stage, Philippine authorities still had to assess land availability and negotiate the project’s details (Reuters, 17 April 2026).

The official formulation appeared to confirm the logic of Pax Silica. A commercial installation located in allied territory would offer companies speed, scale, legal certainty and access to a protected industrial network.

In May, however, a controversy revealed the politics hidden inside the logistics.

The Philippine Bases Conversion and Development Authority, BCDA, said that the United States had requested immunity arrangements for its personnel. The Philippines rejected any special regime and maintained that the project would remain subject to Philippine law and be treated as an ordinary commercial development (Philstar, 2026).

Helberg later denied that the aim had been to obtain diplomatic immunity and explained that Washington sought certainty and predictability for investors. In any event, the final framework remained open and the parties had a period in which to define the protections and activities of the zone (Reuters, 21 May 2026).

The dispute does not prove that the Philippines will lose sovereignty or that the zone will operate under US law. It shows that sovereignty is not preserved or surrendered in a single gesture, but negotiated through plots of land, clauses and exemptions. The decisive issues appear inside the contract: who will exercise jurisdiction, what protections companies will receive, which activities will be permitted, who will approve plans, what environmental and labour obligations will apply, and what will happen if the host country’s priorities diverge from those of the alliance.

Hannah Appel’s contribution takes on a concrete empirical role here. Her ethnography of extraction shows that contracts, technical standards, enclaves and corporate expertise do not merely regulate an economic activity that already exists. They manufacture the legal world in which that activity can appear normal, licit and efficient. In New Clark City, infrastructure does not begin when a machine arrives. It begins when a clause distributes jurisdiction, risk, immunity and responsibility (Appel, 2019).

Opacity is not a secondary communication problem here. It is part of the object under study. The more the zone is presented as a technical solution, the more necessary it becomes to identify which political decisions have been transferred into agreements that are still being negotiated.

New Clark City also demonstrates Philippine agency. The government wants investment and industrialisation, but formally rejects extraterritoriality. It neither accepts nor opposes Pax Silica as a single package. It negotiates the conditions of participation.

The tension is not between absolute sovereignty and total submission. It lies in every condition determining who will govern the infrastructure.

Securing supply while society is eroded

Pax Silica seeks to secure the material conditions for the expansion of artificial intelligence. It protects minerals, energy, manufacturing, trade routes, chips, data centres and models.

Yet that same expansion may destabilise the societies expected to sustain it.

Carson Block has advanced an extreme hypothesis. Automation could displace a significant share of skilled work, reduce contributions to pension funds and trigger capital withdrawals that would hit especially hard the large technology companies whose valuations depend on passive investment flows.

Block estimates that AI could replace around 15 per cent of US knowledge-economy employment within three or four years. Falling contributions and rising withdrawals would force index funds to sell assets, disproportionately affecting Nvidia, Microsoft, Amazon and other companies at the centre of the AI economy (Block, 2026).

This is not a consensus forecast. It is the thesis of an investor specialising in short positions and should be presented as such.

Its relevance to this essay does not depend on accepting the figures. It lies in the contradiction Block identifies: Pax Silica is organised to secure the supply of artificial intelligence, while artificial intelligence itself may erode the social demand sustaining that supply. An economy may possess chips, electricity and data centres while simultaneously reducing the incomes of people who purchase goods, contribute to pension systems and provide political legitimacy for growth.

Block argues that stabilising markets would be easier than reorganising a society in which growth had become detached from employment. The problem would not be merely financial. It would be a rupture in the social contract.

The contradiction can be stated directly. What kind of peace does a technological chain guarantee if the society financing it loses the income needed to participate in the economy that chain produces?

Pax Silica secures materials and platforms but does not yet offer an equivalent answer for labour, redistribution or social protection. Risk once again replaces wealth and is distributed unequally: companies secure supplies, states secure alliances, extractive territories assume environmental damage, and workers face the possibility of displacement.

Supply-chain security can coexist with the insecurity of those living around it.

A peace of assigned positions

In 1949, Truman argued that producing more would lead to peace and prosperity. His speech turned science, capital and technology into instruments of a global mission. It also transformed millions of people into inhabitants of underdeveloped areas, positioned behind a future that others already represented.

Pax Silica does not reproduce that project exactly. It does not promise that every country will travel the same path towards becoming a similar industrial society. It offers a place within a secure architecture. That change appears more modest and realistic because it recognises that no economy can independently control all the minerals, processes, machines and forms of knowledge required for artificial intelligence. Yet abandoning self-sufficiency does not remove the issue of equality; it can make it more urgent.

Interdependence can be mutually sovereign only if every party has some capacity to negotiate its functions, reject conditions, change the rules and leave the architecture without suffering disproportionate harm.

Pax Silica promises access, trust and complementarity, but it has not yet demonstrated an equal distribution of replaceability, value, knowledge and decision-making power. A territory providing a mineral may be replaced by another, while the actor controlling the standard that organises the chain becomes more indispensable with every new accession.

This is the difference between participating in an order and being able to govern it. In this new order, power does not reside only in possession of territory or conventional military capacity, but in the ability to design, co-ordinate and control the technological flows that make contemporary economic and political life possible. That power nevertheless becomes fragile when the infrastructure being secured erodes the employment, demand and social contract on which its legitimacy depends.

The anthropology of development does not require us to deny the usefulness of co-operation, investment or infrastructure. Nor does it demand that every international initiative be interpreted as a coherent conspiracy, or every accession by the Global South as an act of submission. Its warning is more precise: when a policy presents a particular organisation of the world as a technical necessity, we should examine which power relations disappear from the diagram; when it promises security, which risks it displaces and towards whom; when it speaks of complementarity, whether the assigned functions can change; when it offers trust, who can demand accountability; and when it distributes places along a chain, who retains the power to redesign it.

Pax Silica does not promise that every country will arrive at the same future. It offers each a position within an architecture that already imagines what that future will be and what every territory must contribute to sustain it.

The decisive issue is not only who may enter.

It is who will be able to change position once inside.

And who will have the authority to build another architecture when the peace of silicon no longer appears to be a shared promise.

References