Jensen Huang's AI Billions Bypass Latin America — And the Region Is Running Out of Time

When Nvidia CEO Jensen Huang addressed a Paris audience on October 13, 2026 — with French President Emmanuel Macron in the room — he delivered a message that was equal parts pitch and warning: Europe needs to get in the race, and Nvidia intends to fund its entry. Within hours, his company announced plans to deploy $2 billion in new European AI investments within the following week, layering on top of the $8 billion Nvidia had already committed across the continent over the prior twelve months. The day after, in the United Kingdom, Huang committed $250 million in compute access for AI safety research — distributed at $50 million per year over five years to institutions including the University of Oxford, University College London, and Imperial College London. No government in Latin America received a comparable offer.

That asymmetry is not incidental. It reflects a structural divergence in how governments around the world are responding to what Jensen Huang has described as the most transformative technological moment since the invention of electricity. Latin America, home to 650 million people and more than a century of technological dependency, is watching the race from the periphery.

What Huang Is Actually Pitching

The sovereign AI concept Nvidia has championed since early 2024 is built on a foundational reframe: artificial intelligence should be treated the way nations treat roads, power grids, and water systems — as critical infrastructure that a country owns and controls, rather than borrows from a foreign provider. Huang's sharpest argument centers on data: any government that ships its sensitive national information abroad in exchange for AI-generated insights is surrendering something it may never fully recover.

The pitch comes with commercial logic embedded. Nvidia's sovereign AI revenue exceeded $30 billion in fiscal year 2026 — more than 13% of its total annual revenue — with the United Kingdom, France, the Netherlands, Canada, and Singapore named as primary drivers. Bank of America projects the global sovereign AI infrastructure market could reach $50 billion per year over the long term. Against that backdrop, Europe's tech leaders are not just strategic partners; they are a major revenue opportunity. Latin America, by contrast, currently accounts for less than 5% of Nvidia's global revenue.

South Korea illustrates the scale of what competition looks like. In October 2025, at the APEC Summit, Nvidia announced it would supply more than 260,000 advanced GPUs to South Korean government agencies and companies including Samsung, SK Group, Hyundai, and NAVER Cloud — one of the largest single national AI infrastructure commitments in history. Active sovereign AI programs have now launched across more than three dozen countries. Latin America's presence in that cohort remains thin.

The Numbers Expose a Structural Problem

The most precise measurement of the gap comes from the ECLAC-CENIA Latin American Artificial Intelligence Index (ILIA 2025), a joint assessment of 19 regional economies published in October 2025. Its central finding: the region produces 6.6% of global economic output but captures barely 1.1% of worldwide AI investment. No country in the region reaches the global average for AI investment per GDP per capita; the regional benchmark sits roughly six times below that standard.

The infrastructure picture is equally stark. A 2025 analysis by the United Nations Development Programme, drawing on global data center mapping, found that Latin America and the Caribbean host just 4.8% of the world's data center infrastructure — compared to 38.5% for the United States. Nearly all of the compute capacity expected to power the region's projected $30.2 billion AI market by 2033 is owned and operated by a handful of American and Chinese technology firms.

The market opportunity is real: research firm IMARC Group values the regional AI market at approximately $4.7 billion in 2024 and projects a near-sevenfold expansion to $30.2 billion by 2033, driven by digital transformation and AI adoption across key industries. The financing, governance, and hardware foundations needed to capture that growth domestically, however, remain largely absent.

Brazil: The Test Case and the Warning

Brazil is where sovereign AI in Latin America will either gain its foothold or confirm its structural weakness. Márcio Aguiar, Director of Nvidia's Enterprise Division for Latin America, was direct in an interview with Bloomberg Línea about what delay means in practice, warning that Brazil is "missing the opportunity to board this passing train" — and that each month of inaction pushes the country into a new technological era it is not equipped to compete in.

Foreign capital has recognized Brazil's potential. Microsoft committed $2.7 billion to cloud and AI infrastructure across several data center campuses in São Paulo state. Amazon Web Services announced a $5 billion data center buildout in Querétaro, Mexico, demonstrating that hyperscaler appetite for Latin American AI investment exists — even if it flows more readily to Mexico than to Brazil.

A July 2026 Getulio Vargas Foundation study, commissioned by Scala Data Centers and Norgás, quantified the opportunity: adding 12.7 gigawatts of new data center capacity by 2035 would require between $431.8 billion and $698.5 billion in total investment, with each 100-megawatt facility generating an estimated 12,560 direct and indirect jobs. The potential is enormous. The regulatory scaffolding is not in place. Brazil's Redata tax incentive bill — which would create a special tax regime for data center services — remains in Congress, and import costs on GPU hardware continue to constrain domestic deployment.

One landmark transaction does stand out. On June 29, 2026, MeetKai Brasil won SERPRO's sovereign AI tender — the "IA Soberana" selection process finalized by Brazil's federal technology agency on June 26 — becoming the private-sector partner tasked with constructing the country's national sovereign AI platform on Nvidia accelerated computing. The award stands as one of the most significant sovereign AI contract wins in the region's history. But as Japan, Malaysia, and the UAE have demonstrated through two years of committed policy execution, a single government contract cannot substitute for the legislative and regulatory architecture that makes systematic buildout possible.

Where the Region Is Actually Building

Outside Brazil, a handful of countries are showing what execution looks like when regulatory and institutional conditions align.

In El Salvador, the National AI Agency (ANIA), in collaboration with Nvidia and WideLabs, launched an open sovereign AI dataset in early June 2026. Named Nemotron-Personas-El Salvador, the dataset contains approximately one million synthetic citizen profiles generated from official national statistics and built on Nvidia B300 Blackwell Ultra processors — the first initiative of its kind in Central America. By joining Nvidia's global Nemotron-Personas program, El Salvador now builds AI from a data foundation rooted in its own population, rather than relying on models trained overwhelmingly on data from wealthier nations.

In Mexico, AI Green Data Centers is developing an Nvidia-aligned AI Factory in Monterrey, with construction scheduled to begin in 2026 and a target of 100 megawatts of capacity by 2030. The project, undertaken with CIPRE Holding and certified through Nvidia's partner program, is explicitly designed to process regional data within regional borders.

Ecuador's Telconet, a longstanding Nvidia partner, has expanded its data center footprint with a new facility in Guayaquil opened in September 2026 and is advancing a next-generation AI campus in Quito — building toward 100 megawatts of capacity and bringing high-density GPU infrastructure to a country that rarely features in the region's AI investment conversation.

Talent and Training: A Foundation Without the Building

On human capital, Nvidia and SoftServe ran AI Week LATAM 2026 across Colombia, Mexico, and Chile from late September through October 3, offering free technical workshops and hands-on sessions supported by four regional universities, with 400 Nvidia Deep Learning Institute certification slots available. The initiative reflects real commitment to the region's talent base.

But workforce development without corresponding infrastructure creates a pipeline with nowhere to go. Aguiar himself warned that without domestic AI data centers, the region's most skilled engineers, machine learning researchers, and data scientists migrate abroad to follow the opportunities. Training talent for export is not a sovereign AI strategy.

The Clock Is Ticking

There is a paradox at the center of this moment that Nvidia does not obscure. According to the CNAS Sovereign AI Index, Nvidia supplies GPUs in 45% of all tracked sovereign AI infrastructure projects globally. Even the most ambitious national AI programs — including those explicitly designed to reduce foreign technology dependence — are overwhelmingly powered by hardware from one American company. In practice, sovereignty still runs on chips from Santa Clara.

For Latin American governments, this paradox does not diminish the urgency — it sharpens it. The path to controlling national AI capability may run through Nvidia hardware, but the regulatory frameworks, data governance laws, and compute capacity built today will determine whether the intelligence generated from a country's own data belongs to that country in the decade ahead.

The window remains open. But the nations that act within the next 12 to 18 months will control the digital infrastructure of the 2030s. Those that delay will spend it purchasing intelligence built from their own exported data.

© {{Year}} Latin Times. All rights reserved. Do not reproduce without permission.