There are names that concentrate an era. In artificial intelligence, that name is Nvidia: $46.7 billion in quarterly revenue, 56 percent year-over-year growth, a market capitalization exceeding $4.4 trillion. When Nvidia moves, the technological architecture of the next decade moves with it.
That concentration generates both confidence and dependence. The United States recognizes Nvidia’s strategic character and has restricted the sale of its most advanced products to China. Beijing responds with massive subsidies to Huawei and SMIC for alternatives. The European Union advances its Chips Act to reduce vulnerabilities. Each decision reflects what is actually at stake: not a market, but which countries will have the capacity to innovate autonomously and which will depend on external suppliers for the technological foundation.
An inevitable question emerged in 2025: is this concentration sustainable? Some feared speculative excess. Many argued it was a supercycle driven by genuine global demand for AI adoption across industries. Unlike the .com era, where value often attached to unproven ideas, Nvidia’s growth is directly linked to the massive deployment of AI. Nvidia’s earnings reports are not merely accounting statements. They are global trend indicators.
While a handful of nations lead this race, Latin America occupies another position: observer and recipient of consequences. Without capacity for advanced chip design or fabrication, the region experiences technological dependency as higher hardware costs and restricted access. But this does not mean passivity. Startups and researchers leverage available tools to build applications and business models that insert them into the global AI ecosystem, though always from the periphery.
Yet something else has moved. Governments are no longer discussing strategic technology. They are acquiring it.
In August 2025, the Trump administration took a 9.9 percent stake in Intel for $8.9 billion. This was not a subsidy or a loan. It was the government entering as a shareholder, with direct ownership of one of the critical assets of the digital economy. The action was swift and decisive. The government has no board representation or formal governance rights, but it holds equity in a company that sits at the foundation of semiconductor manufacturing—the backbone of artificial intelligence.
This is disruptive but not unprecedented. Latin American history shows that governments intervene when a resource is perceived as vital. Mexico and Peru nationalized oil industries. Chile consolidated control over strategic lithium reserves. More recently, China offers a template: through massive financing and strategic control, the state shaped companies like Huawei and SMIC, acting as venture investor and industrial strategist.
Intel’s case fits that logic, positioned in the semiconductor value chain. This chain defines which countries can innovate with autonomy and which depend on external suppliers. What is at stake is not merely one company’s competitiveness. It is the technological architecture that will shape the decades ahead. The United States has decided it cannot afford to lose Intel. It will own a piece of it instead.
The question that follows is no longer hypothetical. It is already being answered by the actions of major powers. The global semiconductor dynamics and state intervention in technology giants expose a structural vulnerability that runs through Latin America: dependence on external actors for the technological foundation that will shape the coming decades. The region cannot simply observe this game. It must learn from its own history and chart a different course—one that builds genuine technological capacity while avoiding the mistakes of politicization and inefficiency.
For Peru, this moment offers clarity. The private sector is already leading innovation, investing in talent and partnering with universities to build homegrown technological capacity. The state must create frameworks that attract investment, strengthen human capital, and facilitate integration into global value chains. Neither state ownership nor passive dependence. But intentional positioning: clarity about what role the region will play in an architecture it did not design and cannot control, but in which it must still choose to move.
The architecture of power in AI is not being set. It is being built. Some countries are building it. Others are being built around it. A few might negotiate their way to something between. The difference lies not in possession, but in the intentionality with which a nation thinks about its position within the system and acts on that thinking.