In a bustling lab in Santiago, Chile, dedicated researchers are racing against time, eager to harness the power of artificial intelligence before the global tech wave leaves them behind.
Meanwhile, in the southern neighborhood of Cerrillos, local activists are fiercely protesting against the construction of new data centers – the very infrastructure that fuels AI.
Inside the presidential palace, government officials are strategizing. Their goal: expand Chile’s technological footprint with AI, but do so responsibly, managing precious resources and avoiding public discontent.
This complex interplay of ambition and apprehension has ignited a nationwide debate about artificial intelligence. Chile, a South American nation of 20 million people, now finds itself an unexpected focal point, grappling with the profound trade-offs inherent in the global AI revolution.
Chile has actively sought foreign investment and cultivated local talent to build its AI capabilities. While these efforts promise significant economic growth, they also raise concerns about environmental impact and increased reliance on powerful U.S. technology corporations. A government proposal to regulate new data centers has already triggered widespread protests and intense parliamentary debates.
Many ordinary Chileans, who often view artificial intelligence with skepticism or indifference, question whether these potential benefits outweigh the costs.
Rodrigo Cavieres, a prominent voice from the Socio-Environmental Community Movement for Water and Land (MOSACAT), criticized the trend, stating that AI is being treated “like a new form of fetishism.” He emphasized that “data centers are being given priority over the population’s needs.”
In Santiago, passionate protesters successfully demonstrated against the proliferation of data centers, compelling Google to abandon its plans for a second facility in Chile.
These mounting tensions in Chile mirror similar struggles unfolding worldwide. Nations, from the resource-rich United Arab Emirates to the environmentally conscious Netherlands, are confronting a stark choice: aggressively invest in AI, risking overextension, ecological damage, and public discontent, or face the prospect of being technologically sidelined.
These national deliberations often trace back to pivotal moments, such as the one experienced by Álvaro Soto, director of the Chilean National Center for Artificial Intelligence, in 2023. His concerns ignited when an early version of the ChatGPT chatbot, questioned about Chilean literature, predominantly cited only Pablo Neruda. This incident highlighted a critical flaw: existing AI models were not being developed to genuinely represent the unique culture and linguistic nuances of regions like Chile.
In response, Soto and his dedicated research team are now meticulously training their own AI model using previously overlooked data from across Latin America. Recognizing the urgency, Chilean President Gabriel Boric, in his June state of the union address, underscored the nation’s imperative to embrace AI. His administration is actively working to simplify the process for foreign companies establishing data centers and to embed AI tools into everyday governance.
However, despite this strong political drive, communities directly impacted by AI data centers remain profoundly discontent. In northern Santiago, a local group is vocally opposing an Amazon site, labeling it environmentally harmful gentrification. Elsewhere, another group is protesting a Google data center, fearing its potential adverse effects on a nearby wetland. A third, highly organized group in Santiago’s southern suburbs successfully pressured Google to retract its plans for a second data center in the country.
In light of these challenges, President Boric’s government aims to redirect future data center development away from the densely populated Santiago region towards the more sparsely inhabited north. This shift, however, raises new alarms among environmentalists who fear the potential ecological damage to the fragile Atacama Desert, an area already under stress from extensive mining operations.
“Chile often offers a glimpse into the future,” commented Marina Otero, an architect and Harvard lecturer specializing in data centers. “The ongoing debate surrounding AI will only intensify. It’s a clear preview of what’s to come globally.”
The Ecstasy and Agony
Matías Muchnick, the visionary CEO and founder of NotCo, at his Santiago laboratory. His company leverages artificial intelligence, fed by precise food preparation data, to power its innovative recipe engine.
One morning in a futuristic kitchen-lab in Santiago, the team at food-tech startup NotCo was hard at work. Their mission, powered by artificial intelligence, was to crack a complex challenge for major food and snack brands: finding natural replacements for food dyes recently prohibited by U.S. Health Secretary Robert F. Kennedy Jr.
NotCo’s innovative approach utilizes an AI engine, meticulously trained on molecular data, to re-engineer ingredients for some of the world’s leading food companies. To construct this sophisticated system, the $1.5 billion-valued company required substantial computing power unavailable domestically. In 2018, CEO and founder Matías Muchnick reached out to Google, securing access to specialized AI-designed microchips.
“Our ambition was to create a formidable research and development hub, and that collaboration vastly simplified our journey,” Mr. Muchnick explained.
NotCo stands as a shining example of Chilean ingenuity in leveraging AI. However, the very effort to build the foundational infrastructure necessary to support such companies has ignited a significant public outcry.
NotCo’s modern offices in Santiago.
The company has emerged as a beacon of Chilean technological expertise.
In 2015, Google established its inaugural Latin American data center in Quilicura, a Santiago suburb nestled alongside a vital wetland. According to environmental impact assessments filed during the project’s planning, this facility consumes 50 liters of water per second for cooling—a volume comparable to the daily usage of approximately 8,000 Chilean households. A Google spokesperson later stated that last year’s water consumption was significantly lower, equating to that of a single golf course.
Historically, data centers often rely on evaporation to cool their heat-generating computers. While newer technologies aim to conserve and recycle water, environmental advocates point out that many facilities continue to demand substantial water resources.
Local activist Rodrigo Vallejos presented a compelling video showcasing the area’s past, when it boasted vibrant marshlands and serene lagoons. Today, a stark contrast reveals much of the landscape parched, even during the typically rainy season.
Before and after footage vividly illustrates the environmental changes to the wetlands in Quilicura, Chile. (Videos by Rodrigo Vallejos and Marcos Zegers)
For Mr. Vallejos, the balance sheet of progress is heavily skewed. He highlights that the data center provides minimal employment, and the designated community “offset”—a park adjacent to a cemetery—sees very little use.
“Ultimately, we risk transforming into little more than a global warehouse for artificial intelligence,” he asserted.
Rodrigo Vallejos, a dedicated activist and law student, has emerged as a key figure in the campaign to protect the Quilicura wetland.
In many respects, Chile is already on this path. The nation serves as a regional AI hub, currently hosting 33 data centers, a figure projected to double by 2030, according to Chile Data Centers, a leading industry organization.
However, many Chileans are only just beginning to grasp the true function and wide-ranging impact of these expansive facilities. When Google unveiled plans in 2019 for an additional data center in nearby Cerrillos, residents initially anticipated new job prospects. Yet, environmental documents soon revealed a facility requiring minimal staff, designed to consume an astonishing 228 liters of water per second—equivalent to the consumption of approximately 40,000 households.
Local resident Tania Rodríguez took the initiative, going door-to-door to inform her community about the project’s potential issues and rallying her neighbors to oppose it. Their collective efforts paid off: last year, Google publicly announced the withdrawal of its plans.
Yet, Rodríguez’s relief is short-lived. Recently, representatives from a Google subsidiary quietly resumed community engagement for a different data center project.
“We are not against artificial intelligence—it is a technology we absolutely must develop,” Ms. Rodríguez affirmed. Nevertheless, she remains resolute in her intention to continue protesting against Google, asserting, “We must be profoundly efficient in our use of natural resources; it is paramount for human survival.”
The Astronomy Model
The Vera Rubin Observatory, perched in Cerro Pachón, Chile. During the 1990s, international astronomers eagerly competed to construct telescopes in the nation’s pristine desert landscapes.
In 2024, Sebastián Howard, an official within Chile’s Ministry of Science, Technology, Knowledge and Innovation, made a startling discovery: tech companies were planning to build 30 new data centers nationwide over the next four years. Almost all of these proposed facilities were slated for the Santiago region, an area already severely impacted by drought.
“We simply lacked the energy capacity for such an expansion,” Mr. Howard stated regarding the capital. “More critically, we lacked the water resources.”
Howard spearheaded a government initiative to re-route data center development. He and his team engineered a sophisticated mapping tool to identify locations where these facilities would incur minimal environmental and social harm. Their findings pointed to Antofagasta, a northern desert city boasting abundant solar energy potential.
This innovative plan draws inspiration from Chile’s highly successful astronomy model. Back in the 1990s, when international astronomers eagerly competed to erect telescopes in the country’s famously clear desert skies, the Chilean government implemented a groundbreaking regulation: 10 percent of all telescope observation time had to be allocated to local researchers. This policy propelled Chile into a global leadership position in astronomy.
Mr. Howard and his team envision a similar future for AI. “If these companies are committed to investing here,” he explained, “we must ensure their infrastructure also serves our domestic universities and businesses.”
While many nations globally are constructing their own data centers to secure AI access, Chile’s strategy differs. Under their proposed model, local enterprises and academic institutions would gain access to the advanced computing power developed by foreign tech companies, rather than building their own.
However, promoting this vision has proven challenging. Many critics view the government’s initiatives as attempts to favor large corporations, potentially at the expense of public welfare. Activists have particularly highlighted a proposed environmental regulation amendment that could significantly diminish transparency regarding water and electricity consumption by data centers.
The intensity of public sentiment was evident at a workshop last year, where Mr. Howard attempted to explain AI development to community leaders. He was met with shouts and outrage after remarking, “It’s a privatized country. These companies can do whatever they want.”
Aisén Etcheverry, Chile’s former Minister of Science, Technology, Knowledge and Innovation, warned, “The moment you lose the capability to understand how your machine is working, or the ability to even build your own machine, that’s the moment you lose.”
The willingness of tech giants to adopt this northward shift for data centers remains uncertain. Felipe Ramírez, head of Amazon Web Services in Chile, raised concerns that situating data centers nearly 680 miles from Santiago could introduce significant internet latency. This delay would be particularly detrimental for voice-based AI applications, where even a millisecond’s pause can frustrate users.
“While the concept is sound on paper,” he admitted, “I’m unsure if we will ultimately see global AI models being trained in Antofagasta.”
Despite these hesitations, the government is forging ahead. Aisén Etcheverry, a presidential adviser and former Minister of Science, Technology, Knowledge and Innovation, confirmed that discussions with leading tech firms are ongoing. She expressed optimism that a comprehensive plan would be unveiled before the national elections next month.
Etcheverry underscored the profound stakes: if Chile fails to actively influence how AI systems interpret its language, historical context, and societal structures, the nation risks being excluded from shaping its own future.
“The moment you lose the capability to understand how your machine is working or the ability to even build your own machine, that’s the moment you lose,” she reiterated. “And that is a future we absolutely want to avoid.”
Pascale Bonnefoy contributed to this report.