Billions will use AI without knowing they're doing so
Durante décadas, Google Search ha sido el catálogo universal de la humanidad: una puerta hacia la información. Ahora, la compañía anuncia que ese catálogo se convertirá en un agente capaz de actuar en nombre de quien pregunta, integrando su modelo de inteligencia artificial Gemini en Search y en más de trece productos con miles de millones de usuarios. La búsqueda deja de ser un espejo que refleja el mundo y se convierte en una mano que lo transforma. Es un momento que redefine no solo cómo encontramos información, sino cómo tomamos decisiones y nos relacionamos con el entorno digital.
- Google ha anunciado que su modelo Gemini operará como una capa agéntica dentro de Search, capaz de reservar viajes, ejecutar operaciones bursátiles y completar tareas complejas mediante conversación, sin que el usuario abandone la interfaz.
- La integración se extiende a Gmail, Maps, Drive y YouTube —trece productos con más de mil millones de usuarios cada uno—, convirtiendo la IA agéntica en infraestructura invisible para buena parte de la humanidad conectada.
- El modelo publicitario que ha financiado a Google durante décadas enfrenta una tensión existencial: si el usuario ya no navega entre resultados sino que delega en un agente, el espacio tradicional del anuncio desaparece o debe reinventarse.
- Para sostener esta transformación, Google multiplicará por seis su inversión en centros de datos y chips TPU personalizados, una apuesta de capital sin precedentes en la historia de la compañía.
- El verdadero punto de llegada aún es incierto: miles de millones de personas comenzarán a usar inteligencia artificial agéntica sin ser plenamente conscientes de ello, lo que plantea preguntas abiertas sobre autonomía, confianza y poder corporativo.
Google Search está a punto de dejar de ser lo que cuatro mil millones de personas han conocido. Durante décadas, buscar significaba recibir enlaces. Pronto, significará recibir acciones. La compañía ha anunciado la integración de Gemini, su modelo de lenguaje, como una capa agéntica dentro de Search: un sistema que no solo responde preguntas, sino que ejecuta tareas en nombre del usuario. La diferencia es la que existe entre un catálogo de biblioteca y un asistente personal.
La escala del cambio resulta difícil de dimensionar. Google Search procesa cinco billones de consultas al año y alcanza a más personas que la suma de las poblaciones de China, India, Estados Unidos y la Unión Europea. Ahora ese motor reescribe su propia naturaleza: en lugar de comparar precios de cruceros, el usuario podrá reservar el crucero directamente, conversando con la IA. En lugar de consultar cotizaciones bursátiles, podrá fijar parámetros de compra y venta que el sistema ejecutará de forma automática.
Pero la transformación no se detiene en Search. Gemini se desplegará por Gmail, Maps, Drive y YouTube —este último acaba de lanzar Ask YouTube, que devuelve respuestas en formato de vídeo—. En todo ese ecosistema, la misma capa agéntica operará de manera invisible, tomando decisiones y completando acciones en nombre de los usuarios.
El riesgo para Google es real. Search es el producto que genera los ingresos publicitarios que financian todo lo demás. La transición hacia la IA agéntica exige multiplicar por seis el gasto de capital en centros de datos y en los chips TPU que la compañía fabrica para ser más eficientes energéticamente que los de Nvidia. Y queda abierta una pregunta fundamental: si el usuario ya no navega entre resultados sino que delega en un agente, ¿dónde se coloca el anuncio? ¿Cómo se monetiza un mundo en el que la interfaz de búsqueda se ha vuelto infraestructura invisible?
Google apuesta a que el beneficio justifica el riesgo. Gemini dejará de ser una aplicación que se abre conscientemente —como Claude o ChatGPT— para convertirse en parte del sustrato digital que miles de millones de personas usan sin pensar en ello. El cambio más profundo no será tecnológico, sino humano: afectará a cómo las personas adquieren conocimiento y toman decisiones. Esa transformación apenas comienza.
Google Search is about to become something different than what four billion people have known it to be. Right now, when you search, you get links. Soon, you'll get action. The company announced this week that it's embedding Gemini, its own large language model, directly into Search as what engineers call an agentic layer—a system that doesn't just answer questions but executes tasks based on what you ask it to do.
The scale of this shift is hard to overstate. Google Search processes five trillion queries annually. It reaches more people than the combined populations of China, India, the United States, and the European Union. For decades, it has been the engine that reshaped advertising, commerce, and how humans and organizations find information. Now Google is fundamentally rewriting what search means.
The difference between old search and what's coming is the difference between a library catalog and a personal assistant. Before, you could compare cruise prices. Now you'll book the cruise directly through the search interface, specifying your preferences in conversation and letting the AI handle the transaction. You could research stock prices. Now you'll set parameters—buy if the price drops below a certain level, sell if it rises above another—and the system executes the trade automatically. Search becomes not a tool for finding information but a tool for getting things done.
But Google isn't stopping at Search. The company maintains thirteen products with over a billion users each, and five products with more than three billion. Gemini will spread across all of them. It will live in Gmail, reshaping how you manage messages. It will inhabit Google Maps, changing how you navigate. It will integrate into Drive, transforming how you organize files. YouTube just launched a feature called Ask YouTube that returns answers in video form. Across this entire ecosystem, the same agentic layer will operate, invisible to most users, making decisions and taking actions on their behalf.
This transformation carries genuine risk for Google itself. Search is the company's golden goose—the product that generates the advertising revenue that funds everything else, that paid for acquisitions of competitors, that bankrolled experimental moonshots. The shift to agentic AI demands something the company has never done before: multiplying its capital expenditure by six. Data centers will need to expand dramatically. The company will need to manufacture far more of its custom TPU chips, the processors it designed to be more energy-efficient than Nvidia's GPUs. The infrastructure costs are staggering.
There's also a question of whether the old advertising model survives intact. If users are no longer browsing lists of results but asking an AI agent to complete a task, where does the ad go? How does Google monetize a world where the search interface has become invisible infrastructure?
Yet the company is betting that the upside justifies the risk. Gemini will stop being merely an application—a thing you open, like Claude or ChatGPT—and become part of the digital foundation that billions of people use without thinking about it. Billions of people will soon be using agentic artificial intelligence without knowing they're doing so. The relationship between customers and companies will shift. But the deeper change will be in how people themselves acquire knowledge and make decisions. That transformation is only beginning.
Notable Quotes
The transformation will reshape how customers relate to companies and how people acquire knowledge itself— Analysis from El Mundo reporting on Google's strategic shift
The Hearth Conversation Another angle on the story
When you say the AI becomes invisible infrastructure, what does that actually mean for someone using Google Maps?
It means the system starts anticipating what you need before you fully ask for it. You're navigating to a restaurant, and Gemini notices you're running late, so it automatically texts the restaurant to push your reservation. You didn't ask it to do that—you just described the problem in conversation.
So the user loses some control?
Not loses—trades it. You're trading the friction of managing details for the convenience of the system handling them. But yes, you're trusting the AI with more decisions.
And Google's advertising business—how does that survive if search results disappear?
That's the real gamble. The company is betting that agentic AI creates new kinds of commerce that are more valuable than ad clicks. If the AI books your vacation or executes your stock trade, Google takes a cut of that transaction. It's a shift from attention-based revenue to transaction-based revenue.
But what if users don't want that? What if they want to see options?
Then Google has a problem. The entire bet assumes people prefer convenience over choice. History suggests they often do, but not always. And regulators are watching closely.
What about the energy cost? Six times the capital expenditure?
That's the infrastructure bet. Google believes the market will bear those costs because the value created is so much larger. But it also means the company is betting heavily on its own chip design working at scale. If it doesn't, the whole equation breaks.