Gemini and the Energy Cost Behind Smart Replies

Google has just surprised the technology world and users when it officially integrated the AI ​​assistant Gemini into the Google Maps application. This is a bold move, not only changing the way we use maps, navigate or search for information, but also opening the door to a new era, where artificial intelligence is present in every aspect of digital life. However, along with the promise of unlimited convenience and creativity, Gemini also raises many big questions related to energy consumption, climate change and the responsibility of technology giants in the context of global efforts for a sustainable future.

1. Gemini enters Google Maps

For many years, Google Maps has been a familiar “pocket map” for hundreds of millions of users. People use it to find directions, look up locations, find restaurants, hotels or plan trips. However, with the appearance of Gemini, Google Maps is no longer just a navigation tool but becomes a smart assistant that accompanies users throughout their journey.

The biggest difference is the ability to understand and respond to open-ended questions. Users do not need to type in a specific location, but can ask: “If I only have 5 hours in Saigon, where should I go to enjoy the cuisine and take beautiful souvenir photos?” or “I want to explore a corner of Paris with few tourists, what do you suggest?”. Gemini will analyze interests, context and Google’s huge data to make reasonable suggestions, even arrange a time-appropriate route. This is a step forward from a “lookup” application to a real “travel advisor”.

Gemini also adds visuals by creating illustrations right in Maps. For example, if you want to know what it feels like to stand on the Golden Bridge in Da Nang at dawn, Gemini can create a simulation image for you to visualize. When choosing a hotel, you can request an image of the view from a room facing the sea for easy comparison. These images do not replace the real experience, but they help reduce mistakes and inspire your trip.

Gemini also changes the concept of navigation. It not only guides you from A to B as quickly as possible, but also takes into account the experiences along the way. A foodie might be suggested a route that includes popular eateries, while a photography enthusiast might be guided through scenic spots. This makes Google Maps a highly personalized tool, where each trip is unique to each person.

2. Gemini's expansion ambitions from Google Maps to Google Home

If Google Maps is the starting point, Google Home is the important destination in the ambition to make Gemini a part of modern life. Google Home is used by many families to control devices by voice but is still mainly a “command execution” device. When Gemini appears, it can turn the home into a smart living space, where technology thinks, plans and accompanies the owner.

One morning, you wake up and Gemini has your day’s summary ready: meetings, weather, traffic, and a reminder to leave 10 minutes early to avoid traffic. If you’re on a healthy diet, Gemini suggests a suitable breakfast and starts the coffee maker before you even get to the kitchen. When you leave the house, Gemini automatically syncs with Google Maps to guide you to a route that’s not only fast but also relaxing, like passing through the park so you can take a few minutes to relax.

In the evening, when you return home, Gemini can automatically switch to “relaxation mode.” Soft lights, soft background music, comfortable air conditioning, and a gentle reminder: “You have a call scheduled for 9 o’clock. Do you want me to remind you?” This is how Gemini gradually becomes a “digital butler” in the family.

Not stopping there, Gemini also has the potential to support many other aspects: monitoring the sleep and health of members, reminding the elderly to take medication on time, supporting children's learning by answering difficult questions, or even suggesting recreational activities for the whole family.

Google’s ultimate ambition is to turn Gemini into a comprehensive assistant, present from the moment you leave home until you return. If successful, Gemini will no longer be a mindless piece of software but a “digital family member” that is part companion, part advisor, part helper.

3. Energy and climate change concerns

Along with the excitement about Gemini’s new capabilities, the tech community and society have begun to worry about the environmental impact of new-generation AI models. This is not without reason. AI, especially large language models (LLMs) like Gemini, ChatGPT, or Claude, require a huge amount of computation to operate. Every time a user sends a question and receives an answer, thousands of calculations are performed in energy-consuming data centers.

Google, in a report released in August 2025, provided a clearer picture of Gemini’s energy consumption. It found that each text prompt consumed only 0.24 watt-hours of electricity, equivalent to watching TV for less than 9 seconds. Each such interaction emitted 0.03 grams of carbon dioxide and used about 0.26 milliliters of water, most of which went to cooling the server system.

At first glance, these numbers may seem small and insignificant. But when multiplied on a global scale, the problem becomes quite different. Gemini is expected to have around 47 million active users by 2025. If each person sends just 10 requests per day, the energy consumption and emissions will become a huge number.

Google also claims that it has made great strides in reducing its environmental impact. In one year, the Gemini app’s carbon footprint has been reduced by 44 times thanks to algorithmic optimization and improvements to data center infrastructure. Google is also exploring ways to reduce the amount of water needed for cooling, such as recycling water or switching to air-cooling methods.

But that’s just the tip of the iceberg. Because in addition to the energy consumed each time a user sends a request, there’s a much bigger factor: the AI ​​model training process. To train a system like Gemini, Google has to use millions of GPUs for weeks, even months. This process consumes a huge amount of electricity and creates a significant amount of carbon emissions. This is a point that is rarely mentioned in reports, but it is a real challenge in the problem of balancing AI development and environmental protection.

4. Warnings from MIT and the truth about data centers

According to research from the Massachusetts Institute of Technology (MIT), the energy demand of data centers in North America will increase from 2,688 megawatts in 2022 to 5,341 megawatts in 2023, nearly doubling in just one year. AI is one of the main reasons for this. Large language models, recommendation systems, image and video generation all require a significant amount of computing power compared to traditional applications.

In particular, MIT warns that the vast majority of electricity used to power new data centers today still comes from fossil fuel power plants. This poses a paradox: while tech companies are constantly promoting their commitment to sustainability, renewable energy, and emissions reduction, the reality is that they still rely heavily on climate-damaging energy sources.

Globally, the number of data centers is exploding to meet the demand for AI computing. Countries from the US to Europe to Asia are building new data centers. Without clean energy solutions to go along with it, this boom could become a huge burden on global emissions reduction goals.

5. Does AI really accompany environmental protection efforts?

The big question is: Will the rapid development of AI be accompanied by advances in environmental protection? On the one hand, AI brings many benefits in energy optimization, renewable energy technology development, smart power system management, and improved production efficiency. But on the other hand, the process of operating and developing AI itself is consuming a huge amount of resources.

We can look at solar energy and electric vehicles as examples. These are both technologies that are expected to help reduce dependence on fossil fuels. However, as AI develops at a rapid pace, the benefits of reducing emissions from electric vehicles or solar power may be “eroded” by the huge amount of energy consumed by AI systems.

That puts a huge onus on tech companies. Google, Microsoft, OpenAI, and Meta all need to be more transparent about their true carbon footprints, including the training and operation of their AI. And investing more heavily in renewable energy, like wind, solar, and energy storage, must become a top priority.

AI can only truly contribute to environmental protection when it provides a tool to solve climate problems while minimizing its own impact on the planet. Otherwise, we could fall into a vicious cycle: using AI to optimize the environment but allowing AI to destroy the climate balance.