Artificial intelligence is becoming one of the defining technologies of the 2020s. AI assistants can write, create images, analyze documents, generate software and increasingly perform tasks on behalf of their users. But behind every AI interaction lies something less visible: enormous computing infrastructure.

As the global AI race accelerates, technology companies are building data centers at unprecedented speed. These facilities require huge amounts of electricity, water and advanced computer chips. The result is a new question that is becoming increasingly important around the world:

Can the planet generate enough energy to power the AI revolution?

The answer could shape not only the future of technology, but also energy markets, climate policy, infrastructure and geopolitics.

The Hidden Infrastructure Behind AI

When people use an AI chatbot, the experience appears almost effortless. A question is typed into a phone or computer and an answer appears seconds later.

Behind that simple interaction, however, are thousands of servers working inside massive data centers.

Training advanced AI models requires enormous computational resources. Running those models for millions of users also requires continuous electricity. As AI becomes integrated into search engines, smartphones, office software, autonomous systems, financial services and industrial operations, demand for computing power is growing rapidly.

Gartner estimates that worldwide spending on AI-optimized infrastructure as a service could reach approximately $42 billion in 2026, representing a 96% increase over the previous year.

That growth is transforming AI from a purely technological story into an energy story.

Data Centers Are Becoming Strategic Infrastructure

For decades, countries primarily competed over oil, gas, electricity grids and industrial capacity.

Now another resource is becoming strategically important: computing power.

The largest technology companies are investing billions of dollars in new data centers. These facilities are increasingly being built close to reliable electricity sources and major network infrastructure.

The trend is changing local economies as well.

A new data center can bring investment, construction jobs and tax revenue. But it can also create pressure on electricity grids and water supplies.

This is why communities in several parts of the world are beginning to ask whether AI infrastructure should have unrestricted access to local resources.

The debate is no longer simply about whether AI is useful.

It is becoming a debate about who pays for the infrastructure that makes AI possible.

AI's Growing Electricity Appetite

The International Energy Agency has highlighted the rapid growth of electricity consumption by data centers. Global electricity demand from data centers increased by around 17% in 2025.

The number becomes more significant when considering what may happen next.

AI models are becoming more powerful. At the same time, more people are using them and companies are integrating AI into everyday software.

Instead of asking a chatbot one question, an AI agent may eventually perform dozens of actions automatically: searching the internet, analyzing files, booking appointments, writing emails, monitoring markets or managing business processes.

That means the future of AI may involve not just more users, but more continuous computation.

The shift from "asking AI" to "AI doing things" is already becoming visible. OpenAI recently described this transition as AI moving from something people ask questions to something they actively put to work.

More useful AI could therefore also mean more demand for electricity.

The Climate Paradox

This creates one of the most fascinating contradictions surrounding artificial intelligence.

AI could help reduce emissions.

It can optimize electricity grids, improve weather forecasting, make industrial processes more efficient, reduce unnecessary transportation and help researchers develop new technologies.

AI could also accelerate scientific research into batteries, renewable energy and climate adaptation.

The World Economic Forum has argued that AI can play an important role in accelerating the energy transition and improving climate adaptation.

But AI itself requires enormous infrastructure.

And that infrastructure consumes electricity.

If the additional electricity comes from low-carbon sources, the environmental impact can be significantly reduced.

If it comes from coal, oil or natural gas, the situation becomes much more complicated.

This creates what could become the AI climate paradox:

The technology that could help humanity fight climate change could also increase energy demand enough to make the problem harder.

The Return of Fossil Fuels?

The AI boom is already influencing discussions about the future of natural gas, nuclear power and renewable energy.

Technology companies want reliable electricity available 24 hours a day. Solar and wind power are expanding rapidly, but their output depends on weather conditions.

That makes energy storage, nuclear power, hydroelectricity and natural gas increasingly important parts of the conversation.

Some recent analyses have raised concerns that AI could increase demand for fossil-fuel energy rather than immediately replacing it with renewable alternatives.

This does not mean AI will inevitably increase global emissions.

The outcome will depend heavily on how quickly electricity generation becomes cleaner and how efficiently AI systems operate.

In other words, the environmental impact of AI may ultimately depend as much on the energy system surrounding the technology as on the technology itself.

Water: The Other Resource Nobody Sees

Electricity is not the only resource required by modern data centers.

Cooling powerful computers can require significant amounts of water.

As AI infrastructure expands, water consumption is becoming another part of the debate.

This matters particularly in regions already suffering from drought or water shortages.

A data center may appear to be a highly digital facility, but its physical infrastructure can have a very real impact on the surrounding environment.

The United Nations University has highlighted the growing environmental footprint of AI infrastructure, including carbon emissions, water consumption and land use.

This means that the AI revolution is not happening entirely in the cloud.

It is happening in buildings, power stations, transmission networks, semiconductor factories and cooling systems around the world.

The New Geopolitics of Computing

Energy is only one side of the AI competition.

Countries are also competing for advanced semiconductors, data centers, research talent and access to computing infrastructure.

The United States and China remain central players in this competition, while Europe, Japan, South Korea, Taiwan and other economies are trying to strengthen their own technological capabilities.

Computing power is increasingly becoming a strategic asset.

The geopolitical consequences could be enormous.

A country with abundant electricity, advanced chips and powerful AI infrastructure may gain an advantage in manufacturing, defense, science and financial services.

This means that the global AI race could eventually resemble previous races over oil, telecommunications or space technology.

The difference is that the new strategic resource is not something extracted from the ground.

It is computation.

AI Could Change the Electricity Grid

There is another possibility that is often overlooked.

AI may not simply consume electricity.

It could also help transform the electricity system itself.

AI can analyze enormous quantities of data from power grids and predict demand more accurately. It can help operators identify failures, optimize energy distribution and coordinate renewable generation with consumption.

In the future, AI-powered systems could potentially help households and businesses automatically decide when to charge batteries, use appliances or consume electricity based on real-time prices and availability.

That could make the grid more flexible.

In this scenario, AI becomes both a major consumer of electricity and a tool for managing electricity more intelligently.

The Race for More Efficient AI

The solution may not simply be to build more power plants.

Another approach is to make AI itself more efficient.

Researchers and technology companies are working on smaller models, specialized chips and more efficient methods of running AI systems.

If the amount of computing required to perform a task falls dramatically, the energy cost of that task could also decline.

This could become one of the most important technological competitions of the next decade.

The winning AI company may not necessarily be the one with the largest model.

It could be the company that produces the most capable AI with the least computing power.

A New Kind of Digital Infrastructure

The first internet revolution required cables, routers, mobile networks and data centers.

The AI revolution is building another layer on top of that infrastructure.

It requires specialized chips, enormous data centers, high-capacity electricity connections and increasingly sophisticated cooling systems.

That infrastructure is becoming so important that it may influence where factories are built, where energy infrastructure is developed and which regions become technology hubs.

The geography of AI could therefore become just as important as the software itself.

What Happens Next?

The next few years could determine whether the AI boom becomes compatible with a cleaner energy future.

Three things will matter enormously.

First, how quickly renewable and low-carbon electricity expands.

Second, how efficiently AI models and data centers use energy and water.

Third, whether governments and technology companies can coordinate infrastructure development before electricity and water shortages become serious constraints.

The stakes are high.

AI could help humanity solve some of its biggest problems, from energy optimization to scientific discovery.

But technological progress has never been free.

Every digital service ultimately depends on physical resources.

The smartphone in your hand may feel weightless.

The cloud may sound invisible.

Artificial intelligence may appear almost magical.

But somewhere in the world, machines are consuming electricity to make it work.

And as AI becomes more powerful, that invisible energy bill could become one of the biggest stories of the decade.

The Real AI Question

For years, the central question surrounding artificial intelligence was:

How intelligent can machines become?

That question is still important.

But a new question is emerging:

How much energy will the world need to support them?

The answer could determine the speed of the AI revolution, the future of electricity markets and even the environmental consequences of one of the most transformative technologies in human history.

AI may ultimately become greener, smarter and more efficient.

But for now, the world's most powerful technology is discovering a very physical limitation:

intelligence needs energy.

#AI #artificialintelligence #technology #energy #climatechange #future

Hollywood writers went on strike over pay this week, but the decision by companies like Netflix and Disney to use artificial intelligence to replace human writers in the future only fuels anger and fear among the protesters.

AI programs such as ChatGPT, which have an increasing ability to amazingly mimic human conversation, are raising concerns across many industries. As such, the White House called a meeting with Big Tech this week to discuss potential threats. As part of weekly talks with studios and streaming platforms that ended Monday, the Writers Guild of America has asked for binding contracts to govern the use of artificial intelligence. According to the proposal, nothing written by AI can be considered "literary" or "source" material - industry terms that govern who receives royalties - and scripts written by WGA members cannot be "used to train AI." However, according to the WGA, the studios "rejected our proposal" and responded with a proposal to meet once a year to "discuss advances in technology".

The writers' strike is gaining momentum after the statements of streaming platforms that writers in the future may be replaced by artificial intelligence. Netflix and Disney have confirmed that they allow the use of algorithms.

"It's nice that they're offering a meeting about how they're using this against us!" joked WGA negotiating committee member Eric Heisserer, who wrote the script for Netflix's Bird Box. “Art cannot be made by a machine. You are losing the heart and soul of history... I mean the first word is 'artificial', he told AFP during a picket outside the streaming giant's headquarters. While writers already know this, the danger is that "we have to watch tech companies destroy business trying to find out," he said.

While few of the television and film writers who spoke to AFP believe their work could be replaced by computers, for many it was an added blow that studios and streaming platforms believe in the possibility of such a replacement. Screenwriters fear that Hollywood executives, where Silicon Valley companies have overturned many traditional practices such as long-term contracts for screenwriters, may try to cut costs even further by forcing computers to write more blockbusters. The statements of Hollywood's top executives at the Milken Institute Global Conference in Beverly Hills did not dispel the writers' fears in any way. "In the next three years, you're going to see a movie that's been written by AI and shot…good," said film producer Todd Lieberman. “Not just scripts. Editing, everything… scriptwriting, whatever,” added Fox Entertainment CEO Rob Wade. "AI in the future, maybe not next year or next year, but if we're talking about 10 years? Artificial intelligence will be able to do absolutely all of these things.”

Even for sci-fi writer Ben Ripley, who believes AI plays no role in writing, the introduction of legislation "to put up barriers" is "very necessary". Writers "have to be original," he said. "Artificial intelligence is the antithesis of originality."

#netflix #disney #ai #artificialintelligence #movies #film #technology

The system developed at the University of Texas in Austin (USA) - an artificial intelligence connected to an MRI scanner read the thoughts of volunteers. For now, it still makes mistakes, but in the future, similar devices could help people who are unable to communicate normally.

During the experiment, volunteers listened to a story or imagined telling it. Their brains were observed with the help of magnetic resonance imaging (MRI), and the artificial intelligence connected to it turned people's thoughts into text consistent with them. Importantly, no electrodes were needed to be implanted in the brain, and the person using the invention was not limited to a predetermined list of words, the researchers emphasize. However, the system requires special training - each person using it listens to podcasts for several hours, during which the computer watches his brain. "For non-invasive methods, this is a real leap forward compared to what was previously achieved, i.e. usually reading single words or short sentences" - says Prof. Alex Huth, author of the paper, which appeared in the journal "Nature Neuroscience". "Our model decodes long-term, continuous speech on complex topics," he emphasizes.

For now, the computer is not very accurate - it reads thoughts quite precisely in about 50 percent. However, he can often convey the meaning of the utterance. For example, he translated the thought, "I don't have a driving license yet," into "she hasn't started learning to drive yet." Listening to the thought, "I didn't know whether to scream, cry, or run away," he read as "She started screaming and crying and then she said 'I told you to leave me." The system was also good at reading the minds of the volunteers when they watched the videos. The creators of the program also addressed the topic of its possible abuses. They assure that at least currently it is impossible to read someone's mind if the person does not want to. It is also impossible to do it with someone with whom the system has not undergone a long training. “We take concerns about abuse very seriously and are working to prevent them. We want to make sure that people use these technologies when they want to, and that it helps them," said Jerry Tang, who led the research. Scientists hope that their idea will allow for the construction of devices that will allow people currently unable to communicate with the world, e.g. after severe strokes. The current version of the system only allows for its use in the laboratory, but according to the researchers, this can be changed - for example, instead of a large MRI scanner, it is likely that a much smaller device designed for non-invasive functional near-infrared spectroscopy can also be used.

#austin #usa #ai #artificialintelligence #technology #tech #future #science