Elon Musk’s AI Power Bet Puts His Solar-Electric Vision Under Pressure

Elon Musk spent years selling a simple idea: the world could run on solar power, batteries, and electric transport. Now his AI business is leaning hard into natural gas. That contradiction is getting harder to ignore.

Musk built much of his public image around moving the world away from fossil fuels. Tesla’s original 2006 Master Plan described the company’s purpose as helping shift the economy away from a “mine-and-burn hydrocarbon economy” and toward a “solar electric economy.” Musk also pushed the idea that a relatively small area of solar panels in the U.S. could produce enough energy to power the country, with battery storage filling the gaps.

That was the clean thesis.

Build electric cars.
Build solar panels.
Build batteries.
Stop burning fossil fuels.

But in 2026, the story looks different. Musk’s artificial intelligence company, xAI, is using natural gas infrastructure to power its fast-growing data center operations. The move has triggered criticism from clean energy watchers, who say Musk is walking away from the very solar future he spent years promoting.

From Solar Evangelist to Gas-Powered AI

The heart of the criticism is simple: Musk’s AI ambitions need huge amounts of power, and xAI is choosing gas.

According to Electrek, xAI is operating 62 methane gas turbines across two data centers in Memphis, Tennessee, and Southaven, Mississippi. These turbines are used to power xAI’s Colossus supercomputers, which support Grok, Musk’s AI chatbot. Electrek also reported that the combined facilities could emit more than 6 million tons of greenhouse gases per year, based on xAI permit applications, along with more than 1,300 tons of health-harming air pollutants.

The same report says xAI has been buying more gas turbines, including a reported $2.8 billion purchase. That suggests gas is not just a temporary backup plan. It is part of the growth strategy.

That is where the tension starts.

Tesla still sells solar panels, Powerwall batteries, and Megapack energy storage systems. Its public mission remains tied to sustainable energy. But Musk’s AI operation is scaling around fossil fuel power because AI data centers need electricity now, not later.

This is the hard truth of AI infrastructure.

It does not wait politely for the perfect clean energy plan.

Why AI Is Pulling Tech Back Toward Fossil Fuels

Training and running large AI models requires enormous amounts of electricity. These systems need stable, around-the-clock power. Data centers cannot run only when the sun is shining. They cannot afford long interruptions. They need electricity at huge scale, every hour of the day.

Solar can help. Wind can help. Batteries can help. But building a fully renewable system with enough storage for massive AI workloads is expensive, complex, and slower than plugging into gas-fired power.

That is why AI companies are scrambling for power from almost anywhere.

TechBuzz described xAI’s natural gas move as part of a larger pattern across the AI industry. Microsoft and Google also face tension between climate promises and rising data center electricity demand. Google’s emissions have reportedly jumped since 2019, driven heavily by AI infrastructure, while Microsoft has pursued nuclear power deals to support its own data center needs.

So Musk is not alone.

But the criticism is sharper because he is not just another tech executive. He made climate and electrification part of his brand.

That makes the gas pivot more visible.

Tesla’s Solar Products Are Not Powering the AI Buildout

One of the most pointed criticisms is that xAI appears to be using Tesla’s battery products, but not much of Tesla’s solar hardware.

Electrek reported that xAI spent $697 million on Tesla Megapacks for its data centers, but did not buy a materially significant amount of Tesla solar panels, according to the SpaceX S-1 filing cited by the outlet. The same report says SpaceX spent $131 million on 1,279 Cybertrucks, while solar panels for the gas-powered AI data center were apparently not a major priority.

That is the part that looks awkward.

If Tesla has the solar panels and batteries, why is Musk’s AI business not building the clean energy model he has long claimed is possible?

The practical answer is speed and cost.

Gas turbines can be deployed to supply firm power faster than building a fully solar-and-storage-backed data center at the scale xAI wants. But the brand answer is harder. Musk’s companies have spent years arguing that sustainable energy is not just possible, but inevitable.

Now AI is exposing the gap between the pitch and the buildout.

Grok’s Demand Problem Adds Another Layer

Electrek also argues that xAI’s energy use is especially difficult to justify because Grok has not become the dominant AI product Musk hoped for. The report says Grok’s monthly active users fell in April and that downloads dropped sharply, while rivals such as ChatGPT, Claude, Gemini, and DeepSeek remained ahead.

That matters because energy use in AI is already controversial. If a company is burning large amounts of fossil fuel to train and run a widely used, market-leading tool, supporters may argue the trade-off is necessary. But if the product is struggling to gain traction, the power consumption looks more wasteful.

Electrek’s framing is harsh: Musk is burning fossil fuels to run a chatbot that fewer people are using.

That is not neutral language. It is opinionated. But it captures the core criticism.

The energy cost is real. The public benefit is still debated.

Anthropic Deal Raises More Questions

The story also includes a strange business turn.

Electrek reported that xAI leased major compute capacity to Anthropic, a company Musk had publicly criticized months earlier. According to the report, Anthropic was paying xAI $1.25 billion per month for Colossus compute, with the deal potentially worth more than $40 billion through 2029.

That shift matters because it changes the business story.

If Grok could not fully use the data center capacity, selling compute to another AI company helps recover value from expensive infrastructure. But it also makes the energy issue broader. The gas-powered compute is not only supporting Musk’s own chatbot. It is becoming part of the wider AI infrastructure market.

That means xAI’s gas turbines are not just a side project. They are part of a business model.

Space-Based Solar: Innovation or Convenient Rebrand?

The second major thread is SpaceX’s reported interest in space-based solar and orbital data centers.

Electrek reported that SpaceX’s S-1 filing included a new energy thesis around space-based solar power, arguing that solar arrays in orbit can generate far more energy than ground-based panels because they can receive sunlight around the clock. The filing also reportedly discusses an orbital data center constellation that could involve up to one million satellites in low Earth orbit.

The pitch sounds futuristic. Solar panels in space get more consistent sunlight. They avoid night, weather, clouds, and land constraints. For AI data centers, that could sound like the perfect answer.

But critics say the timing is suspicious.

SpaceX is reportedly seeking a huge IPO valuation, and a space-based solar/data center thesis could help position the company as more than a rocket and satellite business. It creates a reason for more Starship launches, more orbital infrastructure, and more long-term revenue.

That is where the criticism lands.

If Musk previously argued that solar on Earth was abundant and practical, why is the new answer to launch solar panels into orbit?

The Earth-Based Solar Argument Has Not Gone Away

Musk’s older argument was not weak. It was strong.

He argued that solar energy on Earth is plentiful, and that storage could make it reliable. In 2017, he told the National Governors Association that a relatively small section of Nevada, Texas, or Utah could power the United States with solar panels, and that battery storage would take far less space than most people imagine.

The basic logic still holds.

Solar panel costs have fallen. Battery technology has improved. Grid-scale storage is growing. Tesla’s energy business has become one of the company’s stronger divisions.

So critics are asking: why not prove the old thesis now?

Build solar-powered AI data centers on Earth. Pair them with batteries. Use Megapacks. Use Tesla solar. Use the same model sold to households and utilities.

The answer may be uncomfortable.

It may not be the fastest or cheapest option for xAI’s immediate race against OpenAI, Google, Anthropic, and others.

That is the trade-off.

The Bigger Problem: AI Is Testing Climate Promises

This story is not only about Musk. It is about the whole tech industry.

AI companies are making huge climate claims while building infrastructure that consumes massive amounts of electricity. Many still talk about carbon neutrality. But when power demand gets urgent, they often choose what is available, not what is cleanest.

That is the new pressure point.

AI wants:

• Constant power
• Cheap electricity
• Fast deployment
• Large-scale data centers
• Low downtime
• Competitive speed

Climate goals want:

• Lower emissions
• Cleaner power
• Better grid planning
• More storage
• Less fossil fuel dependence

Those goals can overlap. But they do not always move at the same speed.

Right now, AI is moving faster.

Key Facts

• Musk spent years promoting a solar-electric economy through Tesla and SolarCity.
• xAI is now using natural gas turbines to power major AI data centers.
• Electrek reports xAI is operating 62 methane gas turbines across facilities in Tennessee and Mississippi.
• The facilities could emit more than 6 million tons of greenhouse gases per year, according to xAI permit applications cited by Electrek.
• xAI has reportedly spent heavily on Tesla Megapacks but not a significant amount on Tesla solar panels.
• SpaceX is now pitching space-based solar and orbital data centers.
• Critics argue ground-based solar remains cheaper and more practical than launching panels into orbit.
• The broader AI industry is facing the same conflict between climate commitments and electricity demand.

The Bottom Line

Elon Musk once made solar power on Earth sound obvious. Now his AI business is burning natural gas, and SpaceX is selling a future where solar moves into orbit.

That does not mean solar is dead. It means Musk’s priorities have changed.

AI needs power immediately. Gas gives it power quickly. Space solar gives SpaceX a bigger story for investors. Ground solar gives Tesla’s old mission credibility, but it may not move fast enough for the AI race Musk wants to win.

That is the contradiction.

The man who sold the solar-electric future is now choosing gas for AI and space solar for valuation.

The old mission was simple: stop burning fossil fuels.

The new strategy is messier: burn gas now, promise orbit later.

And that is why the criticism is growing.