By Laura Sanicola

Elon Musk's vision to put data centers in space got a $75 billion booster through the SpaceX IPO. The event may also be a watershed moment for the energy industry as Musk seeks to disrupt the AI power business.

Musk aims to launch one million satellites that would function like a vast data center in orbit. The idea is to use solar radiation that is far more available and more intense than on earth. The NIMBY problem plaguing terrestrial data centers wouldn't be an issue, assuming any aliens don't mind.

For now, the plan looks like a moonshot. Manufacturing so many satellites and solar panels is no small feat. SpaceX aims to launch multiple rockets a day with its Starship system, which is still in test phase. The economics only work if SpaceX can figure out how to get payload costs down sharply.

Yet analysts call the IPO a "scaling" event since it brought in so much cash to accelerate the project. SpaceX says it's aiming for a 2027-2028 time frame to start launching the satellites on a regular basis. And SpaceX isn't alone in this race; others developing space-based data centers include Alphabet and Jeff Bezos' Blue Origin.

What does the IPO means for power companies? Not much for now.

The AI industry is still short of power, demand is accelerating, and growth forecasts are moving up. Even if all goes as planned for SpaceX, the real threat may not come for at least decade, according to analysts and scientists we interviewed.

Yet SpaceX is now worth watching for its disruptive potential to the power industry. Long-term "terminal values" for power companies could start to wane, pressuring their multiples. Utilities, pipeline operators, and equipment makers like Caterpillar, GE Vernova, and Bloom Energy could be impacted.

Here's what to know about the IPO's impact and some investing takeaways.

Why put data centers in space?

The biggest obstacle to building on earth now is quick access to low-cost, reliable power. Large terrestrial campuses can spend years waiting for grid connections, turbine equipment, permits, and water. A public backlash is growing and costs are escalating.

Many of those obstacles are aren't a problem in space. And once the satellites are running, operating costs would be far lower than on earth.

The problem, of course, is getting the satellites into orbit without breaking the economics.

Musk's idea is to use Starship -- a fleet of massive, reusable rockets -- to launch thousands of satellites a year. Each one would carry a server rack, liquid ammonia for cooling and large radiators to dissipate the heat. Attached solar arrays would provide the power.

Musk's long-term ambition is to build as many as 10,000 Starship rockets. He has applied to deploy up to one million satellites that could eventually provide 100 gigawatts of orbital-computing capacity annually. That would be more than the entire installed base of U.S. nuclear power capacity -- every year.

The technical and practical hurdles are daunting, starting with Starship. The rocket has completed only 12 test flights. The underlying technologies have never been combined at data-center scale in space, and difficult problems such as cooling still need to be solved.

Data Center Power Demand Is Rising

Are solar panels up for the AI space job?

It's a tall order.

Nearly all satellites, including SpaceX's Starlink constellation, run on solar arrays. Historically, launching satellites that use those arrays was so expensive that buyers paid heavily for compact, highly efficient panels that could withstand radiation for many years.

That technology can't easily supply thousands of AI satellites, says Seth Hubbard, a Rochester Institute of Technology professor who has studied space solar for two decades. The higher-tech cells are costly, rely on constrained materials such as gallium and germanium, and are manufactured in relatively small volumes.

The only real alternative is silicon, the material used in ordinary panels and believed to power Starlink satellites. Silicon is cheaper and easier to manufacture, but it produces less power for its weight and degrades more quickly under space radiation.

"There's a big push to make silicon last longer in space, but right now silicon solar cells in low Earth orbit may last only one to three years, " Hubbard said in an interview.

Silicon is probably the only practical option near term, says Nicholas Rolston, an Arizona State University professor who studies solar materials. Newer perovskite cells could eventually be lighter, cheaper and more radiation-resistant, but they remain commercially unproven, he notes.

Overview Energy CEO Marc Berte thinks Musk will figure it out. Berte's start-up plans to collect solar energy in orbit and transmit it to users on Earth. Berte argues that falling launch costs could make cheaper silicon panels viable even if they require roughly three times the area of premium space cells.

"Betting against SpaceX is probably not the best idea," Berte says.

Will orbital data centers be cheaper than those on earth?

Not soon. Wood Mackenzie estimates that orbital data centers now cost at least three times as much as terrestrial ones, largely because of satellite and launch expenses. Those costs would need to fall by roughly 70% to reach parity, according to Robert Liew, the firm's director of global integrated research.

The timing depends heavily on Starship. SpaceX needs the rocket to become fully reusable, reliable and cheap enough to launch so many satellites. It must then repeat those launches at a pace no rocket operator has achieved.

"Making space solar cost-competitive with anything on Earth is very unlikely in the next 10 to 15 years," Rolston tells us.

Even then, orbital compute is more likely to supplement terrestrial data centers than replace them. Analysts don't see a meaningful threat to utilities and terrestrial power companies until the late 2030s or early 2040s.

Who else may benefit as SpaceX ramps up?

Industrial gas company Linde is in a good spot.

The firm is SpaceX's main supplier of liquid oxygen and nitrogen at launch sites in Texas and Florida. Linde supports roughly 70% of SpaceX launches and earns about $400 million a year from the space market, equivalent to 1% of its total revenue, according to Morningstar analyst Nicolas Owens.

SpaceX Beneficiaries

Starship requires far more propellant than its other reusable rocket, Falcon 9, and liquid oxygen makes up a large part of that load. Linde doesn't supply the methane that Starship burns, but Owens expects oxygen demand to be the main source of its growth.

A new Linde air-separation plant began operating near SpaceX's Starbase facility in Texas, earlier this year. Another expansion in Florida is scheduled to start up in 2027.

Owens expects Linde's space revenue to increase up to sevenfold by 2030. Space represents only about 1% of revenue today, but Owens estimates that it could grow to roughly 8% by 2040.

Shares are up 20% this year and trade near all-time highs around $515, close to analysts' average target.

Rocket Lab is a more speculative orbital power stock. The company produces premium, space-grade solar cells and panels. In February the company introduced lower-cost silicon arrays specifically designed for gigawatt-scale orbital data centers.

That gives it exposure to the space race, whether customers favor highly efficient traditional cells or cheaper silicon systems.

Investors are paying heavily for that possibility: shares trade at more than 70 times estimated 2026 revenue. Analysts don't expect positive free cash flow until 2029.

Musk's other big company, Tesla, is also a beneficiary as Musk keeps SpaceX vertically integrated. Tesla's battery systems have already racked up nearly 700 million of orders from SpaceX and xAI for terrestrial data centers.

What's the power outlook for data centers on earth?

Very healthy. Gartner expects data centers worldwide to use 702 terawatt-hours of electricity in 2027, up 57% from 2025. The firm expects the amount of power capacity needed to serve data centers to nearly triple from 104 gigawatts in 2025 to 290 gigawatts by 2030.

SpaceX itself owns two data centers: Colossus in Memphis, Tenn., and Colossus II in Mississippi. As of March, the campuses had enough installed servers to draw roughly one gigawatt of electricity, about as much as a large U.S. city.

SpaceX is using mobile natural gas turbines in Memphis while waiting for more grid power. Colossus II is still adding generation as it expands, and SpaceX has committed another $2.8 billion to turbine purchases through 2029, including a pending $2 billion order for mobile units and related equipment. Solaris Energy Infrastructure and Caterpillar are among its suppliers.

But don't expect SpaceX to announce more Colossus data centers. Its S-1 filing emphasizes orbital computing as a major source of future AI capacity.

Now that SpaceX is a publicly traded company, investors are likely to get far more detailed and regular updates on its progress. As the company keeps testing Starship and building out its solar and satellite manufacturing, investors in power stocks should keep a close watch.

OPIS: Plastics Prices Pull Back

The post-Iran War rally in plastics producers such as Dow and LyondellBasell may be losing momentum.

Producers secured price increases of roughly 40 cents a pound in March and April as the conflict disrupted supplies of key chemical building blocks, according to OPIS.

But some of those gains have evaporated. Higher prices weakened demand, while China -- a major buyer -- has increased production of coal-based chemicals, reducing its need for imports.

North American producers still retain an important cost advantage. Their plants largely run on natural gas liquids such as ethane, whose prices have barely moved since the war began as U.S. ethane production exceeds available export-terminal and shipping capacity.

(MORE TO FOLLOW) Dow Jones Newswires

June 15, 2026 05:55 ET (09:55 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.