Elon Musk recently announced his own company, which seems like an odd combination. SpaceX is acquiring xAI, an artificial intelligence company that also owns social media platform X.
The merger would bring together a profitable rocket company and a multibillion-dollar AI startup competing with OpenAI and other rivals for artificial intelligence supremacy.
The merger comes as SpaceX, whose main activity these days has been the launch of Musk’s Starlink broadband satellites, is preparing for an initial public offering, likely later this year.
The combined company will be valued at $1.25 trillion. This is higher than SpaceX’s recent individual valuation of $800 billion and xAI’s holding company valuation of $230 billion. The IPO stock price is expected to be approximately $525.
“I think this gives a story to the SpaceX and xAI merger, as a business that generates cash and a business that needs cash,” said Steve Coller, chairman of Swiss satellite maker Swissto12.
Mergers may make economic sense, but that’s not all. Here’s what you need to know about this plan.
This seems like an impossible marriage. Why is Mr. Musk doing this?
Apart from economic advantages, there are also technical reasons. Musk envisions a future where up to one million satellites, called data centers, could be sent into orbit and powered by sunlight day and night while they perform artificial intelligence calculations.
What exactly is a satellite data center?
The basic concept is simple. The satellite’s main function is to perform calculations on special chips powered by solar panels in a sun-synchronous orbit, where it is exposed to the sun’s powerful rays 24/7.
Aren’t hundreds of billions of dollars already being spent building AI data centers around the planet?
Musk claims that within three years, it will be cheaper to set up AI data centers in space than the hyperscale data centers currently being built on Earth, citing strain on the power grid, a timeline that some experts have questioned.
but it doesn’t start 1 Is it incredible that a million satellites will be in orbit?
Some people think so. In fact, the number of proposed launches is nearly 100 times the roughly 11,000 Starlink broadband satellites that SpaceX has sent into space with its flagship Falcon 9 since 2019.
But Musk is hoping to use the giant Starship rocket he is developing that can lift more than five times its payload.
“If we’re going to deploy a lot of additional satellites, there’s going to be a huge need for Starship,” Culler said.
Still, this seems like a tremendous number of launches.
There’s no doubt about that. In a blog post, Musk talked about launching millions of tons of satellites every hour into orbit every year.
But Collar said a network of AI data centers operating in space could be much smaller, with 10,000, 50,000 or even as many as 100,000 satellites. “With Elon, I think it’s kind of all-or-nothing,” he said.
How far along is the development of space-based AI data centers?
Not really. Google announced Project Suncatcher in November, saying it was considering an “interconnected network of solar-powered satellites” powered by its chips. And StarCloud, a one-year-old startup in Redmond, Washington, launched a demonstration satellite in November with backing from Andreessen Horowitz and other major venture capital firms.
“I think you’re talking about a 2040 time frame before we have any usable systems,” Culler said.
What are the technical challenges?
The first hurdle to overcome is developing computer chips that can operate in the extreme conditions of space, including exposure to much higher levels of radiation.
“If you want an Nvidia chip to work in space, you’ll probably have to invest several years and millions of dollars to do what’s called space certification, which makes sure there’s no radiation damage,” said Ehud Behar, an astrophysicist at the Technion (Israel Institute of Technology).
Are there any drawbacks to sun-synchronous orbit?
yes. Satellites are exposed to a lot of heat, and the way to dissipate heat generated by the sun and computer chips is to use large aluminum panels called radiators, which is inefficient, Behar said.
“You need to build a large surface to radiate heat. That’s part of the challenge. I don’t think these are small satellites like this. The mass has to be large enough to be cost effective,” he said.
How does SpaceX propose to bring all that data back to Earth?
In a filing with the Federal Communications Commission, SpaceX describes how the center will send data via optical or laser signals to the Starlink network and back to the ground while the entire system is in low orbit, about 300 to 1,200 miles above Earth.
Are there any technical challenges to that as well?
yes. Kaller said satellites in these orbits orbit the Earth approximately every 90 minutes, making it difficult for them to communicate with each other. A potentially better idea is to place terrestrial communications satellites in higher orbits, orbiting the Earth at a much slower pace, so that their transmissions to the ground cover a wider area.
“But this is all huge because launching into low Earth orbit is so cheap [trade-off] “That’s what’s really interesting,” he said.
In a blog post, Musk also talked about how everything will be done. these The data on the network he is building will eventually be accessible by mobile phone from anywhere on the planet. How would it work?
SpaceX bought $17 billion worth of radio spectrum from EchoStar in September and strengthened its partnership with T-Mobile, which allows customers to send text messages when they’re off the grid. A month later, it bought more EchoStar spectrum for $2.6 billion.
But MoffettNathanson telecommunications analyst Craig Moffett said Musk’s satellite network would be far less efficient than terrestrial cell phone networks given distance and would most benefit areas without access to existing networks.
Starlink doesn’t have “the right set of assets to be more than a niche player in the wireless market,” he said.
