According to new reports, SpaceX is targeting an initial public offering (IPO) that would value the company at $1.75 trillion -- the highest valuation for an IPO in history.
Earlier this year, the company combined with Elon Musk's artificial intelligence (AI) business, xAI, in an all-stock deal that valued the combined company at $1.25 trillion. Why combine a rocket and satellite company with an AI business so close to a potential IPO? The reason may surprise you.
SpaceX could upend the AI data center industry with one move
Bloomberg recently reported that, previous to its xAI acquisition, SpaceX was generating around $8 billion in annual profit on $15 billion to $16 billion of revenue. Between 50% and 80% of this revenue was generated by the company's Starlink satellite internet service. "Musk also expects to test launch a new version of SpaceX's next-generation Starship rocket in March, incorporating hundreds of upgrades, following a months-long launch hiatus while the company worked to resolve engineering challenges," the reporting concluded.
There are two major takeaways from this reporting. First, SpaceX's Starlink service has been a massive success. The Starlink division began launching test satellites in 2018. Actual deployment launches ramped up in 2019 and 2020. Six years later, the division is a highly profitable business with a global reach, potentially generating more than $10 billion per year in revenue. This is quite a feat. While Musk is known for his overly optimistic projections, the scaling of Starlink as a business outpaced many experts' predictions -- mine included.
In short, SpaceX was able to upend a decades-old industry over a handful of years. And given the company's plans for its Starlink service, further expansion will be pursued aggressively.
Image source: Getty Images.
The second thing to note from Bloomberg's reporting is the imminent restart of SpaceX's Starship rocket testing. The Starship rocket can theoretically carry a total payload of 100 to 150 metric tons. SpaceX's current rocket, the Falcon 9, can carry just 22 metric tons. If Starship testing succeeds, it could dramatically lower the price of bringing payloads into low-earth orbit. That should help the company scale its Starlink service faster than ever before, with lower costs and higher net profits.
But there's one other thing Starship could unlock for SpaceX: the ability to launch data centers into space for hyper-efficient compute infrastructure that can power the AI revolution.
One of the biggest problems in the AI world right now is the acquisition of energy. Data centers -- the infrastructure that makes AI development applications possible -- require huge amounts of energy to run. And while efficiency gains are being made, one of the main costs involves the electricity required to keep this hardware from overheating. That's why many data center start-ups aim to build infrastructure in cold places like Alaska and Greenland.
Space, meanwhile, is often -400 degrees Fahrenheit or colder. For running AI infrastructure, it could prove the most efficient place imaginable. The issue, of course, is getting giant data centers to space. With operational Starships, SpaceX could unlock this potential. Solar arrays -- the same kind that power space stations -- could theoretically fuel these space data centers, strongly mitigating their climate impact.
Of course, there are also some serious engineering challenges involved. Space, while cold, is also a vacuum. According to the Associated Press, this vacuum "trap[s] heat inside objects in the same way that a Thermos keeps coffee hot using double walls with no air between them." That's a big problem. "An uncooled computer chip in space would overheat and melt much faster than one on Earth," explains Josep Jornet, a professor at Northeastern University.
Still, Musk is unfazed. "Space-based AI is obviously the only way to scale," he recently wrote on SpaceX's website. "It's always sunny in space!"
As with Starlink, putting data centers into space will prove very challenging. But SpaceX's historic successes buffer these doubts. Musk's success with other businesses like Tesla (TSLA 2.97%) add credence to his claims.
Of course, the idea could prove very profitable if executed, which is likely why Musk is touting the possibility before the company's expected IPO. While I remain skeptical, data centers in space could ultimately change the game for the AI industry, solving the data center energy problem in a way few people could imagine possible.
