YOU ARE AT:Network InfrastructureOpinion: The SpaceX-xAI mega-merger is an opening gambit to a bigger play

Opinion: The SpaceX-xAI mega-merger is an opening gambit to a bigger play

The deal creates the first-ever $1.25 trillion private company, marking a new chapter in AI compute race

Confirming previous reports, SpaceX just acquired Elon Musk-owned startup, xAI, that you may know as the maker of the Grok chatbot, in an all-stock deal. 

Although financial details are undisclosed, according to multiple sources, the combined company is valued at a whopping $1.25 trillion (SpaceX at $1 trillion and xAI at $250 billion) — making it the world’s most valuable private company.

What’s driving the mega-merger? 

In a missive on Tuesday, Musk said that the acquisition creates “the most ambitious, vertically-integrated innovation engine on (and off) Earth” that marks not only a new chapter for the two companies involved, but a new era of space-based AI for mankind. 

“Current advances in AI are dependent on large terrestrial data centers, which require immense amounts of power and cooling. Global electricity demand for AI simply cannot be met with terrestrial solutions, even in the near term, without imposing hardship on communities and the environment,” he said.

To make sense of what the deal accomplishes, one needs to look at what the two companies separately bring to offer and why combining their portfolios make sense. Musk’s SpaceX is an aerospace company — the biggest in the world — that also owns Starlink — the world’s largest satellite internet provider. 

xAI, by contrast, is a much smaller initiative, and is known for the social media platform X (a rebrand of Twitter which Musk acquired in 2022) and the aforementioned Grok chatbot which has recently come under fire for producing sexualized images of women. 

In a recent filing to FCC, SpaceX sought approval to build a “space cloud” that will constitute up to a million satellites — an Orbital Data Center system, if you will — that will support the world’s heaviest AI workloads with solar power and space cooling. 

All SpaceX’s current programs, including Falcon, Dragon, Starlink, and the new Starship that boosts 20x the capacity of Falcon, have the technological muscle to put this idea in motion.

So, if SpaceX is to turn into a full-blown AI data center company for which it has all the satellite technology and rocket launch capabilities, it only makes sense to bring its AI subsidiary into the fold for the AI firepower. 

The business angle

Investors are speculating whether the deal may complicate SpaceX’s upcoming IPO. Inside sources told the New York Times with the merger signed ahead of the IPO, Musk will likely take the combined company public — supposedly around June, with a target to raise $50 billion. 

Some say that merging the two firms makes it an easy sell to investors as it leaves Musk fully focused on one business as opposed to two. Others argue that it complicates the story because of the valuation mismatch. Ergo, bringing the highly profitable SpaceX and the cash-burning xAI under one roof is just bad business.

Swapnil Amin, Ex-Tesla director, says that the move is supposed to be a financial rescue. In a LinkedIn article, he writes, “Every headline says “Musk consolidates power.” That’s the wrong story. Here’s the real one: xAI burns $1B/month. It can’t IPO alone — not when OpenAI is raising $100B at $830B with 10x the revenue. So Musk did what he always does. He restructured the capital stack.”

With that, Musk can freely pour SpaceX’s capital into the laboring AI business and give it the boost of life it needs. 

Bigger worries

Separately, another camp is decrying Musk’s ambition to launch floating supercomputers in space. “The Singularity’s path to disassembling the Solar System for compute has found a corporate home,” writes investor Alex Wissner-Gross on LinkedIn.

Further, it seems like the project has some entry barriers to cross. While the space offers
an endless supply of solar energy and zero land footprint, it brings some steep challenges as well. The traditional air- and liquid-cooling systems used in terrestrial data centers (yes, that’s a term now!) are redundant in space. That leaves radiation as the only option to dissipate the waste heat from servers.

To cool an orbital data center with radiators, it will take a cooling deck the size of a tennis court, to say nothing of the trapped heat inside the server which will require its own internal liquid-cooling systems internally. 

Regardless of what engineering marvels that may lead to, for now everybody seems to be on the same page about one thing: the AI race is about to heat up real fast, real soon.

ABOUT AUTHOR

Sulagna Saha
Sulagna Saha
Sulagna Saha is a technology editor at RCR. She covers network test and validation, AI infrastructure assurance, fiber optics, non-terrestrial networks, and more on RCR Wireless News. Before joining RCR, she led coverage for Techstrong.ai and Techstrong.it at The Futurum Group, writing about AI, cloud and edge computing, cybersecurity, data storage, networking, and mobile and wireless. Her work has also appeared in Fierce Network, Security Boulevard, Cloud Native Now, DevOps.com and other leading tech publications. Based out of Cleveland, Sulagna holds a Master's degree in English.