Google signs $920M monthly compute deal with SpaceX — company’s projected annual data center revenue to exceed its combined proceeds from Starlink, launch services, and AI in 2025

News
By
Published

SpaceX is set to make big money from its data centers.

a data center surrounded by some homes
(Image credit: Getty Images)

SpaceX just announced that it closed a multi-year deal to provide compute capacity to Google. The agreement, which is worth $920 million per month, will begin in October 2026 and is expected to continue until June 2029. Reuters said that the transaction includes 110,000 Nvidia GPUs, plus CPUs, memory, and all other components needed for AI processing.

It appears that Elon Musk's company will not deliver the entire 110,000-strong GPU compute capacity in one go — Google will pay a reduced monthly fee as the company brings more server racks online through September 30, 2027. If SpaceX cannot hit the 110,000-GPU target on that date plus a one-month grace period, then Google can cancel the agreement or settle for the lower number of available GPUs “with a corresponding pro-rata reduction in the monthly fees.” It also gave the two parties the option to cancel the deal altogether after December 31, 2027, provided that they give a 90-day notice to the other.

This is the second major deal that SpaceX announced in months, as Anthropic secured the entire computing power of SpaceX’s Colossus 1 data center in early May. This was a surprising move, especially as Colossus 1 is one of the company’s most hyped assets, which Elon Musk launched in just 19 days. It turns out that launching it at such speed meant that it has a mix of H100, H200, and GB200 GPUs, which is resulting in efficiencies for training AI LLMs as the faster GB200 GPUs end up waiting for the older, slower GPUs before it can complete each computational step. Anthropic is instead using it for inferencing, especially as it is struggling to keep up with the demands of its growing user base.

Latest Videos From

The combined annual value of just these two deals is already worth more than SpaceX’s entire revenue for 2025. Reuters estimated that they would bring in more than $25 billion annually to the company, compared to the less than $20 billion that it made from Starlink, launch services, and AI revenue.

These massive deals, worth more than $70 billion in total, will lift SpaceX as it targets a $1.75 trillion IPO on June 12, 2026. While it started out as a space exploration company and is known for commercially launching satellites at a fraction of the cost compared to NASA and providing relatively affordable and stable satellite internet, it’s actively expanding towards orbital data centers. SpaceX acquired xAI earlier this year to help achieve that dream and has even filed some documents at the FCC detailing its plans. Google is also reportedly in talks with the company for a slice of the orbital data center pie.

Google Preferred Source

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Jowi Morales
Jowi Morales
Contributing Writer

Jowi Morales is a tech enthusiast with years of experience working in the industry. He’s been writing with several tech publications since 2021, where he’s been interested in tech hardware and consumer electronics.

23 Comments Comment from the forums
  • Notton
    I don't particularly care for either company, but SpaceX IPO might just tank everything.
    Is the AI bubble gonna burst?
    SpaceX biggest rug pull in history?

    https://finance.yahoo.com/markets/stocks/most-active/At the time of posting:
    Nvidia -13.56
    Intel -12.61
    Marvell -52.96
    Micron -131.99
    Tesla -27.45
    Amazon -7.76
    SMCI -5.26
    Broadcom -33.18
    AMD -56.82
    Microsoft -11.38
    #NotFinancialAdvice

    This guy seems to do a good breakdown of what just happened
    hMArdJ1qJT0
    Reply
  • Zaranthos
    Looking at the big picture of xAI, Tesla, and SpaceX it really looks like Elon has the potential to accomplish some pretty amazing things. I don't see much chance of an AI bubble bursting anytime soon because it has actually gotten more useful for the average person and despite the problems all the AI models continue to advance, improve, and get better. The expansion goals of AI have barely begun and the amount of compute that is expected to come online by all the major players is mind boggling. The sheer brute force power that compute will bring to bear to power future models is hard to comprehend. The combined goals of Musk's companies and his track record of completely breaking expected timelines gives him a huge potential advantage over most of his competition. What happens when Terafab comes online if it meets the planned goals? He definitely has an advantage over most of the competition when it comes to deploying orbital or space based data centers. Not sure anyone really knows how Tesla bots might factor into the equation either.

    I completely understand why some people are skeptical of AI, I was late to the game of even trying it because it was frequently so terribly bad in the early days and even sometimes still is. AI continues to get more useful even though you have to correct it, and fight with it for accurate data. One of the things that is likely to fix a lot of what's bad with AI is more compute so the algorithms can do their own parallel fact checking which has already happened in part with Grok which vastly improved its accuracy.

    In the future there may be hobbyist scientists who own Tesla bots (or something similar) that put them to work in a home lab reproducing scientific experiments to verify the validity of who knows what. Or they'll just be making paper airplanes of all kinds and testing them to debunk or prove claims made online. All this and more will be fed back into AI model data. Someone will have theirs out in their garden planting and growing new hybrid plants. Human imagination will still drive technology to accomplish new things, and technology will accelerate our ability to execute those ideas.

    I think it's an exiting time.
    Reply
  • Ron Olbrey
    Zaranthos said:
    Looking at the big picture of xAI, Tesla, and SpaceX it really looks like Elon has the potential to accomplish some pretty amazing things. I don't see much chance of an AI bubble bursting anytime soon because it has actually gotten more useful for the average person and despite the problems all the AI models continue to advance, improve, and get better. The expansion goals of AI have barely begun and the amount of compute that is expected to come online by all the major players is mind boggling. The sheer brute force power that compute will bring to bear to power future models is hard to comprehend. The combined goals of Musk's companies and his track record of completely breaking expected timelines gives him a huge potential advantage over most of his competition. What happens when Terafab comes online if it meets the planned goals? He definitely has an advantage over most of the competition when it comes to deploying orbital or space based data centers. Not sure anyone really knows how Tesla bots might factor into the equation either.

    I completely understand why some people are skeptical of AI, I was late to the game of even trying it because it was frequently so terribly bad in the early days and even sometimes still is. AI continues to get more useful even though you have to correct it, and fight with it for accurate data. One of the things that is likely to fix a lot of what's bad with AI is more compute so the algorithms can do their own parallel fact checking which has already happened in part with Grok which vastly improved its accuracy.

    In the future there may be hobbyist scientists who own Tesla bots (or something similar) that put them to work in a home lab reproducing scientific experiments to verify the validity of who knows what. Or they'll just be making paper airplanes of all kinds and testing them to debunk or prove claims made online. All this and more will be fed back into AI model data. Someone will have theirs out in their garden planting and growing new hybrid plants. Human imagination will still drive technology to accomplish new things, and technology will accelerate our ability to execute those ideas.

    I think it's an exiting time.
    Hey, optimism is not allowed on Musk articles. It's really just not allowed for anything, but mainly Musk articles. Thanks.
    Reply
  • wiscovitch
    Zaranthos said:
    Looking at the big picture of xAI, Tesla, and SpaceX it really looks like Elon has the potential to accomplish some pretty amazing things. I don't see much chance of an AI bubble bursting anytime soon because it has actually gotten more useful for the average person and despite the problems all the AI models continue to advance, improve, and get better. The expansion goals of AI have barely begun and the amount of compute that is expected to come online by all the major players is mind boggling. The sheer brute force power that compute will bring to bear to power future models is hard to comprehend. The combined goals of Musk's companies and his track record of completely breaking expected timelines gives him a huge potential advantage over most of his competition. What happens when Terafab comes online if it meets the planned goals? He definitely has an advantage over most of the competition when it comes to deploying orbital or space based data centers. Not sure anyone really knows how Tesla bots might factor into the equation either.

    I completely understand why some people are skeptical of AI, I was late to the game of even trying it because it was frequently so terribly bad in the early days and even sometimes still is. AI continues to get more useful even though you have to correct it, and fight with it for accurate data. One of the things that is likely to fix a lot of what's bad with AI is more compute so the algorithms can do their own parallel fact checking which has already happened in part with Grok which vastly improved its accuracy.

    In the future there may be hobbyist scientists who own Tesla bots (or something similar) that put them to work in a home lab reproducing scientific experiments to verify the validity of who knows what. Or they'll just be making paper airplanes of all kinds and testing them to debunk or prove claims made online. All this and more will be fed back into AI model data. Someone will have theirs out in their garden planting and growing new hybrid plants. Human imagination will still drive technology to accomplish new things, and technology will accelerate our ability to execute those ideas.

    I think it's an exiting time.
    Musk does have a "track record of completely breaking expected timelines"...Where's that FSD at?
    Reply
  • Zaranthos
    wiscovitch said:
    Musk does have a "track record of completely breaking expected timelines"...Where's that FSD at?

    Well where I live would benefit from FSD because I've been nearly killed by human drivers many times just driving my short commute to and from work. FSD without AI is nearly impossible because of "edge cases". I would also wonder if cameras are enough to truly meet political and regulatory safety scrutiny when dealing with unknowns like fog, heavy rain, blizzards, and even humans and animals running into traffic. Driving is pretty dangerous and humans fatally screw it up pretty often. It's also going to be hard for people to trust AI which probably isn't mature enough for the task yet either. Can older vehicles be programmed with software alone to deal with variability like road construction, map changes, accident lane closures, etc?
    Reply
  • alan.campbell99
    The big picture of xAI? Uh, xAI is one big reason why SpaceX operates at a loss after it was bolted on. Tesla is massively overvalued for what it is now, also don't forget the utter flop that is the cybertruck which was pushed by Elon. Of the three I'd say SpaceX on it's own has done the best, largely because the people in the company that actually know stuff have blunted his more moronic notions ala said Cybertruck.
    And no, cameras aren't enough. You need other sensors like LIDAR that most other manufacturers put in their designs, Elon won't because among other things he's cheaping out.
    It's not a question of Elon 'hate', it's more pointing out how the emperor has no clothes.
    Reply
  • Marlin1975
    Zaranthos said:
    Well where I live would benefit from FSD because I've been nearly killed by human drivers many times just driving my short commute to and from work. FSD without AI is nearly impossible because of "edge cases". I would also wonder if cameras are enough to truly meet political and regulatory safety scrutiny when dealing with unknowns like fog, heavy rain, blizzards, and even humans and animals running into traffic. Driving is pretty dangerous and humans fatally screw it up pretty often. It's also going to be hard for people to trust AI which probably isn't mature enough for the task yet either. Can older vehicles be programmed with software alone to deal with variability like road construction, map changes, accident lane closures, etc?


    Nothing you posted comes anywhere near what he said. You tried to parrot "track record of completely breaking expected timelines". which is wholly not true. If anything its the opposite.

    Musk is the king of over promising and pumping up and then moving on to another outlandish claim before being called out for the last one.
    Reply
  • wiscovitch
    Zaranthos said:
    Well where I live would benefit from FSD because I've been nearly killed by human drivers many times just driving my short commute to and from work. FSD without AI is nearly impossible because of "edge cases". I would also wonder if cameras are enough to truly meet political and regulatory safety scrutiny when dealing with unknowns like fog, heavy rain, blizzards, and even humans and animals running into traffic. Driving is pretty dangerous and humans fatally screw it up pretty often. It's also going to be hard for people to trust AI which probably isn't mature enough for the task yet either. Can older vehicles be programmed with software alone to deal with variability like road construction, map changes, accident lane closures, etc?
    Slick dodge...Again, where is the FSD that has been promised for almost a decade? I'm not denying the tech is hard, but you have to factor in that quagmire before saying "track record of completely breaking expected timelines" in some sort of positive way.
    Reply
  • Zaranthos
    Marlin1975 said:
    Nothing you posted comes anywhere near what he said. You tried to parrot "track record of completely breaking expected timelines". which is wholly not true. If anything its the opposite.

    Musk is the king of over promising and pumping up and then moving on to another outlandish claim before being called out for the last one.
    It's entirely subjective. People that hate Musk will form a completely different opinion than people who don't.

    Colossus 1 broke all records for timelines compared to other similar projects, and now it likely could break records for time to ROI if the deals with Anthropic and Google last. Tesla wasn't supposed to be possible either.

    Does Musk pontificate about all kinds of stuff, sure there is no denying that. He's also the richest human being on earth by net worth and likely has the potential to rapidly expand that wealth. I don't have the money to bet against him and neither do most of his haters.
    Reply
  • wiscovitch
    Zaranthos said:
    It's entirely subjective. People that hate Musk will form a completely different opinion than people who don't.

    Colossus 1 broke all records for timelines compared to other similar projects, and now it likely could break records for time to ROI if the deals with Anthropic and Google last. Tesla wasn't supposed to be possible either.

    Does Musk pontificate about all kinds of stuff, sure there is no denying that. He's also the richest human being on earth by net worth and likely has the potential to rapidly expand that wealth. I don't have the money to bet against him and neither do most of his haters.
    Well if it's subjective then what are we talking about here?!?
    Reply