SpaceX commits exclusively to Nvidia GPUs as Musk plans orbital AI data centers with Vera Rubin

Elon Musk says SpaceX will exclusively use Nvidia GPUs 'because they are the best' -- says optimized Vera Rubin NVL72 will be launched into space next year

No to AMD, Cerebras, D-Matrix, Rebellions, SambaNova, and Tenstorrent? Elon Musk on Tuesday said in an X post that SpaceX and xAI will exclusively use Nvidia GPUs because 'they are the best.' He later clarified during SpaceX's earnings call that Nvidia's Vera Rubin NVL72 rack-scale system's design is above everything else that is available today, which is certainly praise for Nvidia, but not such a good sign for other developers of merchant AI accelerators. "Going forward, we have decided to build exclusively on Nvidia, because we think the Vera Rubin architecture is the best architecture," Elon Musk said during SpaceX's earnings call. "We think it is the best AI computer, and we greatly value our close cooperation and partnership on many levels with Nvidia. We are exclusive to Nvidia. [...] We think the design of the NVL72 VR[200] computer is a much better design than, say, having a standard rack style design." Historically, xAI has exclusively used Nvidia's Hopper and, more recently, Blackwell hardware to train multiple generations of Grok. Although AMD has used Grok-1 on its Instinct MI300X accelerators, there has never been a public announcement or credible evidence that xAI has evaluated or used AMD Instinct, or other non-Nvidia accelerators in production. xAI's Colossus supercomputers have been using Nvidia's accelerators for years, so the official exclusivity looks more like a formality that gives a strong testament for Nvidia rather than a decision that was hard to make. When it comes to the praise of the cable-less design of compute trays in Nvidia's Vera Rubin NVL72 VR200 rack system, then Musk's admiration of this architecture is understandable, as while expensive, such trays greatly improve serviceability, assembly speed, and reliability by eliminating a large number of manual cable and hose connections that are common sources of installation errors and failures. Interestingly, but Musk's SpaceX plans to deploy Vera Rubin not only in its own and xAI's data centers, but also in space. "With respect to the Starmind AI satellite, which will be essentially an optimized Vera Rubin NVL72 computer, this is not some sort of far future distant thing; we expect to start launching this next year," Musk said. "We think the design of the NVL72 VR[200] computer is a much better design than, say, having a standard rack style design. We expect to actually deploy this on the ground as well as in orbit, because we think it is going to be a radical simplification of the normal NVL72 rack." Deploying an NVL72-scale machine will be by far a more ambitious project than Nvidia has in mind with its Space-1 Vera Rubin Module that is designed to deploy several, perhaps a dozen, of Rubin AI accelerators in space. 36 Vera CPUs and 72 Rubin AI GPUs offer rather formidable performance, though many questions remain about the cooling and reliability of such racks in space. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Elon pledges to give Nvidia a virtual monopoly over the stars

Nvidia commands about 85 percent of the datacenter GPU market today, and if Elon Musk has his way, the AI infrastructure giant will have a virtual monopoly over the stars. On Tuesday, Musk wrote on the social media network he owns that "SpaceX has committed to using Nvidia GPUs exclusively because they are the best." The collab should surprise absolutely no one. It's not like AMD or anyone else out there is making hardened versions of their GPUs for orbital datacenters. Why, you might ask? Well, to quote one Gartner analyst, the idea that space-based datacenters will ever be economically viable is "peak insanity." Nvidia just happens to have $13.2 billion in cash to burn on pipe dreams like these. Specifically, SpaceX says that it will be deploying Nvidia's Space-1, a specialized version of its upcoming Vera Rubin compute platform designed to operate high above the Earth's atmosphere where no one can hear you scream because the LLM is hallucinating again. In a separate X post Tuesday, SpaceX announced it was "partnering with Nvidia to design the Starmind AI1 satellite compute payload," and that "Each of the Starmind satellites will include Nvidia Rubin GPUs and Vera CPUs for datacenter class space compute." According to SpaceX's website, when fully deployed, the Starmind AI1 will measure 30 meters tall, have a wingspan of 75 meters from solar array to solar array, and support a compute payload of 250 kW. That might sound like a lot of power, but to put that in perspective, that's only enough power for roughly a single Vera Rubin NVL72 rack. So, if you thought Nvidia's rack systems were expensive, just wait till you have to loft them into orbit. Speaking of orbit, before any of this can happen, SpaceX needs to get its super heavy rocket, Starship, into orbit, which still hasn't happened. An even bigger challenge is making it cost-effective to lob the 3.33 ton satellites into orbit. It currently costs about $7,000 to put a kilogram into orbit aboard a Falcon 9 rocket. That puts the current price of launching the AI1 at north of $23 million, assuming you manage to pack the thing into a Falcon 9 in the first place. For orbital datacenters to be cost-effective, Starship is going to need to get the cost per kilogram down to Elon's stated goal of $10, Ortibal CEO told El Reg earlier this year. That would bring the cost to orbit to about $33,300, assuming the target is anything more than aspirational. Now is probably about the time we remind folks that Elon doesn't exactly have the greatest track record of following through on his "commitments." Remember how SpaceX was going to put humans on Mars by 2022 or how Tesla would field a million robo-taxis by 2020? But even if they don't go to space, Musk still might get something valuable out of Starmind AI1's development. In yet another X post, Elon opined that the satellite's architecture was so efficient that SpaceX properties would be deploying them, sans solar and thermal management systems, in their terrestrial datacenters. "It's a major improvement in data center efficiency," he wrote. The notes came on the heels of SpaceX's first earnings report as a public company, during which it said that capital expenditures grew by a factor of 6x versus a year ago to $18.4 billion as the company goes into overdrive building out xAI, the company's artificial intelligence offering. The company posted a quarterly net loss of $541 million on revenues of $7.81 billion. The stock was down more than 10% in mid-day trading on Wednesday, and is down 17% from its listing price of $135 in June. ®

Nvidia to build Starmind AI1 satellite compute payload for SpaceX

The project reaches far beyond a single satellite. SpaceX envisions an orbital network that could eventually grow to one million satellites, creating what would amount to a distributed AI supercomputer in space. If realized, the constellation would dramatically expand computing capacity outside Earth's data centers and rely on laser links to move data between satellites at extremely high speeds. SpaceX plans to power the first Starmind satellites with Nvidia's Vera CPUs and Rubin GPUs, hardware the chipmaker recently positioned as part of its emerging space computing portfolio. Elon Musk also endorsed the partnership, saying SpaceX has committed to using Nvidia GPUs exclusively because "they are the best." The partnership also marks SpaceX's addition to Nvidia's growing list of space computing collaborators. When Nvidia introduced its space platform in March, it named six launch partners but did not include SpaceX. Tuesday's announcement adds the aerospace company to that ecosystem as both firms push AI infrastructure beyond terrestrial networks. According to Nvidia, its Space-1 Vera Rubin module can deliver up to 25 times the AI processing performance of an H100 GPU. Commercial shipments of the platform are expected to begin later this year. Million-satellite vision SpaceX believes space offers several advantages for AI computing. Satellites operating in sun-synchronous orbit receive steady solar energy, providing a consistent power source without depending on Earth's electrical grid. Engineers also designed the Starmind architecture to remain hardware-flexible. Future satellites could replace onboard processors with newer chips from different suppliers without redesigning the entire spacecraft. Earlier this year, SpaceX asked the Federal Communications Commission for permission to deploy as many as one million orbital data center satellites between roughly 311 and 1,243 miles (500 and 2,000 kilometers) above Earth. Company filings describe a petabit-scale optical communications network that would allow satellites to exchange data using high-speed laser links.

SpaceX Executive Says Starship Could Unlock Orbital AI Data Centers as Earth Faces Power, Land and Coolin

Orbital AI Targets Earthbound Power Constraints Speaking at the ITU's AI for Good Global Summit 2026 on Sunday, Bednarek said Earth-based AI infrastructure faces growing resource pressures. "Space will become a solution to these problems. The sun has vast energy. It requires no water or land," she said, adding, "the universe is always clear." The International Energy Agency expects global data-center electricity consumption to roughly double to about 950 terawatt-hours by 2030, with power use by AI-focused facilities tripling. Bednarek said SpaceX combines frequent launches with Starlink-scale manufacturing. "The AI satellites we're currently designing and manufacturing are merely an evolution of the latest Starlink satellites and can be mass-produced very quickly," she said. Starship Unlocks Scale for Space Computing Starship is central to scaling that concept. SpaceX says its fully reusable rocket is designed to carry more than 100 metric tons to orbit. Bednarek said large orbital platforms ultimately require "megaton-class materials," adding, "Starship can make that a reality." SpaceX has asked the FCC to authorize up to 1 million solar-powered data-center satellites and aims to demonstrate orbital AI computing by late 2027. Broader deployments could begin as early as 2028. SpaceX Sees Orbital AI as Major Frontier Musk has called solar-powered orbital data centers a "no-brainer," predicting that space will become "the lowest-cost place to put AI" within two or three years. He has also argued that "space is the only way" to scale AI computing at extreme levels. SpaceX sees enormous potential. Its published plans envision Starship eventually delivering millions of tons to orbit annually and a factory capable of producing thousands of AI satellites beginning as soon as late 2027. Benzinga Edge Rankings show SpaceX stock fails to provide a favorable price trend in the Short, Medium and Long term. Price Action: SpaceX shares were up 2.77% to $136.80 during pre-market trading on Monday. Photo courtesy: Samuel Boivin / Shutterstock.com Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.

Elon Musk Commits SpaceX Exclusively To NVIDIA GPUs Citing "They're The Best", With 10GW Of AI Compute Coming By 2027

SpaceX's Founder & CEO, Elon Musk, has said that the company has committed to using NVIDIA GPUs exclusively. SpaceX Will Leverage NVIDIA GPUs "Exclusively," With Elon Saying That They're "The Best" On The Market Well, it looks like SpaceX has decided to use NVIDIA's GPUs exclusively to power its AI needs. In its latest earnings call, Elon Musk confirmed that NVIDIA's latest Vera Rubin platform will be used at SpaceX data centers for AI workloads, with deployments spanning across the globe and also in space. The NVIDIA Vera Rubin NVL72 rack, codenamed Kyber, will be the one powering SpaceX's AI data centers. Initially, a total of 2GW of compute capacity will be installed by this year's end, and a total of 10 GW of capacity is expected to come online by the end of 2027. This marks a big win for NVIDIA as it continues to gain major deployments for its Vera Rubin platform across big names in the AI cloud provider segment. More interestingly, the fact that SpaceX is already citing deployments in Earth's orbit is the bigger part of the story. NVIDIA has already announced its Space-certified Space-1 Vera Rubin module. It offers four Rubin GPUs, two Vera CPUs, and a hefty amount of memory. Some features of the NVIDIA Space-1 Vera Rubin module include: * Performance: The module offers up to 25 times the AI compute capability of the previous H100 GPU for orbital workloads. * Purpose: It enables real-time AI processing for geospatial intelligence, autonomous operations, and on-orbit analytics. * Architecture: It utilizes the same "Vera Rubin" architecture as NVIDIA's terrestrial chips for a unified development ecosystem. * Partnerships: The hardware is planned for use by firms such as Aetherflux, Axiom Space, and Planet Labs. * Application: It is intended for satellites and on-orbit servicing vehicles requiring high-performance AI inference in space. Previously, SpaceX had announced a partnership with Anthropic to build a multi-Gigawatt "Orbital" AI data center. SpaceXAI also believes that an orbital compute installation will address some of the major bottlenecks of terrestrial-based systems, such as Power, Land, and Cooling. With unlimited power and cooling, AI chips will then become a major bottleneck. Another bottleneck can arise from deliveries to space itself, as the reliance on commercial "outer space" shipments to move many hundreds of thousands of GPUs will be a task in itself. But at the pace at which AI is moving right now, the outer space AI Datacenter dream is far from a dream, and is aiming to become a reality very soon. Follow Wccftech on Google to get more of our news coverage in your feeds.

Elon Musk Envisions Massive Growth for SpaceX's Starlink Amid AI and Robotics Surge: 'It's Not Out of the

Musk said the rapid expansion of artificial intelligence and robotics could drive a massive increase in bandwidth demand, creating an opportunity for Starlink to capture a much larger share of the market. Musk Says Starlink Could Capture 50% of Internet Traffic Reacting to angel investor and All-In Podcast co-host David Friedberg's comments, Musk argued that the rise of AI and robotics will fundamentally change the economics of communications by creating an enormous new demand for data transfer. "Bandwidth demand will increase massively due to AI & robotics. Their need for data transfer is orders of magnitude more than humans!" he said. That could give Starlink room to expand well beyond its current satellite internet business, according to Musk. Even if the overall communications market merely doubles, he believes Starlink could capture a significant portion of it outside China. "Even if the communications market merely doubles in size, I would expect Starlink to reach at least 25% market share outside of China (maybe one day in China too), which would be over half a trillion dollars in revenue per year," Musk said. But his most aggressive projection goes considerably further. "It's not out of the question that Starlink carries more than 50% of Internet traffic long-term," Musk said, adding that such an outcome "would probably be over a trillion/year." "There is no obvious impediment to that outcome so far," he added. Friedberg Sees Starlink Becoming SpaceX's Cash Machine Friedberg argued that Starlink could become a trillion-dollar company within just 18 months. "Starlink alone could be generating on the order of $40 billion of revenue top line, with a huge amount of that flowing to free cash," Friedberg said. "That could be $30 billion in free cash flow within the year." Friedberg argued that Starlink's recurring subscription revenue could warrant a premium valuation, saying, "The Starlink business alone could be a trillion-dollar market cap within 18 months." The investor also pointed to Starlink's disruption of traditional satellite and telecom businesses. "You can see it in AT&T, Verizon, HughesNet, and Viasat... That market got decimated by Starlink," Friedberg said. He sees an even bigger opportunity if Musk moves ahead with a Starlink handset, which could accelerate subscriber growth. "Right now, he's adding 2 million subscribers on the consumer side per quarter. You could see that going to 4 to 5 million a quarter," Friedberg said. "You could actually see an acceleration." For Friedberg, the significance of Starlink goes beyond its standalone valuation. The business could generate enough cash to finance Musk's other ambitious projects. "And the rest of it is like, is Elon going to do well with the excess capital that's spinning out of Starlink?" Friedberg said. "I don't know who else I would give it to, considering what he's doing with Starship, AI compute, and the Terafab." SpaceX Increases Focus On AI Compute Capacity Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors. Photo courtesy: Shutterstock Market News and Data brought to you by Benzinga APIs To add Benzinga News as your preferred source on Google, click here.

Spacex and Nvidia Teaming Up on New Orbital Ai Data Center

Space Exploration Technologies, or SpaceX, CEO Elon Musk just made a huge announcement about his "Starmind" plan to launch a new network of orbital AI data centers. And he's teaming up with Nvidia to make it happen. According to an announcement on SpaceX's social media site X, SpaceX and Nvidia are working together to design the "compute payload" for Starmind AI1, the first in an orbiting network of satellites that will run AI workloads in outer space. One of the biggest expenses for a data center is electricity. SpaceX's goal for its orbiting Starmind AI data centers is to maximize power efficiency, thereby lowering compute costs. Because solar energy is more abundant in space than on the Earth's surface, solar panels on satellites that can be adjusted to face the sun are theoretically more efficient than terrestrial solar panels. Meanwhile, 30% to 40% of a terrestrial data center's power is devoted to running massive cooling systems that prevent the servers from overheating as they process the tremendous amount of data that AI requires. Because outer space is usually incredibly cold, keeping an orbital computer from overheating would theoretically require no electricity at all. This is all theoretical, though: although the vacuum of space is cold, you can't just put your entire computer into that vacuum or it will freeze. And while sunlight is stronger in space, even a large orbital solar panel can only produce a fraction of the electricity of a solar farm. Plus, in space there's no gravity, no air supply, and no technician who can replace a burnt-out fuse. Starmind will need to take all these factors into account. According to SpaceX's announcement, the first Starmind satellite will feature Nvidia's top-of-the-line Vera CPUs and Rubin GPUs, as well as its Vera Rubin NVL72 rackscale system. Nvidia, for its part, hasn't made any official comment regarding the scope of its participation.