You are using an out of date browser. It may not display this or other websites correctly. You should upgrade or use an alternative browser.
ScienceAnthropic, SpaceX announce compute deal that includes space development - SpaceXAI has signed an agreement with Anthropic to provide access to Colossus 1.
SpaceXAI has signed an agreement with Anthropic to provide access to Colossus 1, one of the world’s largest and fastest-deployed AI supercomputers.
Built from the ground up in record time, Colossus delivers unprecedented scale for AI training, fine-tuning, inference, and high-performance computing workloads. Colossus 1 features over 220,000 NVIDIA GPUs, including dense deployments of H100, H200, and next-generation GB200 accelerators. The cluster delivers extreme parallel performance for large language models, multimodal systems, scientific simulations, and generative AI at frontier scale.
Anthropic plans to use this additional compute to directly improve capacity for Claude Pro and Claude Max subscribers.
As part of this agreement, Anthropic also expressed interest in partnering to develop multiple gigawatts of orbital AI compute capacity.
The compute required to train and operate the next generation of these systems is outpacing what terrestrial power, land, and cooling can deliver on the timelines that matter.
SpaceX is the only organization with the launch cadence, mass-to-orbit economics, and constellation operations experience to make orbital compute a near-term engineering program rather than a research concept. If engineering challenges can be overcome, space-based compute offers near-limitless sustainable power with less impact on Earth.
As part of the agreement, Anthropic will get access to more than 300 megawatts of compute capacity, and it also “expressed interest” in working with SpaceX to develop multiple gigawatts of compute capacity in space. Anthropic said the deal will directly improve capacity for its paid Claude Pro and Claude Max subscribers.
The deal comes after Musk, who merged SpaceX with his competing AI startup, xAI, this year, has repeatedly criticized Anthropic because of its clash with the U.S. government. Musk has said Anthropic is “doomed to become the opposite of its name,” which would be misanthropic, and has asked if there’s a “more hypocritical company than Anthropic.”
But on Wednesday, Musk changed his tune. He said in a post on X that he spent a lot of time with senior members of the Anthropic team over the last week, and that he was “impressed.”
“Everyone I met was highly competent and cared a great deal about doing the right thing. No one set off my evil detector,” Musk wrote. “So long as they engage in critical self-examination, Claude will probably be good.”
In a separate post on X on Wednesday, Musk said that xAI “will be dissolved as a separate company” and will be called SpaceXAI.
XAI has been building out its data infrastructure, primarily around Memphis, as it races to compete in the booming AI market against OpenAI, Google and Anthropic. Colossus 1 is its largest facility to date.
When building and running Colossus, xAI and its subsidiary, MZX Tech, LLC, installed and operated dozens of natural gas-burning turbines to power the facility, claiming that no federal permit was required to do so because they were only for temporary use.
The turbines reportedly worsened air pollution problems around Memphis, and led to persistent protests against xAI’s work in the region.
Anthropic was founded in 2021 by a group of executives and researchers who defected from OpenAI. The company is best known for developing a family of AI models called Claude, and it’s currently in talks with investors about raising cash at a valuation of $900 billion, as CNBC previously reported.
Last month, Anthropic said that demand for Claude has led to “inevitable strain on our infrastructure,” which has impacted “reliability and performance” for its users, particularly during peak hours.
The company has been working to increase its capacity by inking several compute deals in recent weeks, including a multibillion-dollar agreement with Amazon.
The company has also been navigating a contentious few months of negotiations with the U.S. government.
In March, the Pentagon declared Anthropic a supply chain risk, blacklisting it from work with the U.S. military, after the two sides failed to reach an agreement on how the company’s models could be used. Anthropic sued the Trump administration in San Francisco and Washington to try to reverse the Pentagon’s actions, and that litigation is ongoing.
Meanwhile, the Defense Department has been embracing xAI’s Grok model, along with offerings from other AI companies.
But to what end?
Is the ultimate goal putting so much compute behind an AI to the point where you can ask it "make society post scarcity and perfect and make me built like prime arnold but keep it natty" and it'll magically somehow do it or is this all in the service of having two billion people go @grok PLEASE PUT HER IN BIKINI instead of one?
Am I just the retard in this equation or is it truly E=mc² + AI?
Someone (Altman?) when asked what the usecase/business case for AGI was replied with "We'll just ask it how to make profit" which is actually close to this.
In reality they don't know. But they might actually believe this. In the meantime, sell AI to businesses.
It is a significant challenge. The solar cells and radiators to do AIsats at scale will be huge. Not undoable, but huge. IMO we're going to need launch costs much lower than even foreseen with upcoming superheavy lift to make it financially make sense.
It is a significant challenge. The solar cells and radiators to do AIsats at scale will be huge. Not undoable, but huge. IMO we're going to need launch costs much lower than even foreseen with upcoming superheavy lift to make it financially make sense.
An awful lot is riding on Starship which sure is going slow. Elon needs to double whatever he's spending on it or something. I know he wants cheap but this is really dragging. New Glenn has a lot of potential and is nailing the landings, but the 2nd stage clearly has major problems getting things to final orbits.
That is one of the bigger issues but there are two big benefits: Infinite uninterrupted power / sunlight and no building permits/laws/citizens to oppose you, There are also other downsides like being unable to swap out failed hardware and cosmic rays however.
Long story story from what I've seen and read it's feasible-ish. And what with clown world being clown world all bets are off.
Its a terrible idea due to the costs associated with putting a data center's worth of equipment into earth orbit. Any cost savings in other areas are going to be wiped out by the costs associated with putting all this stuff into orbit.
The other continuous problem associated with these large AI clusters even on earth is their reliability. The more stuff you cluster together with equipment at this scale, the higher the probability is that something is going to be failing somewhere in the hardware. And the networks associated with these things are designed to a standard where just about not a single packet can be lost. A grain of dust getting into a fiber optic connector can easily become a serious problem. These things could never be made reliable enough to operate in scale in space beyond the ability to do repair or replace components as they fail.
The other absurdity in all this stuff is "line goes up" applied to AI training. Its been the case for a while that scaling up the training associated with AI isn't necessarily leading to a better experience with AI. The technical issues are perhaps elsewhere. But there is a giant industrial complex out there devoted to the religious idea that AI training at greater scale is the only thing that matters.
It will probably common place and worth while 100 years from now when much of industry is space/lunar based but now its just pants on head retarded grifting nonsense.
