Musk and Anthropic, surprisingly, join forces!
In the early hours of May 7th, Beijing time, a joint announcement by Anthropic and SpaceX sent shockwaves through the entire AI community. According to the announcement, the two parties have signed a collaboration agreement. Anthropic will utilize the full computing power of the SpaceX Colossus1 data center, which will provide Anthropic with over 300 megawatts of new computing power in one month (equivalent to over 220,000 NVIDIA GPUs), directly improving the user experience of Claude Pro and Claude Max subscribers.
This marks the first direct, formal, and large-scale cooperation between Musk's business empire and Anthropic. Over the past few years, Amazon and Google have always been Anthropic's key partners, deeply integrated with the AWS and Google TPU ecosystems in cloud infrastructure, chip supply, and model training.
Musk has not only never publicly invested in Anthropic but has also openly criticized its AI safety roadmap and political inclinations. There has been almost no public record of infrastructure collaboration, model collaboration, or business alliances between the two parties.
When the most powerful large-scale model company and the most talked-about world's richest man finally join hands, such a collaboration inherently attracts attention.
Musk's Calculations
For Musk, the subtlety of this collaboration lies in the timing.
Just ten days ago (April 27th), Musk's lawsuit against OpenAI officially entered the trial phase in the Northern District Federal Court of California, with both sides engaging in a heated legal battle in recent days. This lawsuit, considered by the outside world as the "largest case of the AI era," has almost turned Musk's and OpenAI's long-standing grudges into public spectacle.
And Anthropic happens to be one of OpenAI's core and direct competitors. Therefore, this collaboration naturally takes on an intriguing meaning — the enemy's enemy is a friend, and as long as it can make OpenAI uncomfortable, Musk is willing to do anything.
Moreover, if we delve deeper into the AI competitive landscape, this collaboration also foreshadows Musk's strategic evolution in the AI era.
Superficially, this appears to be a standard computational power transaction—SpaceX provides GPU clusters, Anthropic gains more inference resources, and both parties get what they need. But this matter is clearly not so simple.
Because Musk's current actions around AI are no longer just about "doing models." Over the past two years, due to Musk's personal involvement in xAI (now integrated into SpaceX and renamed SpaceXAI), outsiders increasingly view OpenAI, Anthropic, and other large-scale model companies as Musk's potential competitors. However, with the completion and commissioning of the Colossus supercomputing cluster, data center capacity has begun to overflow, and Musk's role positioning has quietly shifted.
Today, Musk is increasingly resembling an "arms dealer" in the AI world—anyone lacking computing power can come to him, even if the other party was once a potential competitor.
Anthropic, deeply embroiled in the "AI Decrement" controversy, has finally found a savior
For Anthropic, the importance of this collaboration may be even higher than imagined by outsiders.
Over the past few months, Claude's reputation has been undergoing subtle changes. On the one hand, Claude Opus 4.7 and the mysterious Mythos are still considered the best models on the market; but on the other hand, voices regarding Claude's "AI Decrement" have begun to appear more frequently in the community.
Especially within the heavy developer community, this sentiment is particularly evident. Some have found that when Claude is dealing with long code and complex engineering tasks, its reasoning ability experiences a "cliff-like decline." Research reports also point out that the "thought budget" or answer length of some Claude models has been significantly reduced. Moreover, more and more retail investors have provided feedback that Claude's hallucinations are becoming more severe, and the model is more likely to "seriously talk nonsense" when processing complex information.
After the "AI Decrement" controversy escalated, the Anthropic official has released a technical retrospective report, acknowledging that from March to April, due to product-level adjustments and bugs, the Claude model experienced "performance degradation" in complex tasks.
However, this explanation did not convince the market. Public opinion still widely believes that high inference costs and a shortage of computing power are the main reasons for the performance fluctuations of large models like Claude in practical applications.
The stronger the model, the higher the inference cost; the larger the user base, the more terrifying the GPU consumption. Every AI company faces a commercial dilemma—users want the model to run at full capacity forever, but the company must control costs. Hence, dynamic throttling, inference budget adjustment, answer length control, priority scheduling... These mechanisms almost inevitably appear, and the user's ultimate perception is that the "model is dumbed down."
This is also why the computing power partnership with SpaceX is so crucial for Anthropic.
In the announcement, Anthropic stated that, with the partnership in place, the core user experience of Claude will be directly improved:
· First, Anthropic will double the five-hour usage quota for Claude Code in the Pro, Max, Team, and seat-based billing Enterprise plans.
· Secondly, Anthropic will lift the throttling measures for Claude Code in Pro and Max accounts during peak hours.
· Thirdly, Anthropic will significantly increase the API rate limit for Claude Opus models.
AI model evaluation expert Alex Finn commented on this collaboration, saying, "Over the past few months, Anthropic has been somewhat struggling with reduced quotas, model dumbing down... and now Musk is coming to the rescue, giving Anthropic the opportunity to use the world's largest supercomputing cluster." Anthropic's computing power crisis had always been the Achilles' heel of the entire company, leading to a decline in user reputation and market sentiment. Musk, solely through a collaboration, has solved this issue.
Alex Finn used an analogy that American sports fans might find easier to understand: "Musk's help is equivalent to directly giving Anthropic a wildcard entry into Wembley Arena!"
Final Fantasy—Space Odyssey for Electricity
In the partnership announcement, there was a small detail that many people overlooked— "Both sides are also interested in collaborating to develop gigawatt-level orbital AI computing power." In plain English, Musk and Anthropic want to move AI data centers to space. This may sound like science fiction, but the underlying issue is very real.
The AI industry's most influential analyst, Aakash Gupta, explained, "The Earth's available electricity, land, and cooling capacity are no longer able to keep up with the demand."
Anthropic has now locked in approximately 15 gigawatts of computing power, equivalent to the electricity consumption of 11 million households. However, this is still not enough... While NVIDIA can produce chips and Anthropic has enough money, what they truly cannot manufacture on time is electricity, land, and cooling capacity. The rate at which model demand is growing far exceeds the pace at which these infrastructures can be built.
The forefront of the computing power race has now started to venture beyond Earth. And globally, there is only one company truly capable of deploying gigawatt-level solar arrays into orbit at scale—SpaceX.
If this interstellar, saga-like story eventually becomes a reality, the only person who may be able to achieve it is Musk.
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia
