Elon Musk "teaming up" with Anthropic, Claude's "dumbing-down crisis" has a chance to be saved

Original Title: “Elon Musk and Anthropic, Going to Space to Find Power”
Original Author: Azuma, Odaily Planet Daily

Elon Musk and Anthropic, have surprisingly teamed up!

On the early morning of May 7th, Beijing time, a joint announcement from Anthropic and SpaceX instantly shocked the entire AI community. According to the announcement, both parties have signed a cooperation agreement, Anthropic will utilize the full computing power of SpaceX’s Colossus1 data center, providing over 300 megawatts of new computing capacity within a month (equivalent to over 220k NVIDIA GPUs), directly improving the user experience for Claude Pro and Claude Max subscribers.

This is the first time Elon Musk’s business empire has reached such a direct, formal, and large-scale cooperation with Anthropic. Over the past years, Anthropic’s core partners have always been Amazon and Google, whether in cloud infrastructure, chip supply, or model training, Anthropic has long been deeply integrated with AWS and Google TPU systems.

Musk has never publicly invested in Anthropic, and has even publicly criticized its AI safety approach and political tendencies multiple times. There had been almost no public record of infrastructure collaboration, model cooperation, or business alliances between the two before.

When the most capable large model company and the most talked-about world’s richest man finally connect, such a partnership naturally carries a lot of buzz.

Musk’s Little Tricks

For Musk, the more subtle aspect of this cooperation lies in the timing.

Just ten days ago (April 27), Musk’s lawsuit against OpenAI officially entered the trial phase in the Northern District of California federal court. In recent court appearances, both sides have exchanged barbs, with a tense atmosphere. This lawsuit, regarded by outsiders as the “biggest AI case of the era,” has almost fully publicized Musk’s long-standing grudge against OpenAI.

And Anthropic is precisely one of OpenAI’s most core and direct competitors at present. So this partnership naturally adds an intriguing layer—the enemy’s enemy is a friend; as long as it makes OpenAI uncomfortable, Musk will do anything.

From the perspective of AI competition dynamics, this cooperation also signals an evolution in Musk’s strategic layout in the AI era.

On the surface, it’s just a standard power trade—SpaceX provides GPU clusters, Anthropic gains more inference resources, each side gets what it needs. But clearly, there’s more to it.

Because Musk’s current AI-related activities are no longer just “building models.” Over the past two years, Musk personally launched xAI (which has now merged into SpaceX and renamed SpaceXAI), and the outside world increasingly views OpenAI, Anthropic, and other large model companies as Musk’s potential competitors. But with the construction and deployment of the supercomputing cluster Colossus, data center capabilities are spilling over, and Musk’s role has quietly shifted.

Today, Musk is increasingly resembling an “arms dealer” in the AI world—anyone lacking computing power can come to him, even former potential competitors.

Anthropic, embroiled in the “dumbing down” controversy, finally has a hero to save the day

For Anthropic, this cooperation might be even more significant than it appears.

In recent months, Claude’s reputation has been subtly changing. On one hand, Claude Opus 4.7 and the mysterious Mythos are still regarded as some of the best models on the market; but on the other hand, voices criticizing Claude’s “dumbing down” have become increasingly frequent in the community.

This sentiment is especially obvious among heavy developers. Some have found that Claude’s reasoning ability sharply declines when handling long code or complex engineering tasks; some reports indicate that Claude’s “thinking budget” or response length has been significantly reduced; more retail users report that Claude’s hallucinations are worsening, and the model is more prone to “seriously talking nonsense” when dealing with complex information.

After the “dumbing down” controversy intensified, Anthropic’s official released a technical review report, admitting that between March and April, due to product adjustments and bugs, Claude models experienced “performance degradation” in complex tasks.

But this explanation did not convince the market; public opinion still generally believes that high inference costs and lack of computing power are the main reasons for the performance fluctuations of large models like Claude in practical applications.

The stronger the model’s capabilities, the higher the inference costs; the larger the user base, the more terrifying the GPU consumption. All AI companies face a fundamental business dilemma—users want models to run “at full capacity” forever, but companies must control costs. So mechanisms like dynamic throttling, inference budget adjustments, response length controls, and priority scheduling almost inevitably appear, and users’ ultimate perception is “the model has become dumber.”

This is why the power collaboration with SpaceX is so critical for Anthropic.

Anthropic stated in the announcement that, with this cooperation, the core user experience of Claude will be directly improved:

· First, Anthropic will double the five-hour usage limit of Claude Code for Pro, Max, Team, and seat-based Enterprise plans.

· Second, Anthropic will remove the throttling measures on Claude Code for Pro and Max accounts during peak hours.

· Third, Anthropic will significantly increase the API rate limits for Claude Opus models.

AI model evaluation expert Alex Finn commented on this cooperation, “Anthropic has been somewhat dormant in recent months, with reduced quotas and a duller model… Now Musk comes to the rescue, giving Anthropic the chance to use the world’s largest supercomputing cluster.” Anthropic’s computing power crisis has long been a Achilles’ heel for the company, and market sentiment and user reputation have been declining accordingly. Musk’s one deal has solved this problem.

Alex Finn used a metaphor more familiar to American sports fans: “Musk’s help is like giving Anthropic a Vembanyama!”

The ultimate fantasy—going to space to find power

In the cooperation announcement, there was a small but often overlooked clause—“Both parties are also interested in jointly developing several gigawatt-level orbital AI computing power.” Translated into plain language, Musk and Anthropic want to move AI data centers into space. It sounds like science fiction, but the underlying issues are very real.

The top analyst in the AI circle, Aakash Gupta, explained: “On Earth, electricity, land, and heat dissipation capacity can no longer meet the rapidly growing demand.”

Anthropic has now secured about 15 gigawatts of computing capacity, equivalent to the power consumption of 11 million households, but that’s still not enough… Nvidia can produce chips, and Anthropic has enough money, but what cannot be manufactured on time are the electricity, land, and heat dissipation capabilities—while the growth rate of model demand far exceeds the speed of infrastructure development.

The frontier of the power race has now begun to leap off the planet. And only one company in the world has the capability to send gigawatt-scale solar arrays into orbit—SpaceX.

If this interstellar story finally becomes reality, only Musk might be able to do it.

Original link

Click to learn about the job openings at BlockBeats

Join the official BlockBeats community:

Telegram subscription group: https://t.me/theblockbeats

Telegram discussion group: https://t.me/BlockBeats_App

Twitter official account: https://twitter.com/BlockBeatsAsia