Anthropic Locks Up SpaceX’s Colossus 1 Capacity. Is It Pulling Ahead in the AI Race?

Anthropic said on May 6 that it agreed to a deal with SpaceX to secure all the compute capacity at SpaceX’s Colossus 1 data center near Memphis. The move gives the company behind the Claude series of artificial intelligence (“AI”) models access to 300 megawatts within a month. That includes access to more than 220,000 graphics processing units (“GPUs”) from market leader Nvidia (NVDA).

While the deal may sound like business as usual, it’s the kind of move that, in retrospect, may feel like a strategic shift in the AI wars. Is SpaceX making a strategic pivot from its Grok AI chatbot?

But Anthropic’s press release gave second billing to this starring attraction.

Instead, it highlighted that the SpaceX tie-up and increased capacity would offer higher limits for Claude users. For example, Anthropic doubled Claude Code’s five-hour rate limits for certain plans and removed the peak-hour limit reduction for Claude Code on Pro and Max accounts.

That’s all good news for users, but the real story here is Anthropic’s acceleration in the AI race and SpaceX’s pivot. AI companies are in “land-grab” mode for computing resources. They’re racing to lock down the crucial data-center capacity needed to run their models over the next three to five years. Usage is expected to soar, and these AI companies need capacity in place now so they’re not constrained later.

Meanwhile, SpaceX is ceding computing capacity that might otherwise be used for Grok. The move suggests that SpaceX is less interested in developing the AI assistant – where growth has slowed in recent months – than it is in providing the raw computing power needed by AI.

It all comes as SpaceX’s Grok is losing steam. Downloads of Grok have fallen from a high of more than 20 million in January to 8.3 million in April, according to AppMagic. Moreover, in a survey of more than 260,000 U.S. AI users in the second quarter of 2026, just 0.174% paid for Grok, in line with a year ago. This compares with more than 6% for OpenAI’s ChatGPT.

Grok is doing poorly with enterprise users, the most crucial market segment. Their big bucks and the potential to lock in AI business for years make them a coveted segment. It’s one reason that OpenAI and Anthropic have been racing to sign up corporate clients. Both have formed big joint ventures with private equity firms to accelerate rollouts across their portfolio companies.

A SpaceX retrenchment on Grok could make sense, given the demands of SpaceX’s stable of capital-hungry businesses, including Starlink and Terafab. Still, it was only back in March that SpaceX CEO Elon Musk said, “SpaceX would far exceed everyone combined.” We’ll see…

Meanwhile, Anthropic is racing ahead now in key market segments such as enterprise AI. Nearly half (48%) of respondents to a March survey of some 500 people said their firm was using Anthropic’s Claude and planned to continue to use it, according to market research firm Enterprise Technology Research. That was a marked increase from 21% in the previous year.

Anthropic Is Rapidly Expanding Its Competitive Position

As part of the AI race, Anthropic has been quickly locking down computing capacity, beyond what it secured in its recent SpaceX deal.

It made a special point of noting its recent capacity deals in its latest release, highlighting:

• An agreement with Amazon (AMZN) for up to 5 gigawatts, with about 1 gigawatt of that scheduled by the end of 2026.

• A 5-gigawatt deal with Alphabet (GOOGL) and Broadcom (AVGO), slated to come online in 2027.

• A strategic partnership with Microsoft (MSFT) and Nvidia, including $30 billion worth of capacity.

• Anthropic’s $50 billion investment in American AI infrastructure with Fluidstack.

In their press releases, Anthropic and SpaceX paid lip service to an interest in a partnership “to develop multiple gigawatts of orbital AI compute capacity.” Clearly, SpaceX would set up this capacity in orbit while powering Anthropic’s models – and maybe Grok, too?

In any case, the SpaceX press release tipped its hand about the current state of so-called “orbital AI” and the likelihood of a deal. It noted optimistically, “If engineering challenges can be overcome, space-based compute offers near-limitless sustainable power with less impact on Earth.”

Tellingly, SpaceX’s release suggests that it sees itself more as a supplier of computing capacity. The first phrase – “If engineering challenges can be overcome” – reveals just how early in the process SpaceX is here. It’s not about financing, operations, sales, or even a real business model. It’s about engineering. In other words, whether it can be physically done at all.

Meanwhile, the announcement of more capacity – and the potential elbowing-out of a rival – comes as Anthropic’s Claude Mythos model created a fury over its ability to detect security vulnerabilities. The news has had Anthropic in the headlines for weeks, bringing massive attention to the startup.

Anthropic Looking to Raise More Money, Valuation Surging

All this expansion costs money, and Anthropic is going back to the well for even more cash. Back in February, the company raised $30 billion, and it’s currently in discussions to raise at least $30 billion more, at a valuation of more than $900 billion, according to Bloomberg.

It’s an opportune moment to raise as much money as possible for the money-hungry business model. Anthropic’s growing competitive position gives it a greater chance of becoming the dominant AI company – evidenced by a potential valuation that would exceed OpenAI’s. At the same time, investors have never been more receptive and willing to fund AI companies.

So, Anthropic appears to be clearly pulling ahead of OpenAI. In fact, we’re seeing signs that the investment case for OpenAI is becoming less attractive for sophisticated investors.

The valuation for Anthropic’s next fundraising round has been all over the place in recent weeks. April reports pegged a valuation round near $800 billion, while recent news suggested a $900 billion valuation or even a $1 trillion. Anthropic’s private valuation now sits near $1 trillion on Forge Global, a share exchange. In comparison, OpenAI’s last valuation hit $852 billion.

A new round at $900 billion would see Anthropic’s valuation more than double from just months ago, continuing a string of impressive valuation growth:

• November 2024: Raised $4 billion at a $40 billion valuation

• March 2025: Raised $3.5 billion at a $61.5 billion valuation

• September 2025: Raised $13 billion at a $183 billion valuation

• February 2026: Raised $30 billion at a $380 billion valuation

A new valuation at $900 billion would mean Anthropic’s value had increased about 22 times in a span of 18 months or so. Absolutely torrid.

The company is reportedly still considering an initial public offering (“IPO”) in 2026, with October mentioned as a tentative month for trading to begin. An IPO would boost the valuation even further, bring tens of billions more into the company’s coffers, and rank it among the largest IPOs of all time. Before public investors could buy it, the stock would have an even higher valuation – potentially $1.5 trillion, $2 trillion?

Even if Anthropic is the hottest thing going, serious investors will need to see a credible path to profitability before they plunk down real money on the stock.

Regards,

James Royal, Ph.D.

Marc Chaikin, the founder of Chaikin Analytics, has built an award-winning system that turned “bearish” on software stocks two months before they crashed this year. Now, he’s warning that one AI lab’s breakthrough could CRASH the Nasdaq while igniting a $500 trillion wealth transfer. This 60-year Wall Street legend has found a little-known $40 “pre-IPO backdoor” into the private startup behind this economic sea change. Click here for its name and ticker symbol before June 16.