The largest capital arbitrage in history.

Written by: Kuri, Deep Tide TechFlow

On February 2, SpaceX announced the acquisition of xAI.

The post-merger valuation is $1.25 trillion, bringing Musk's rockets, Starlink, 𝕏 platform, and Grok all under one roof.

In a statement on the official website, Musk referred to the new company as "the most ambitious vertically integrated innovation engine on and off Earth."

However, public data shows that xAI's revenue last quarter was $107 million, with a net loss of $1.46 billion, burning nearly $10 billion in cash over the first nine months, averaging close to $1 billion per month. Meanwhile, SpaceX's profit in 2025 is estimated to be around $8 billion.

A company burning $1 billion a month is being integrated into a company that earns $8 billion a year, and then preparing for an IPO, planning to raise $50 billion.

For xAI's shareholders, this deal has a more practical name: a lifeline.

Left Hand to Right Hand

This is not the first time Musk has shuffled things between his companies.

In March 2025, xAI acquired 𝕏. Musk's merger valuation was $113 billion, which many believed was overvalued, as 𝕏's advertising revenue had not yet recovered to the levels seen during the Twitter era. But the focus of this deal is not how much 𝕏 is worth; the real-time data generated by its 600 million active users is all funneled into Grok's training pipeline, which is xAI's biggest exclusive resource compared to OpenAI and Anthropic.

In January 2026, Tesla invested $2 billion in xAI's Series E funding round. The reason was the deep integration of Grok's capabilities, as both the vehicle system and the training of the Optimus robot require it. Tesla shareholders voted last year on whether to invest in xAI, with more votes in favor than against, but the board ultimately did not approve it; this $2 billion investment was directly realized.

On February 2, SpaceX acquired xAI. Tesla, holding shares in xAI, indirectly acquired a minority stake in SpaceX after the merger.

Meanwhile, xAI spent hundreds of millions over the past year purchasing Tesla's Megapack battery systems to power the Colossus supercomputing center. Tesla, in turn, procured Starlink services from SpaceX to provide internet for its vehicles.

X has already become the "dowry" in Musk's business empire, first marrying xAI for algorithmic recognition, and now "marrying into" SpaceX alongside xAI, all to complete the final narrative puzzle for that $1.25 trillion check at the IPO feast in June.

Money and people are circulating between Musk's companies, and there is a term in investor circles that specifically describes this system: Muskonomy.

Ross Gerber, an investor in Tesla and xAI, stated the obvious: "It's like a bunch of overvalued companies merging into a larger overvalued mess, run by Elon. But from another perspective, it's now a pure Musk concept stock; if you want to invest in Elon, it's all here."

In a publicly traded company system, transferring assets and contracts between affiliated companies under a single controlling entity usually invites regulatory scrutiny. But almost all of Musk's companies are private. There are no obligations for public financial disclosures, no independent board oversight, and the shareholders are a small group of VCs and sovereign funds who won't scrutinize things like public shareholders would.

As long as the company is not publicly listed, this left-hand-to-right-hand game can continue.

But after the IPO, the story changes.

xAI's Life and Death Crisis

Why does SpaceX want to swallow xAI?

Peeling away the grand narratives of space, the core reason is simple: to save itself.

$1 billion per month, that is xAI's current burn rate.

$33 million a day, $1.4 million an hour, $23,000 a minute. By the time you finish reading this paragraph, xAI has burned another $60,000.

Where is the money going? The vast majority is being poured into Colossus, the supercomputing cluster xAI built in Memphis, Tennessee. It has already installed over 200,000 Nvidia H100 GPUs equivalent in computing power, with a target power of 2 gigawatts. Just the procurement of chips and batteries has consumed tens of billions of dollars. Additionally, in the first nine months, stock incentive expenses were nearly $160 million, with the price of AI talent in the ongoing talent war.

But on the revenue side, there is hardly any growth that matches this burn rate.

xAI's total revenue for 2025 is expected to be around $500 million, mainly from Grok's API calls and 𝕏 Premium subscription revenue sharing. The management's guidance to investors is for revenue to grow to $2 billion in 2026, with profitability expected in 2027.

$2 billion in revenue sounds like a lot, but OpenAI's annualized revenue for 2025 has already exceeded $20 billion. Even if xAI reaches its target, its scale would only be one-tenth of OpenAI's.

Founded less than three years ago, xAI's valuation skyrocketed from $0 to $250 billion, going through at least six rounds of financing.

The most recent round was the E round in January this year, at $20 billion, with investors including Nvidia, Valor Equity Partners, and the Qatar Investment Authority. Coupled with the $5 billion debt financing arranged by Morgan Stanley earlier, xAI has raised over $40 billion in total.

A company losing tens of billions a year with $500 million in revenue, sporting a $250 billion valuation, equates to a 500x price-to-sales ratio.

If it were to IPO separately, the secondary market would find it hard to accept this valuation.

A pattern in U.S. IPOs in 2025 is: almost all companies' trading prices after going public are lower than the last round of private valuation.

In contrast to xAI, SpaceX's situation is completely opposite.

It may be one of the most profitable private companies in the world. It is the only commercial rocket company in the U.S. capable of regularly sending astronauts to the International Space Station. Starlink's revenue has already surpassed that of rocket launches and is a recurring income stream. With 9 million paying users, they pay monthly. This is the business model that the secondary market loves the most.

However, SpaceX also has issues if it goes public on its own.

Wall Street's valuation logic is harsh: rocket companies are valued based on cash flow, while AI companies sell imagination.

Taking traditional aerospace giant Lockheed Martin as a reference, the market typically maintains a PE multiple of 20-30; even if SpaceX, the unicorn king, is given a "hardcore tech premium" of 50 times, with its $8 billion profit in 2025, its market value would hover around $400 billion. Even with a "new space economy" premium giving it a 100x PE, it would only reach an $800 billion valuation.

But what about AI companies? OpenAI is seeking an $830 billion valuation while in a loss state, and Anthropic is valued at $350 billion.

Musk wants over $1 trillion.

Thus, the financial logic behind the merger of SpaceX and xAI is not complicated: xAI cannot sustain its valuation alone, and SpaceX's profits provide a safety net, turning the package into a combination of "profit, growth story, and competitive moat."

For IPO underwriters, this is much easier to sell than selling them separately.

But Musk also needs a narrative to make the combination of a rocket company and an AI company make sense.

So he found one: space data centers.

Space Data Centers: A New Story for the IPO

Musk wrote in the merger statement: "The advancement of AI relies on large ground-based data centers, which require massive amounts of power and cooling. The global power demand for AI cannot be met by ground solutions, and even in the short term, it will burden communities and the environment."

This statement has an ironic background. xAI's Colossus supercomputing center is located in Memphis, where the local community has been protesting its pollution issues. The NAACP and environmental organizations are preparing to file lawsuits.

Musk claims that ground data centers burden communities, and his own data center is a case in point.

His proposed solution is to move computing power to space.

Powered by solar energy, launched by SpaceX rockets, and using Starlink's satellite network to transmit data.

Last Friday, SpaceX submitted an application to the FCC requesting authorization to launch up to 1 million satellites to support the "orbital data center" plan.

"I estimate that within two to three years, the place with the lowest cost for generating AI computing power will be in space," Musk said last month in Davos.

This vision is grand and still in its early stages.

Currently, no company has ever operated a data center in space. All of xAI's computing power is on the ground. Jeff Bezos's Blue Origin has also announced a similar space backbone network plan, and Google has a space data center research project called Project Suncatcher. Both remain in the conceptual stage.

But an IPO does not require product realization; it needs a sufficiently large story.

If SpaceX goes public alone, the story is about rockets and Starlink. That’s already good, but the growth ceiling is visible. The global commercial launch market is limited, and Starlink's user growth will also hit saturation. With xAI added, the story transforms into "AI + space infrastructure," a narrative worth trillions.

Adding the vision of space data centers, the narrative becomes "the future of human computing power is in orbit, and we are the only company with rockets to send it up there."

For underwriters and roadshow PPTs, this three-layered nesting doll distinction is enormous.

As for when the space data center can be operational, that will be a matter after the IPO.

If the space data center is feasible, xAI will have an infrastructure advantage that other AI companies cannot replicate. OpenAI rents Amazon Web Services and has to share profits with Microsoft; Google negotiates power supply with state governments to deal with environmental reviews. Musk does not need to; he has his own cloud operating in space.

From rocket launches (SpaceX) to satellite networks (Starlink) to data training (xAI) to content distribution (X platform) to application scenarios (Tesla autonomous driving, Optimus robot), the entire industrial chain is under Musk's control.

Tesla Shareholders: The "Gas" on the Road to Mars

In this capital game, there is an invisible loser: Tesla shareholders.

The complaints from the Tesla shareholder community have reached a peak. Several investors have questioned on social media: in 2020, Musk hinted that Tesla shareholders would have priority subscription rights for SpaceX, but after xAI was established in 2023, Tesla's AI team was poached by xAI, and computing resources were redirected to xAI. Now Musk is asking Tesla to invest $2 billion in xAI while also expecting Tesla shareholders to hold indirect shares at the valuation levels of SpaceX at $1.5 trillion and xAI at $250 billion.

Some investors have done the math: when the promise was made in 2020, SpaceX's valuation was $100 billion, now it's $1 trillion, a tenfold increase; when xAI was established in 2023, its valuation was $10 billion, now it's $250 billion, a twenty-fivefold increase. The "priority subscription rights" have turned into "high-level takeover rights."

Tesla is running low on cash. There is still $44 billion in cash on hand, but the automotive business has seen a decline in sales for two consecutive years. This week, Tesla announced a $2 billion investment in xAI and plans to double its capital expenditures. Wall Street analysts predict that due to the massive investment in AI infrastructure, Tesla may face a cash flow deficit of $5-7 billion in 2026.

The timeline is telling:

• December 2025: xAI completes a $20 billion financing round, valued at $230 billion.
• January 2026: Tesla announces a $2 billion investment in xAI.
• January 30, 2026: SpaceX applies to launch 1 million satellites.
• February 2, 2026: Announcement of the acquisition of xAI.

A series of actions within 60 days. Musk is playing a big game, and Tesla shareholders are not on the board; they are merely chips.

In Musk's "Muskonomy" system, resources circulate between companies, and each circulation creates new valuation peaks.

But Tesla shareholders have found that their technology is being siphoned off, funds are being called upon, and in the end, they can only "buy back" these assets through indirect holdings at valuation levels far exceeding the initial promises.

On the other hand, Tesla will benefit from Musk and SpaceX's success, becoming a pure SpaceX concept stock.

A vivid metaphor is:

"Telsa shareholders now resemble an ex-wife whose savings were taken by her ex-husband to start a business. Although she curses Musk for being unethical and misappropriating funds, when she sees her ex-husband (SpaceX) is really about to pull off a $1.5 trillion super IPO, she can't help but rush to remarry, hoping to secure a family seat on that ticket to Mars. A complex emotion mixed with Stockholm syndrome and greed.

The Evolution of Musk's Methodology

Musk wrote a mission statement for the merged company: “To create a sentient sun to understand the universe, extending the light of consciousness to the stars.”

This statement encapsulates the underlying logic of everything Musk does: using a vision so grand it cannot be falsified to transform immediate financial issues into a long-term narrative.

SpaceX survived in its early days in a similar way. When rockets exploded three times in a row and the company was on the brink of bankruptcy, it was the vision of "humanity must become a multi-planetary species" that supported its valuation.

Now this methodology has been upgraded.

Instead of telling one story with one company, it is better to bundle all companies together to tell a larger story.

Rockets handle transportation, Starlink handles transmission, xAI handles intelligence, 𝕏 handles data, Tesla handles energy and robotics… Each piece has its flaws when looked at individually, but when pieced together into the vision of "space AI civilization," the flaws become "unrealized components."

Musk has discovered a secret: during the private company phase, valuation is primarily driven by narrative. As long as the story is big enough and far-reaching enough, investors are willing to believe. SpaceX's valuation rose from billions to $800 billion, and Tesla's rose from the brink of bankruptcy to $1.6 trillion, all following this logic.

However, the narrative of a single company always has a ceiling. Rockets can only talk about Mars, electric vehicles can only talk about autonomous driving. Once they hit the ceiling, valuation growth stagnates.

The solution is to connect the narratives of all companies to build a super narrative.

In this narrative:

SpaceX is the "space infrastructure operator," xAI is the "pioneer of human computing power in space," Tesla is the "carrier of robotics and energy ecology," and 𝕏 is the "training ground for real-time data."

Each story has its issues when viewed separately: SpaceX's Mars colonization is far off, xAI's Grok cannot compete with ChatGPT and Claude, Tesla's sales are in a downward trend, and 𝕏's advertising revenue has plummeted.

But when put together, these problems transform into "a grand vision still under construction."

Is this vision worth $1.25 trillion? There will be a preliminary answer on the day of the IPO pricing in mid-2026. Whether AI has boosted the valuation of rockets or dragged down the IPO of rockets will become clear.

By then, the sentient sun will need to shine through the financial reports.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。