SpaceX acquisition xAI, "Marsk Economics" final

Original by Curly, Deep tide TechFlow

On February 2, SpaceX announced the acquisition of XAI。

The combined valuation was $1.25 trillion, up to which the Mask rockets, chains, X platforms, Grok were all under one roof。

Mask called the new company “the most ambitious vertical integration innovation engine on Earth and beyond”。

However, public data show that XAI earned $107 million in the last quarter, with a net loss of $1.46 billion, and burned nearly billions of cash in the first nine months, almost $1 billion per month. And SpaceX had a profit of about 8 billion in 2025。

BURNING 1 BILLION COMPANIES A MONTH INTO 8 BILLION COMPANIES A YEAR, THEN PREPARING AN IPO, PLANNING TO INTEGRATE $50 BILLION。

For xAI shareholders, the deal has a more practical name: lifeline。

Left to right

It's not the first time that Mask moved between his own company。

In March 2025, XAI acquired X. The combined valuation given by Mask is $11.3 billion, which is widely believed to be overestimated, because X's advertising revenues have not returned to Twitter age levels. But the focus of this transaction is not how much X is worth, or the real-time data generated by 600 million active users, all of which are injected into Grok's training pipeline, which is the largest exclusive resource for xAI relative to OpenAI and Anthropic。

In January 2026, Tesla spent $2 billion on round E finance for xAI. The reason is the ability to call Grok in depth, because the engine system and Optimus robotic training are needed. Tesla shareholders voted last year whether or not to invest in XAI, with more votes in favour than against, but the board finally failed to approve, and this time $2 billion went directly to the ground。

February 2, SpaceX acquired xAI. Tesla indirectly held SpaceX minority shares after the merger because of XAI holdings。

On the other hand, XAI has spent hundreds of millions of dollars over the past year buying Megapack batteries from Tesla to power the Colossus Supercentre. Tesla also procures a satellite chain service from SpaceX to provide a network of vehicles。

X has become the “dowry” of the Mask commercial empire, married to XAI in exchange for algorithms, and is now “comfort” with XAI. SpaceX is only able to complete the final narrative of the $1.25 trillion check at the June IPO feast。

There is a word in the investor circle: Muskonomy。

Tesla and XAI investors Ross Gerber told the truth: “It's like a bunch of high-value companies merged into a bigger high-value mess run by Elon. But in other words, it's now the Pure Musk Concept Unit, and you want to vote for Elon? Here it is."

IN A PUBLICLY LISTED COMPANY SYSTEM, THE TRANSFER OF ASSETS AND CONTRACTS BETWEEN COMPANIES UNDER THE FLAG OF THE SINGLE CONTROLLER USUALLY INVOLVES REGULATORY REVIEW. BUT MASK'S COMPANY IS ALMOST ENTIRELY PRIVATE. WITHOUT A PUBLIC DISCLOSURE OBLIGATION, WITHOUT CHECKS AND BALANCES ON AN INDEPENDENT BOARD OF DIRECTORS, SHAREHOLDERS ARE A SMALL GROUP OF VCS AND SOVEREIGN FUNDS AND ARE LESS SERIOUS THAN PUBLIC SHAREHOLDERS。

As long as the company's not on the market, this left-handed game can continue。

BUT AFTER IPO, THE STORY IS DIFFERENT。

xAI, live or die

SpaceX

There's only one thing at the core:HelpI don't know。

One billion dollars/month, which is the current rate of burning money。

$33 million per day, 1.4 million per hour, 23 million per minute. You read that, XAI burned another $60,000。

Where's the money? The majority of them went into Colossus, XAI, a super-calculable cluster built in Memphis, Tennessee. More than 200,000 Nvidia H100 GPU equivalents, target power 2 Giva. Billions of dollars were swallowed up by purchasing chips and batteries alone. Also, the first nine months of equity incentive spending was close to $160 million, and the price of AI's talent war was there。

But the income side can hardly see a match for the increase in the rate of burning money。

xAI 2025 revenue is expected to be around $500 million for the year, mainly from Grok 's API call and X Premium subscription share. The management ' s guidance to investors was to increase revenue to $2 billion in 2026 and to profit in 2027。

Two billion is a lot to hear, but OpenAI has earned more than 20 billion in 2025. Even if XAI achieves the target, the volume is only one tenth of OpenAI。

xAI, less than three years old, has experienced at least six rounds of financing, with valuations ranging from zero to $250 billion。

The most recent round was in January of this year, Round E, $20 billion, with investments from Nvidia, Valor Equity Partners and the Qatar Investment Authority. Together with the $5 billion debt financing previously arranged by Morgan Stanley, xAI has accumulated over $40 billion。

A company that loses billions of dollars each year and earns 500 billion dollars, with a value of 250 billion, equals 500 times the market sales rate。

IF THE IPO WAS SEPARATE, IT WOULD BE DIFFICULT FOR THE SECONDARY MARKET TO ACCEPT THE VALUATION。

ONE OF THE RULES OF THE IPO IN 2025 IS:Almost all of the companies are listed at lower transaction prices than in the last private round of valuation。

Compared to xAI, SpaceX is in exactly the opposite position。

It could be one of the most profitable private companies in the world. The United States is the only commercial rocket company that can routinely send astronauts to the International Space Station. The income of the chain has exceeded that of the rocket launch operation and is recurrent. 9 million paid users, paying monthly. This is the preferred business model for the secondary market。

But SpaceX has a problem listing alone。

WALL STREET'S VALUATION LOGIC IS BRUTAL: ROCKET COMPANIES COUNT CASH FLOWS, AI SELLS IMAGINATION。

Referring to the traditional airline giant Lockheed Martin, the market's PE multiplier has remained 20-30 times the same year; even with a 50-fold “hard nuclear technology premium” for SpaceX, the king of the unicorn, with a profit of $8 billion in 2025, the market value hovers around $40 billion, and even with a 100-fold PE premium for the New Space Economy, 80 billion。

But what about AI? OpenAI is looking for 83 billion new valuations and Anthropic is valuing 350 billion。

Mask wants more than 1 trillion。

Therefore, the financial logic of SpaceX combined with xAI is not complex:xAI alone could not sustain the valuation, and SpaceX's profits were pushed to the bottom and packed into a combination of “profits, growth stories, moats”。

FOR IPO UNDERSELLERS, THIS IS MUCH BETTER THAN SELLING SEPARATELY。

BUT MASK NEEDS A NARRATIVE THAT MAKES SENSE BY PUTTING ROCKET COMPANIES TOGETHER WITH AI。

So he found one:Space data centres。

SPACE DATA CENTER, NEW STORY FOR IPO

MASK WROTE IN HIS CONSOLIDATED STATEMENT: “AI'S PROGRESS DEPENDS ON LARGE GROUND DATA CENTRES REQUIRING HUGE AMOUNTS OF ELECTRICITY AND HEAT DISPERSION. GLOBAL AI ELECTRICITY DEMAND CANNOT BE MET BY GROUND-BASED PROGRAMMES, EVEN IN THE SHORT TERM PLACING A BURDEN ON COMMUNITIES AND THE ENVIRONMENT.”

There is a ironic background to this. The Colossus Supercentre of xAI was built in Memphis and local communities have been protesting its pollution. NAACP and EPA are ready to file a complaint。

Mask said that the ground data centre placed a burden on the community, as exemplified by his own data centre。

The solution he offered was to move the algorithm to space。

Powered by solar power, launched by SpaceX rockets, and data transmitted by satellite network of the chain。

Last Friday SpaceX submitted an application to the FCC requesting authorization to launch up to 1 million satellites to support the Orbital Data Centre project。

"I ESTIMATE THAT WITHIN TWO TO THREE YEARS, THE LOWEST COST OF GENERATING AI WILL BE IN SPACE." MASK SAID LAST MONTH IN DAVOS。

This vision is ambitious and early。

No company currently operated data centres in space. XAI's all the math is on the ground. Blue Origin of Jeff Bezos has announced a similar space backbone project, Google has a space data centre research project called Project Sunchacher. Both remain at the conceptual stage。

BUT IPO DOESN'T NEED A PRODUCT TO LAND, IT NEEDS A BIG ENOUGH STORY。

SpaceX is listed alone, the story of rockets and chains. It's good, but growing the ceiling is visible. The global commercial launch market is this big a year, and the growth of users of the chain will be saturated. Plus xAI, the story becomes "AI + space infrastructure," a trillion-billion-class narrative。

And with the space data centre vision, the narrative became..“The future of human calculus is in orbit, and we are the only company with rockets that can send it”。

THE DIFFERENCE BETWEEN THESE THREE LAYERS OF DOLLS IS HUGE FOR THE UNDERWRITERS AND THE ROAD SHOW。

AS FOR WHEN SPACE DATA CENTRES COULD RUN, THAT WAS AFTER IPO。

If space data centres were feasible, XAI would have infrastructure advantages that other AI companies could not replicate. OpenAI rents Amazon cloud services to split profits with Microsoft; Google negotiates electricity supply with state governments and deals with environmental screening. Mask doesn't need to, he has his own cloud running in space。

From the rocket launch (SpaceX) to the satellite network (stellar chain) to the data training (xAI) to the content distribution (X platform) to the application scene (Tesla Autopilot, Optimus robot), the entire industrial chain is controlled by Mask。

Tesla shareholders, "Gas" on Mars Road

There's an invisible loser in this capital game:Tesla shareholders。

The complaints of the Tesla shareholder community reached their peak. Multiple investors questioned social media: in 2020, Mask suggested that Tesla shareholders had priority over SpaceX, but in 2023, after XAI was established, Tesla's AI team was dug up and its computing resources were transferred to XAI. Now Mask requests Tesla to contribute 2 billion dollars to investment XAI, and to have Tesla shareholders hold indirect shares at SpaceX 1.5 trillion and XAI 250 billion。

Investors have left their books: by 2020, when SpaceX was committed, it was 100 billion, now 1 trillion, 10 times more; in 2023, XAI was set to value 10 billion, now 250 billion, 25 times more. "Priority of subscriptions" becomes "high-level transfer rights"。

Tesla is running out of money. There was also $44 billion in cash on the books, but there had been a decline in sales for two consecutive years. This week Tesla announced an investment of $2 billion in xAI and plans to double capital spending. Wall Street analyst predicts that Tesla may face a cash-flow deficit of $5-7 billion in 2026 due to the huge investment in AI infrastructure。

The timeline is telling:

• December 2025: xAI completed $20 billion in finance, valued at $23 billion
• January 2026: Tesla announced an investment of 2 billion to xAI
• January 30, 2026: SpaceX applied to launch 1 million satellites
• February 2, 2026: announcement of acquisition xAI

60 days of chain action. Mask is playing chess, Tesla shareholders are not on board, they're just chips。

In Muskonomy, in Mask, resources flow between companies, each of which creates new valuations。

However, Tesla shareholders found that their technology had been withdrawn, funds had been transferred and that they could only “buy back” these assets by indirect holding, far above the originally promised valuation level。

But on the other hand, Tesla would benefit from the success of Mask and SpaceX, a pure SpaceX concept unit。

A metaphor of image is:

“The Tesla shareholder is now like an ex-wife who was taken away by his ex-husband to start a business, and, while he was scolded by Mask for not speaking of Wuddeh and illegally misappropriated public funds, when he saw that SpaceX was really about to create a super IPO of 1.5 trillion, he could not help but remarry, hoping to get a family seat on that ticket to Mars. A complex mood that blends Stockholm syndrome with greed。

The re-evolution of Mask methodology

Mask wrote to the merged company mission:“Make a conscious sun to understand the universe, extending the light of consciousness to the stars.”

This phrase hides the bottom logic of Mask doing everything: turning immediate financial problems into long-term narratives with a vision that is too big to be false。

SpaceX survived this early. The vision that “humans must be multiplanetary species” supported the valuation when the company nearly went bankrupt when the rocket was fired three times。

Now the methodology has escalated。

Instead of using a company to tell a story, all companies should be bound together to tell a bigger story。

The rocket is for power, the chain is for transmission, xAI is for intelligence, X is for data, Tesla is for energy and robots, each piece is defective, but it's a picture of "Ai civilization in space" that becomes "unreal."。

Mask discovered the secret that, at the private company stage, valuation was largely narrative driven. As long as the story is big and far enough, investors are willing to believe it. SpaceX valuation went from billions to $80 billion, and Tesla went from near bankruptcy to 1.6 trillion, which is the logic。

However, individual company narratives always have ceilings. The rocket only talks about Mars, the electric car about autopilot. Once the ceiling has been touched, valuation growth has stalled。

The solution is to connect all corporate narratives and build a super narrative。

In this narrative:

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

Each story is problematic: SpaceX's Mars migrants are distant, XAI's Grok can't hit ChatGPT and Claude, Tesla's sales are falling, and X's advertising revenues are falling。

But together, these problems become “a vision of what is still being built”。

THIS PICTURE IS WORTH $1.25 TRILLION? IPO PRICING DAY IN MID-2026 WILL HAVE A PRELIMINARY ANSWER. IS AI PULLING UP THE ROCKET VALUATION, OR IS AI PULLING DOWN THE IPO。

By that time, the perceived sun will need to be read into the books。