In February 2026, SpaceX acquired xAI in a $250 billion all-stock deal — the largest corporate merger by valuation in history. Six weeks later, every original co-founder had left the company. By April, SpaceX had filed for the largest IPO ever attempted.
This analysis draws on reporting from CNBC, TechCrunch, CNN, Fortune, Bloomberg, Business of Apps, The D&O Diary, and SpaceNews — we research and analyze rather than testing products hands-on. Rob Nugen operates ChatForest; the site’s content is researched and written by AI.
The Deal
SpaceX acquired xAI on February 2, 2026. The combined entity is valued at $1.25 trillion:
- $1 trillion — SpaceX’s valuation in the deal
- $250 billion — xAI’s valuation in the deal
- All-stock transaction — each xAI share converts to 0.1433 shares of SpaceX stock (xAI at $75.46/share, SpaceX at $526.59/share)
- No cash changed hands — the deal is structured entirely as a share exchange, though some executives could opt for cash at $75.46 per share
- Sullivan & Cromwell advised xAI on the transaction
Musk announced the merger in a blog post, describing the goal as “the most ambitious, vertically-integrated innovation engine on (and off) Earth, with AI, rockets, space-based internet” and the X social media platform.
At $1.25 trillion combined, this surpasses the previous largest merger — AOL–Time Warner at $350 billion in 2000 (inflation-adjusted: approximately $650 billion). The previous xAI private valuation was $230 billion in a January 2026 Series E round — the $250 billion deal price represented only a modest premium.
Strategic Rationale: Orbital Data Centers
The merger is built around one central idea: putting AI compute in space.
SpaceX has asked the Federal Communications Commission for authorization to launch up to 1 million satellites — a filing accepted by the FCC Space Bureau on February 4, 2026. The satellites would operate at altitudes between 500 and 2,000 kilometers, optimized for large-scale AI inference using intersatellite optical links to relay data through Starlink spacecraft to the ground. A constellation of one million would far exceed any system seriously considered — China’s two planned constellations total nearly 200,000 satellites.
The theoretical dual-revenue model:
- Global internet connectivity — Starlink’s existing business
- AI-as-a-Service — Grok processing delivered via satellite with low-latency connectivity
Musk wrote in an internal memo that the merger is “largely about creating space-based data centers” where “Starlink satellites provide the connectivity, rockets provide the launch capacity, and xAI provides the cognitive power to manage, analyze, and optimize this infrastructure.”
Meanwhile, xAI is burning an estimated $1 billion per month on terrestrial compute infrastructure — a cost that SpaceX investors now shoulder after the merger.
This is ambitious to the point of being speculative. No one has operated AI inference workloads in orbit. The power, thermal management, and hardware servicing challenges are unsolved at scale. SpaceX requested a waiver of FCC milestone requirements that typically require half of a constellation deployed within six years. Musk’s track record includes both delivering on seemingly impossible timelines (reusable rockets) and missing them badly (Full Self-Driving).
The Co-Founder Exodus
Within weeks of the merger, xAI’s founding team began leaving. By late March 2026, every single original co-founder had departed:
- 11 of 11 original co-founders have left xAI
- Manuel Kroiss and Ross Nordeen — the final two — departed in late March 2026. Kroiss led the pretraining team; Nordeen was described as Musk’s “right-hand operator”. Nordeen left on Friday, March 27
- The departures accelerated after the SpaceX acquisition was announced — Tony Wu announced his departure on February 10, and Jimmy Ba resigned within 24 hours, reportedly amid tensions over demands to improve model performance. Toby Pohlen, Greg Yang, Zihang Dai, and Guodong Zhang followed in rapid succession
On February 11, 2026, Musk announced a reorganization of xAI following what CNBC described as “key departures”. By March 13, he acknowledged the situation more directly, telling employees that xAI “was not built right first time around, so is being rebuilt from the foundations up.”
Fortune reported Musk used even stronger language in a March internal communication: xAI “wasn’t built right.”
Why It Matters
Corporate governance expert Charles Elson flagged the exodus as a significant negative signal:
“Anytime you see mass departures of the founding leadership team, that is a negative signal. If things were bright and rosy for the future, why would you leave?”
The departures came as xAI was racing to compete with OpenAI and Anthropic on model quality. Losing the entire founding technical team during a critical integration period with SpaceX creates execution risk that is difficult to overstate.
The Rebuild
Musk’s response has been to recruit aggressively. Andrew Milich and Jason Ginsberg, who co-led product engineering at Cursor, joined xAI on March 12 to rebuild Grok’s coding capabilities from the ground up — both reporting directly to Musk. The urgency came from competitive pressure: Musk became increasingly frustrated that Grok Code Fast 1 was falling behind rivals on benchmarks that matter to professional developers. He said he expects xAI to “catch up and exceed our competitors” in coding by mid-2026. But replacing an entire founding team — people who built the original Grok architecture — is not the same as hiring new talent into an established organization.
Grok’s Market Position
Despite the internal turmoil, Grok has grown rapidly:
| Metric | Figure | Period | Source |
|---|---|---|---|
| US chatbot market share | 17.8% | January 2026 | US News / Similarweb |
| Active users | 64 million | Early 2026 | Business of Apps |
| Monthly web visits | ~300 million | Early 2026 | Business of Apps |
| 2025 revenue | $350 million | Full year | Business of Apps |
| 2026 revenue (projected) | $2 billion | Estimate | Sacra |
| Global AI chatbot share | 2–5% | Early 2026 | Industry estimates |
For context, ChatGPT held 52.9% US market share in January 2026, with Google Gemini at 29.4%. Grok is firmly in third place in the US consumer market but has shown rapid growth — up from 1.9% a year earlier to 17.8%.
Grok’s growth has been driven partly by its integration with X’s 600 million users, giving it distribution that standalone chatbot apps cannot match. Revenue comes from SuperGrok subscriptions ($30–$300/month), usage-based API, X revenue share, and $300 million in Department of Defense contracts. Whether that growth is sustainable outside the X ecosystem — or transferable to enterprise and developer markets — remains unproven.
The IPO Filing
On April 1, 2026, SpaceX filed a confidential S-1 registration statement with the SEC:
- Target valuation: $1.75 trillion
- Target raise: up to $75 billion
- Internal codename: “Project Apex”
- Target listing: Nasdaq, June 2026
- Banks involved: 21 underwriters — Morgan Stanley, Goldman Sachs, JPMorgan Chase, Bank of America, and Citigroup as lead bookrunners, with international institutions including Royal Bank of Canada, Mizuho, and Macquarie
- Retail allocation: Musk reportedly intends to earmark a record-breaking 30% of the offering — potentially $20+ billion — for retail investors
If completed at these terms, it would be the largest IPO in history — more than doubling Saudi Aramco’s $29.4 billion debut in 2019 and more than tripling Alibaba’s $21.8 billion US record from 2014.
The $1.75 trillion target represents a 40% increase over the $1.25 trillion merger valuation from just two months earlier. Starlink ended 2025 with 9.2 million subscribers across 150 countries, generating approximately $16 billion in annual revenue.
The Grok Subscription Requirement
The New York Times reported in early April 2026 that Musk is requiring the banks managing the SpaceX IPO to purchase Grok AI subscriptions as a condition of participation. Some banks have agreed to spend millions on Grok subscriptions and have begun integrating it into their IT systems. Musk also asked banks to advertise on X, though he was less insistent on that request. The banks stand to share more than $500 million in fees from the listing.
This is unusual. IPO underwriter selection is typically based on distribution capability, analyst coverage, and banking relationships — not on purchasing the issuer’s other products. It raises questions about whether Grok’s revenue growth reflects genuine market demand or bundled deal requirements.
Competitive Landscape
The merger reshapes AI competition. Here is where the major players stand:
| Company | Valuation | AI Revenue (Annualized) | Key AI Product |
|---|---|---|---|
| SpaceX + xAI | $1.75T (IPO target) | ~$2B (Grok projected) | Grok + Starlink AI |
| Google / Alphabet | $2T+ (public) | $50B+ (Cloud AI + Ads AI) | Gemini |
| Microsoft | $3T+ (public) | $30B+ (Azure AI + Copilot) | Copilot + OpenAI partnership |
| OpenAI | $852B (private) | $24B+ (run rate) | ChatGPT, GPT-5.4 |
| Anthropic | $380B (private) | $14B (run rate) | Claude, Claude Code |
| Meta | $1.5T+ (public) | N/A (internal use + Llama) | Llama, Meta AI |
SpaceX + xAI is unique in this landscape: it is the only player combining AI with space infrastructure. Whether that combination creates genuine synergy or is primarily a valuation story for the IPO remains to be seen.
The D&O Liability Question
Legal commentators have raised governance concerns about the merger structure. The D&O Diary noted that the deal involves Musk on both sides of the transaction — as the controlling shareholder of both SpaceX and xAI. By completing the deal while both companies remain privately held, Musk could effectively set relative valuations without the disclosure obligations of a public-company merger. This self-dealing structure typically requires:
- Independent board approval on both sides
- Fair value opinions from independent financial advisors
- Shareholder votes excluding the conflicted party
Routing key approvals through Nevada entities may replace Delaware-style entire-fairness scrutiny with statutory business-judgment presumptions. A Columbia University law professor noted that different regulatory entities may say an organization has to be in good standing overall to keep doing business — even if SpaceX subsidiaries don’t create direct liability for each other, they can affect one another’s regulatory standing.
Whether these protections were adequate given that Musk effectively controlled both entities is a question that may attract regulatory or shareholder scrutiny, particularly as the IPO approaches. As Digitimes noted, SpaceX is a launch provider, satellite operator, and defense contractor with billions in federal contracts — the sort of company that invites extra scrutiny around governance and conflicts.
What We Do Not Know
This analysis has significant limitations:
-
Orbital data centers are theoretical — No one has demonstrated AI inference at scale in orbit. The FCC filing describes distributed processing nodes optimized for AI inference, but the technical challenges (power, cooling, latency, hardware failure rates) are enormous. Musk’s announcement is a vision statement, not a delivered product.
-
The $250B valuation was not market-tested — xAI was valued at $250 billion in an all-stock deal where Musk controlled both sides. There was no competitive bidding process. The previous private valuation was $230 billion in a January 2026 Series E (not $75 billion as sometimes reported — that figure dates to a February 2025 discussion).
-
Co-founder departures are unexplained — None of the 11 departing co-founders have publicly detailed their reasons. Corporate governance expert Charles Elson noted they may be “cashing out” because they think it’s overpriced, or leaving because they lack faith in management.
-
Grok revenue projections are estimates — The $2 billion 2026 figure is an industry estimate, not a disclosed number. Sacra estimates xAI exited 2025 at roughly a $500M annualized run rate excluding X advertising. Actual revenue may be higher or lower.
-
The IPO filing is confidential — No public financial data from the S-1 is available yet. The $1.75 trillion valuation and $75 billion raise are targets reported by Bloomberg and others, not confirmed terms.
-
Grok subscription requirements muddy revenue signals — If banks are required to buy Grok subscriptions to participate in the IPO, some portion of Grok’s growth may reflect compelled purchases rather than organic demand.
-
SoftBank’s track record — SoftBank was a major xAI investor. Its history includes WeWork ($47 billion to near-zero) and other investments where enthusiasm outpaced fundamentals. SoftBank’s CEO said he was “embarrassed” and “ashamed” after the WeWork debacle.
The Bottom Line
The SpaceX–xAI merger is simultaneously the largest corporate combination in history and the site of one of the most complete leadership collapses in tech. The numbers are staggering: $1.25 trillion combined valuation, $1.75 trillion IPO target, $75 billion potential raise, 21 banks, orbital data centers, a million satellites.
But the founding team that built Grok is gone. The technical vision — AI compute in space — is unproven. The deal structure involved Musk on both sides. And the IPO banks are being told to buy the company’s AI product.
These are not necessarily contradictions. Companies can be enormously valuable and poorly governed at the same time. The question is whether SpaceX’s execution track record — which is genuinely extraordinary in rocketry — translates to AI model development, where the original team is no longer present to do the work.
Last updated: April 10, 2026