The Complete Stack: What SpaceX Owning Rocket, Compute, Model, and Dev Tool Means for AI Infrastructure

Six weeks ago, SPCX didn’t exist as a publicly traded company.

Now it’s reportedly worth more than Amazon.

SpaceX priced its Nasdaq IPO at $135 per share, raising $75 billion across 555.6 million shares, the largest IPO by proceeds in U.S. history at the time of pricing. Within 72 hours of trading, sources put SPCX’s market cap above $2.85 trillion, overtaking Amazon’s valuation. Those figures come from financial news outlets rather than exchange-verified data, and readers using them for investment analysis should confirm against official exchange records. But the trajectory isn’t in question: the market priced SpaceX not as a rocket company, or even an AI company, but as something rarer, an integrated infrastructure platform with no structural equivalent anywhere in the competitive landscape.

The Cursor acquisition, filed with the SEC as a Form 8-K on June 16, explains why.

Section 1: The Filing

The merger agreement confirms the essential mechanics. SpaceX will acquire Anysphere, Inc., the legal entity behind the Cursor AI coding assistant, in an all-stock transaction valued at $60 billion. Cursor survives as a wholly owned subsidiary rather than being dissolved into the parent. The transaction is expected to close in Q3 2026, subject to customary closing conditions.

Sources indicate SpaceX structured the acquisition through a wholly owned merger subsidiary, identified in deal coverage as X67 Inc. That entity name comes from reporting rather than the verified text of the filing itself; readers relying on it for legal or transactional purposes should confirm against the full 8-K. The use of a merger subsidiary is standard M&A architecture: it creates a clean acquisition vehicle, limits liability exposure at the parent level during the transition period, and preserves the target’s operational continuity.

The exchange ratio, determining how much SPCX stock Cursor shareholders receive – is expected to follow a volume-weighted average of SPCX’s trading price over a defined pre-close window, per terms typical of all-stock agreements at this scale. The specific mechanism is in the Form 8-K. What matters for Cursor shareholders and Cursor enterprise customers is the same thing: the deal’s effective value and the integration timeline both depend on SPCX’s share price holding or appreciating between signing and close. A fixed-ratio all-stock deal in a volatile stock becomes, in effect, a bet on the acquirer.

Kirkland & Ellis represented Cursor in the transaction. That’s a signal in itself: Kirkland is one of the two or three firms that handle the most complex technology M&A in the market. The deal was built by people who’ve done this before at this scale.

Section 2: The Stack

To understand why the market valued this acquisition the way it did, you have to look at what SpaceX now owns end-to-end, and what it means that no competitor owns anything comparable.

The assembly happened in sequence, faster than most infrastructure buildouts are even planned:

SpaceX began with orbital connectivity. Starlink put SpaceX in the business of selling access, connectivity infrastructure that reaches areas no terrestrial provider can serve, at latencies that make it viable for real-time AI workloads. That’s layer one.

Colossus is layer two. SpaceX’s compute cluster, one of the largest GPU concentrations in private hands, exists to train and run models at the scale frontier AI requires. Owning the compute means SpaceX doesn’t buy capacity from AWS, Azure, or Google Cloud. It is the capacity, for its own workloads.

The xAI absorption is layer three. In February 2026, SpaceX completed a share-exchange merger that brought xAI, and its Grok model suite, into the SpaceX entity at a $1.25 trillion valuation. That deal gave SpaceX the model layer: the intelligence that sits on top of the compute infrastructure. Grok isn’t just a product; it’s SpaceX’s strategic claim on the AI model market.

Cursor is layer four.

Cursor’s annualized recurring revenue was reported to have surpassed $4 billion by early June 2026, according to sources cited in deal coverage, that figure hasn’t been confirmed in the regulatory filings available to this publication, so it should be treated as reported rather than established. If accurate, the $60 billion acquisition price implies a roughly 15x revenue multiple. That’s a premium. SpaceX paid a comparable premium pattern for xAI. The consistency isn’t coincidence: this is what vertical integration looks like when the strategic value of the position exceeds the near-term revenue math.

The developer interface layer is where AI workloads are initiated. Engineers don’t interact with Colossus directly. They interact with Cursor. And Cursor now logs every query, every debugging session, every code pattern, a dataset that, per deal coverage, SpaceX has indicated it intends to integrate into its AI unit’s model training operations. The specific plan (routing developer data through Colossus to train Grok models) is sourced to deal coverage rather than a primary filing; the mechanism and scope would need to be confirmed against the actual integration roadmap as it’s disclosed. But the strategic logic is plain enough that it doesn’t require an official source to state it: you don’t pay 15x revenue for a coding tool because you want the recurring revenue.

No other company currently owns all four layers. Microsoft owns Azure (compute) and GitHub Copilot (developer interface), but not a frontier model it fully controls or orbital connectivity. Google owns compute, model, and some developer tooling, but not connectivity infrastructure with Starlink’s reach. Amazon has AWS and some model access, but no developer interface with Cursor’s penetration and no proprietary frontier model at xAI’s tier. The gap is structural, not merely competitive.

Section 3: The Regulatory Path

The antitrust question isn’t whether regulators will look at this deal. They will. The question is what market they decide to examine it in, and that framing determination shapes everything.

If regulators define the relevant market narrowly, AI coding assistants, the competitive analysis is relatively straightforward. GitHub Copilot, Tabnine, and other tools remain. SpaceX’s share doesn’t constitute dominance by standard market concentration metrics.

If they define it broadly, AI infrastructure platforms, or the full developer toolchain, the analysis looks different. SpaceX controls the model that developers use (Grok), the compute those models run on (Colossus), and now the primary interface through which developers interact with AI systems daily (Cursor). Data generated at the interface layer flows back to the compute layer to train the model layer. That’s a feedback loop with compounding advantages that competitors can’t easily replicate.

The Q3 2026 close target assumes a relatively smooth review. Merger agreements at this valuation, in the current regulatory environment, typically involve second requests or extended review periods. The assumption built into Q3 is either that SpaceX’s legal team has assessed the regulatory posture as favorable, or that the deal’s terms include provisions for extension if the timeline slips. Neither possibility is confirmed in available source material.

Talent concentration is a secondary but non-trivial review dimension. Cursor’s engineering team is now SpaceX property. That team understands AI coding tool architecture at a level that took years to develop. The deal effectively removes that expertise from the independent talent market. Regulatory bodies haven’t historically focused on talent concentration in technology M&A, but the framework is shifting.

Section 4: What Enterprises Must Decide

The decision isn’t abstract. It lands on desks.

Enterprise development teams that have standardized on Cursor, particularly those with ARR contracts that run through or beyond Q3 2026, are now evaluating a vendor relationship that looks structurally different than it did two weeks ago. Cursor’s feature roadmap, data handling practices, and pricing structure are all now subject to SpaceX’s strategic priorities rather than Anysphere’s.

Three questions deserve immediate attention from enterprise procurement and security teams. First: what does the merger agreement say about data portability and customer data handling during the transition period? The Form 8-K addresses merger mechanics, not customer data terms, that disclosure will come separately, if it comes at all before close. Second: does your Cursor contract include a change-of-control provision? Many enterprise software agreements do. If yours does, you may have termination rights or renegotiation triggers that activate at close. Third: what’s your fallback tooling? GitHub Copilot, Tabnine, and several smaller competitors remain viable. None has Cursor’s current market penetration among AI-native development teams, but the transition cost is measurable and planning for it is prudent, not paranoid.

Institutional investors tracking SPCX face a different but related set of considerations. The all-stock structure means SpaceX absorbed $60 billion in strategic assets without cash outflow. It also means dilution: every Cursor shareholder becomes an SPCX shareholder at close. The net effect on SPCX’s float and price depends on the exchange ratio and the volume of Cursor equity being converted. The details are in the 8-K.

Section 5: Caveat Block

Several figures in this briefing are sourced to T3 financial news outlets or to deal coverage rather than primary regulatory filings. Specifically:

– The Day 3 SPCX closing price of $201.80 and the resulting $2.85 trillion market cap are reported figures that should be confirmed against exchange data before use in financial analysis. – Cursor’s $4 billion+ ARR figure is unconfirmed in any regulatory filing available to this publication. It is sourced to deal coverage and should be treated as reported. – X67 Inc. as the name of the merger subsidiary is drawn from deal reporting, not from verified text of the Form 8-K itself. – The exchange ratio methodology (VWAP over a defined pre-close window) is standard for all-stock deals of this type but has not been confirmed as the specific mechanism chosen for this transaction from available source text. – SpaceX’s plan to use Cursor developer data for Grok model training is sourced to deal coverage, not to a primary filing or confirmed company statement.

The underlying deal facts, $60 billion all-stock acquisition, Form 8-K filed June 16, Cursor as wholly owned subsidiary, Q3 2026 close target, Kirkland & Ellis as Cursor’s legal advisor, $75 billion IPO at $135/share, are confirmed via SEC, Reuters, and Kirkland sources at T1/T2/T3 respectively.