TL;DR

SpaceX exercised its option on June 16 to buy Anysphere, the maker of Cursor, in a $60 billion all-stock deal expected to close in Q3 2026. The acquisition gives SpaceX a profitable AI coding app and developer distribution channel, but the central test remains whether Grok can become a leading foundation model.

SpaceX exercised its option on June 16 to buy Anysphere, the company behind the AI coding agent Cursor, in a $60 billion all-stock deal that would make Cursor a wholly owned SpaceX subsidiary and give Elon Musk’s company a major revenue-generating AI application on top of its compute, power, model and distribution assets.

Under the reported terms, each Cursor share will convert into SpaceX Class A shares, with closing expected in the third quarter of 2026. SpaceX had previously secured an option to buy Cursor for $60 billion or pay a $10 billion alternative fee; this week, it chose the acquisition route.

Cursor brings SpaceX one of the strongest commercial products in AI software. Founded in 2022 by four MIT graduates, Anysphere reached about $4 billion in annualized revenue by early June, according to the source material, up from roughly $2 billion in February. Its growth has come in AI-assisted coding, one of the few AI software categories where companies are paying at scale.

The deal follows signs of closer technical ties between Cursor and Musk’s AI operations. Cursor trained its newest model on tens of thousands of xAI chips, according to the source material, and two senior Cursor engineers had already moved to xAI. Cursor CEO Michael Truell framed the transaction as a joint effort to build “the world’s most useful AI models,” with a co-trained model expected to ship into Cursor and Grok Build soon.

AI Dispatch · Infrastructure & Strategy

SpaceX owns every layer
of AI now

The $60B Cursor buy completes the stack: power, compute, research, model, app, distribution. But owning every layer isn’t winning every layer — and the model is the weak one.

$60B
all-stock · Cursor
(Anysphere)
The stack, layer by layer
06
Distribution
X · Tesla · Optimus · Cursor’s developer base
Strong
05
Application — Cursor
~$4B annualized revenue · just acquired
Bought
04
Model — Grok  ← the weak link
Underdelivered vs compute; training moved to Colossus 2
Weak
03
Research — xAI
Folded into SpaceX, Feb 2026
Mid
02
Compute — Colossus 1 & 2
~555K GPUs · orbital data-center plans filed
Dominant
01
Power
On-site gas generation, built faster than utilities interconnect
Dominant
The landlord pivot — renting Colossus 1 to rivals
Colossus 1 · Memphis
220,000+ GPUs · 300 MW
xAI couldn’t parallelize Grok on its mixed H100/H200/GB200 build, so it moved training to Colossus 2 and leased the rest out.
⚠ ran at ~11% utilization — “embarrassingly low”
Anthropicthru May 2029
$1.25Bper month
Googlethru June 2029
$920Mper month
combined ≈ $26B / year in compute revenue
122
days to build the first 100K-GPU cluster
~555K
Nvidia GPUs across the Memphis site
~2 GW
total power capacity
~$18B
in silicon (phase 1 alone ~$4B)
The take

You can buy a coding app and a model team. You can’t buy the research lead that makes your foundation model the one everyone else builds on — which is why Anthropic pays Musk $1.25B/month, not the other way around. Owning every layer bought SpaceX the right to attempt the hard thing. It hasn’t done it yet.

Sources: SpaceX S-1 & SEC filings; WSJ; Reuters; CBS; TechCrunch; Forbes; Business Insider; Introl; Built In (Feb–Jun 2026). Lease figures per SpaceX filings; utilization per a reported internal xAI memo.
thorstenmeyerai.com

A Full AI Stack Test

The acquisition matters because SpaceX is assembling an unusually integrated AI business. The company now has exposure to power generation, large-scale GPU infrastructure, a research lab through xAI, the Grok model line, major consumer and enterprise distribution through X, Tesla and Cursor, and now a profitable developer product.

That structure could reduce dependence on outside infrastructure providers and give SpaceX direct control over how models are trained, packaged and sold. It also changes the competitive map: rivals such as Anthropic and Google may compete with SpaceX in models while also depending on SpaceX-controlled compute capacity.

But the deal also highlights the gap SpaceX has yet to close. Owning the stack does not mean leading every layer. The source material describes Grok as the weaker part of SpaceX’s AI position, saying it has underdelivered relative to the scale of compute available. That judgment remains an analysis, not a settled market verdict.

Agentic Coding with Claude Code: The everyday developer's guide to agentic coding with Claude Code

Agentic Coding with Claude Code: The everyday developer's guide to agentic coding with Claude Code

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Cursor Fills The App Gap

Before the Cursor deal, SpaceX’s strongest AI asset was infrastructure. Its Colossus supercomputer complex in Memphis reportedly grew from the first 100,000-GPU cluster to 200,000 GPUs within months, and the broader site is described as having about 555,000 Nvidia GPUs across H100, H200 and GB200 systems, with roughly 2 gigawatts of total power capacity.

Power has become a limiting factor for AI expansion, and SpaceX has leaned on on-site gas generation to move faster than utility interconnections. The source material says the initial Colossus phase cost as much as $4 billion, with silicon across the expanded site pegged near $18 billion.

SpaceX’s model strategy changed earlier in 2026 when xAI was folded into SpaceX, according to the source material. Grok became the company’s foundation-model layer, while Cursor now adds a paid software surface used by developers. That makes the acquisition less about adding another AI brand and more about connecting compute, model training, product usage and revenue.

“the world’s most useful AI models”

— Michael Truell, Cursor CEO

Building AI-Powered Products: The Essential Guide to AI and GenAI Product Management

Building AI-Powered Products: The Essential Guide to AI and GenAI Product Management

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Grok Remains The Open Question

The main unresolved issue is whether SpaceX can turn its infrastructure advantage into a model advantage. The source material says xAI struggled to parallelize Grok training on a mixed H100, H200 and GB200 build at Colossus 1, then moved training to Colossus 2 and leased remaining capacity to other companies. That account is attributed to reported internal material and has not been independently detailed in the provided source text.

It is also not yet clear how quickly Cursor and Grok Build will share a co-trained model, how Cursor’s existing customers will respond to SpaceX ownership, or whether regulators will review the deal closely given SpaceX’s position across compute, software and distribution.

The economics of the compute leases are another area to watch. The source material cites leases to Anthropic and Google totaling about $26 billion a year in compute revenue, but the durability of that revenue depends on contract terms, capacity delivery, customer demand and whether rivals keep renting from a direct competitor.

AI in Embedded Systems: Types, Techniques, Machine Learning, Model Training vs. On-device Inference, Algorithms, Frameworks and Tools.

AI in Embedded Systems: Types, Techniques, Machine Learning, Model Training vs. On-device Inference, Algorithms, Frameworks and Tools.

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Deal Closing And Model Delivery

The next milestone is the expected closing of the Anysphere acquisition in Q3 2026. Until then, investors, developers and AI competitors will watch for regulatory filings, integration details and any customer reaction from Cursor’s developer base.

After closing, the test moves from dealmaking to execution. SpaceX will need to show that Cursor can keep growing under new ownership, that Grok can improve against rival models, and that its compute empire can support both internal training and outside rental demand without starving either side.

Large-Scale AI Engineering: Design, Train, and Optimize Foundation Models on NVIDIA GPU Clusters

Large-Scale AI Engineering: Design, Train, and Optimize Foundation Models on NVIDIA GPU Clusters

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

What did SpaceX buy?

SpaceX exercised an option to buy Anysphere, the maker of the AI coding agent Cursor, for $60 billion in an all-stock transaction.

When is the deal expected to close?

The transaction is expected to close in the third quarter of 2026, according to the source material.

Why is Cursor valuable to SpaceX?

Cursor gives SpaceX a fast-growing paid AI application, a developer user base and a model team that can connect directly to SpaceX-controlled compute and Grok-related products.

What is the weak point in SpaceX’s AI strategy?

The weak point identified in the source material is the model layer. Grok has access to major compute resources, but the analysis says it has not yet matched the lead of stronger frontier-model competitors.

What remains unknown?

It is not yet clear how regulators will treat the deal, how Cursor customers will respond, when the co-trained model will ship, or whether Grok can turn SpaceX’s infrastructure advantage into model leadership.

Source: Thorsten Meyer AI

You May Also Like

What a High-End Home Office Says About Modern Professional Identity

Just how a high-end home office reflects your professional identity reveals your commitment to excellence and innovation, but there’s more to discover.

I think Anthropic and OpenAI have found product-market fit

Recent pricing moves and enterprise investments suggest Anthropic and OpenAI have found strong product-market fit with their coding and general-purpose AI agents, signaling a shift toward revenue generation.

The last six months in LLMs in five minutes

A concise overview of the major advances in LLMs over the past six months, including model rankings, coding capabilities, and new projects.

When AI Builds Itself: Inside Anthropic’s Evidence on Recursive Self-Improvement

Anthropic says Claude now writes most merged code and is improving AI research workflows, while warning full recursive self-improvement is unproven.