SpaceX Owns Every Layer of AI Now. The Model Is Still the Weak Link.

📊 Full opportunity report: SpaceX Owns Every Layer of AI Now. The Model Is Still the Weak Link. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

TL;DR

SpaceX has completed its $60 billion acquisition of Cursor, owning every layer of the AI stack from hardware to applications. Despite this, the company’s AI models remain a weak link, with low utilization and ongoing development challenges.

SpaceX has finalized the acquisition of Cursor for $60 billion in all-stock, gaining control over every layer of the AI stack, from hardware to applications. This move makes SpaceX the only company with such comprehensive vertical integration in AI, but the company’s models still face significant performance and utilization challenges, raising questions about the effectiveness of their AI strategy.

On June 16, SpaceX announced it exercised its option to buy Cursor, a profitable AI coding startup, for $60 billion, after earlier plans to acquire it. The deal consolidates SpaceX’s position as a fully integrated AI powerhouse, controlling hardware, data centers, research labs, models, and applications. Cursor, founded in 2022 by MIT graduates, had achieved approximately $4 billion in annual revenue primarily from its AI coding product, making it a rare profitable player in AI development.

With the acquisition, SpaceX now owns the entire AI stack, including the supercomputers in Memphis, the Grok model line, and its own application, Cursor. The company has built the Colossus supercomputers, which operate nearly 555,000 Nvidia GPUs, representing a multi-billion-dollar infrastructure capable of training large models. SpaceX has also secured long-term compute agreements with rivals like Anthropic and Google, leasing excess capacity at its supercomputers, which indicates a strategic move to monetize idle compute resources.

Despite owning every layer, the models themselves remain a weak link. Internal reports reveal that the utilization rate of SpaceX’s supercomputers for model training is only about 11%, far below the 35–45% considered production-ready. The low utilization stems from inefficiencies in parallelizing training across the mixed GPU architecture, prompting SpaceX to lease out its supercomputers to other AI labs, including rivals, to maximize revenue. Elon Musk has publicly expressed confidence in leasing the compute, citing the possibility of reclaiming resources if certain AI applications pose risks.

At a glance
breakingWhen: announced June 16, 2026, expected to cl…
The developmentSpaceX announced the acquisition of Cursor for $60 billion, completing its control over the entire AI infrastructure, from compute to application, but the AI models still face performance issues.
SpaceX owns every layer of AI — the stack, the rentals, the weak link
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

Why Full AI Stack Ownership Doesn’t Guarantee Success

Owning every layer of the AI stack positions SpaceX as a unique, vertically integrated AI conglomerate, giving it control over hardware, data, research, and applications. However, the persistent issues with model efficiency and low utilization highlight that infrastructure alone does not ensure AI performance or leadership. This development underscores the ongoing challenge of developing effective, scalable AI models, even when the underlying compute and data are under tight control.

Amazon

Nvidia GPU supercomputer for AI training

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

The Evolution of SpaceX’s AI Ambitions and Infrastructure

SpaceX’s move into AI has been strategic, with the company building the Colossus supercomputers in Memphis, capable of training massive models at unprecedented speeds. The company’s integration of xAI, its research lab, and the Grok model line has aimed to establish a comprehensive AI ecosystem. Prior to the Cursor deal, SpaceX had already secured significant compute agreements, leasing capacity to rivals like Anthropic and Google, which indicates a shift towards monetizing idle infrastructure while developing its own models.

The acquisition of Cursor marks a milestone in SpaceX’s ambition to control the entire AI supply chain, from silicon to application, but the models’ limited efficiency remains a challenge. The low utilization rates and the need to lease out supercomputing resources reflect ongoing hurdles in AI model development, despite the company’s technological and financial resources.

“Leasing our excess compute is a strategic move; we can reclaim resources if AI ever poses a threat to humanity.”

— Elon Musk, SpaceX CEO

AI Systems Performance Engineering: Optimizing Model Training and Inference Workloads with GPUs, CUDA, and PyTorch

AI Systems Performance Engineering: Optimizing Model Training and Inference Workloads with GPUs, CUDA, and PyTorch

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Unresolved Challenges in AI Model Performance and Utilization

It is still unclear how effectively SpaceX’s models will improve in performance and efficiency over time. The low utilization rates suggest persistent technical hurdles, and the long-term impact of owning every layer of AI remains uncertain. Additionally, the strategic implications of leasing capacity to rivals and the potential risks associated with AI development are still developing areas of concern.

Amazon

high performance AI computing servers

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Next Steps in SpaceX’s AI Strategy and Model Development

SpaceX is expected to continue investing in improving its AI models, aiming to increase training efficiency and utilization. The company’s upcoming model releases and potential infrastructure upgrades will be closely watched, alongside regulatory developments and the evolving AI landscape. The integration of Cursor into SpaceX’s broader ecosystem will likely influence the company’s AI ambitions and competitive positioning.

Amazon

enterprise AI supercomputers

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

Key Questions

What does SpaceX’s acquisition of Cursor mean for the AI industry?

It signifies a move toward fully integrated AI infrastructure, combining hardware, research, and applications under one company, which could reshape competitive dynamics in AI development.

Despite owning the infrastructure, the models have low utilization rates (~11%) and technical inefficiencies, limiting their effectiveness and scalability.

Will owning all layers of AI guarantee SpaceX’s success?

Not necessarily; infrastructure control does not automatically translate into superior AI performance, especially if models remain inefficient or unscalable.

How does leasing compute to rivals impact SpaceX’s strategy?

Leasing excess capacity generates revenue and maximizes infrastructure utilization but may also complicate control over AI development and pose strategic risks.

What are the next major developments to watch for?

Upcoming model releases, infrastructure upgrades, and regulatory decisions will be key indicators of how SpaceX’s AI ambitions unfold in the coming months.

Source: ThorstenMeyerAI.com

This content is for general information only and is not financial, tax or legal advice. Consult a qualified professional for decisions about your money.
You May Also Like

Forward-Deployed Engineer Economics 2.0: The Unit Economics Math, Six Months Later

An update on FDE economics reveals profitability at scale, driven by actual contract sizes and costs, with implications for enterprise AI deployment and lab scaling.

Candor as a Moat: A Critical Reading of Dario Amodei and Anthropic

Examining how Dario Amodei’s transparency and policy proposals serve as a strategic barrier for Anthropic amid AI regulation debates.

Open-source sponsor update generator

A new tool for open-source maintainers to automate sponsor updates is entering initial testing, aiming to improve communication and support sustainability.

Understanding Anthropic’s $965B Series H: The Compute Revolution

Anthropic’s latest funding round emphasizes hardware infrastructure—chips, memory, power—to support AI scaling, marking a shift from valuation to physical capacity.