Cursor Could Become the Key Pivot for a Valuation Re-rating of SpaceX's AI Platform

Overview

This report conducts an in-depth study of SpaceX's proposed acquisition of the AI coding platform Cursor. Morgan Stanley believes the market currently mainly views SpaceX's artificial intelligence business as a compute-rental-style new cloud service and has not yet fully reflected Cursor's value in developer workflows, enterprise distribution, proprietary models, interaction data, and application-layer network effects. The combination of Cursor with Grok, X real-time data, SpaceXAI compute, Starlink connectivity, and potential collaboration with Tesla physical AI could form an end-to-end enterprise platform for builders in software and physical industries.

Core views

First, Cursor's value comes not only from underlying models, but more importantly from its model-agnostic orchestration layer that controls developer workflows, context, retrieval, tool calling, and deployment entry points. Second, enterprise AI spending and code-generation penetration remain in the early stages of rapid growth, and Cursor has strong expansion potential given its broad enterprise coverage and usage-based billing. Third, shifting more tasks from expensive third-party frontier models to Composer and Grok is expected to significantly improve unit inference costs and gross margin. Fourth, after deducting the value of the Space and Connectivity businesses from the current share price, the implied value of the consumer and enterprise AI businesses is only approximately $12 per share, slightly above 1x 2028 sales and materially below new cloud service peers. Fifth, business disclosure after transaction completion and subsequent model releases could serve as catalysts for narrowing the valuation discount of the AI business.

Analysis framework

The report uses bottom-up operating forecasts, modeling Cursor's subscription users, revenue per user, usage, token consumption per user, inference costs, ARR, revenue, and gross margin on a quarterly basis through 2027 and annually extrapolating to 2030. It also measures the implied valuation of SpaceX's AI business through a sum-of-the-parts value breakdown, and compares the cost-capability efficiency of Cursor, Grok, and third-party frontier models using coding intelligence indices, cost per task, and token consumption.

Methodology notes

• Operating forecastBottom-up ARR and gross margin model

  Derive revenue and margin from paid seats, enterprise penetration, revenue per user, usage, and inference costs.

  The model forecasts quarterly through 2027 and annually through 2030; enterprise customers are expected to contribute more than approximately three-quarters of ARR, while the share of usage-based billing gradually increases.

• Valuation analysisSum-of-the-parts and implied valuation

  Deduct the value of the Space and Connectivity businesses from the share price to estimate the implied value the market assigns to the AI business.

  Based on Morgan Stanley's valuation, after deducting approximately $127 per share of value for the Space and Connectivity businesses, the current implied value of the AI business is approximately $12 per share.

• Competitive analysisCost-intelligence Pareto frontier

  Compare coding capability, cost per task, and token efficiency simultaneously, rather than judging competitiveness solely by model benchmark scores.

  Grok 4.5 achieves coding capability close to top-tier models at a cost per task of approximately $2.59; the Composer model has lower capability, but its cost per task is only approximately $0.04 to $0.08, making it suitable for high-frequency routine tasks.

• Data limitationsTriangulation of public information

  Use company disclosures, media reports, and third-party industry data to cross-estimate Cursor operating metrics.

  Cursor remains a private company with limited official financial data; the relevant forecasts have not been reviewed by Cursor management and do not include non-public financial information, so there is relatively high estimation uncertainty.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• SpaceX (SPCX.US)

  Core research subject and proposed acquirer of Cursor

  Strengths

  Has Space, Starlink connectivity, AI compute, Grok models, and potential physical AI synergies, which can provide Cursor with compute, data, distribution, and capital support.

  Weaknesses

  Business portfolio is complex, capital expenditure is high, AI platform synergies still require execution validation, and current operating forecasts include many long-term assumptions.

  Comparison

  The market currently evaluates its AI business more through a new cloud service framework, while the report believes it could develop into an end-to-end platform with application entry points and proprietary data.

  Risks

  Acquisition integration, model competition, returns on compute investment, valuation volatility, product synergies below expectations, and uncertainty around transaction completion timing.

• Cursor

  Enterprise AI coding platform proposed to be acquired by SpaceX

  Strengths

  Has developer workflow entry points, broad enterprise coverage, a model-agnostic orchestration layer, proprietary Composer models, and rapidly growing ARR.

  Weaknesses

  Historical gross margin has been weighed down by third-party model inference costs, official financial data is limited, and individual developer accounts may still be loss-making.

  Comparison

  Compared with pure model companies, Cursor is closer to an AI-native development environment and SaaS platform; compared with new cloud services, its subscription revenue has higher revenue intensity per unit of compute.

  Risks

  Commoditization of low-cost coding models, gaps in frontier model capability, bifurcated developer adoption of model strategies, price competition, and customer migration.

Key data

• Cursor reported ARRapproximately $4 billionAs of early June 2026, up significantly from approximately $500 million a year earlier.
• Cursor ARR forecastapproximately $8 billion by end-2026; approximately $17 billion in 2027; approximately $33 billion in 2030Mainly driven by increased enterprise penetration and growth in usage-based billing.
• Proposed acquisition consideration$60 billionSpaceX plans to acquire Cursor with Class A common stock, with completion expected in the third quarter of 2026.
• Enterprise coveragemore than 50,000 enterprises and more than 64% of the Fortune 500Customers include Samsung, Adobe, Figma, Stripe, AMD, PwC, Shopify, Nvidia, and others.
• Enterprise ARR shareapproximately 75% to 80%Enterprise customers are expected to contribute most of Cursor's ARR over the long term.
• Cursor revenue contribution forecastapproximately $2.5 billion in 2026; approximately $13 billion in 2027Representing approximately 10% and 19% of SpaceX AI revenue forecasts, respectively, and approximately 15% in 2030.
• Compute utilizationapproximately 7% to 10%Refers to Cursor's share of total SpaceXAI nominal compute during the forecast period; its revenue intensity per unit of power is close to twice the average workload.
• Gross margin pathturns positive in the third quarter of 2026, reaching the low 60% range by 2030Improvement depends on more requests shifting to Composer, Grok, and lower-cost third-party models.
• Implied value of the AI businessapproximately $12 per shareDerived after deducting approximately $127 per share of value for the Space and Connectivity businesses, slightly above 1x 2028 sales.
• Target price and bull-case scenario$300 / $600$300 is the base target price; the $600 scenario reflects higher value from Cursor's evolution from an orchestration layer toward frontier models and a platform.

Impact & implications

If Cursor continues to expand enterprise customers after the transaction closes and leverages SpaceXAI compute and Grok models to reduce inference costs, SpaceX's revenue structure could shift further from capital-intensive compute rental toward high-margin subscription and application-layer revenue. Because the current valuation reflects this path only to a limited extent, transaction completion, operating data disclosure, and model performance improvement could bring a significant valuation re-rating. Conversely, if model capability is insufficient, cost advantages are replicated by competitors, or enterprise usage growth slows, Cursor may struggle to meet the report's high growth and gross margin assumptions.

Risks

• The SpaceX-Cursor transaction may not close as expected in the third quarter of 2026, or integration progress may fall short of expectations.
• Cursor still relies on third-party frontier models, and inference costs may continue to pressure gross margin.
• Composer and Grok may sit in the middle ground between high-end capability and lowest-cost models, making it difficult to form stable differentiation.
• Competition from open-weight, subsidized, and low-cost coding models could trigger price wars and compress margins.
• Enterprise AI budgets, paid seats, or usage growth may fall below the report's assumptions.
• Cursor is a private company, and ARR, revenue, and cost forecasts are mainly estimated based on public information and have not been reviewed by management.
• SpaceX's high capital expenditure and compute expansion may weaken free cash flow and investment returns.
• The share price has recently risen by more than 25%, and failure of near-term catalysts to materialize could lead to significant volatility.

What to watch

• The formal completion timing, transaction terms, and integration arrangements for the Cursor acquisition.
• Cursor revenue, ARR, customer mix, inference costs, and gross margin first disclosed in third-quarter 2026 results.
• Coding performance and cost profile of Grok 4.6 and Grok 4.7 after incorporating Cursor interaction data.
• Release progress of Grok 5 before the end of 2026 and its adoption within Cursor.
• Performance, pricing, and inference efficiency of Composer 3 or subsequent Composer models.
• Cursor enterprise customer count, paid seats, share of usage-based billing, and ARR growth rate.
• Traffic routing ratio between proprietary models and third-party models, and whether gross margin can turn positive in the third quarter of 2026.
• Cursor's share of SpaceXAI compute and revenue intensity per unit of compute.
• Whether the market's implied valuation discount for SpaceX's AI business narrows.