SpaceX Plans to Vertically Integrate Gas-Turbine Blade Supply to Address the Power-Delivery Bottleneck for AI Compute Expansion

Overview

The report focuses on the strategic significance of SpaceX establishing blade and vane casting capabilities for its ground-based AI data centers and Starship Raptor engines. Morgan Stanley believes that the key constraint on large-scale AI compute expansion is not simply megawatts, but rather gas-turbine supplier production schedules, configurations, and delivery lead times; building in-house capacity for critical hot-section castings could strengthen SpaceX's control over power-deployment timing, product synergies, and costs.

Core views

SpaceX's vertical-integration rationale is to bring the nickel-based superalloy castings needed by AI data-center gas turbines and Raptor turbopumps onto a common manufacturing platform, spreading fixed costs while relieving external supply constraints. The report notes that the few specialized foundries in the industry have operated at full capacity for extended periods and carry order backlogs, with gas-turbine delivery windows generally extending to 2029 or 2030. If the Bastrop facility successfully reaches volume production, SpaceX could accelerate its conversion of energy into intelligence; however, construction, certification, equipment lead times, yields, and workforce development create substantial near-term execution risk.

Analysis framework

The report assesses the AI power bottleneck and its implications for company value using SpaceX job postings, publicly disclosed equipment-purchase commitments, market order and capacity data, management interviews, and a sum-of-the-parts valuation framework.

Methodology notes

• Valuation methodsSum-of-the-Parts Valuation

  Value Space, Connectivity Services, X & Grok, and Enterprise AI separately and then aggregate them

  The $300 target price comprises $8 for Space, $118 for Connectivity Services, $8 for X & Grok, and $165 for Enterprise AI; the valuation date is 2027-06-30, with forecasts extending through 2040.

• Valuation methodsDiscounted Cash Flow

  Value each business segment using cost of capital and terminal growth rates

  The report uses an 11.1% WACC and an 11.9% cost of equity; terminal growth rates are 4.0% for Space, 4.5% for Connectivity Services, 3.0% for X & Grok, and 5.0% for Enterprise AI.

• Strategic AnalysisVertical Integration

  Shorten delivery times and strengthen supply-chain control by internalizing scarce component supply

  The report views blades and vanes as a shared critical bottleneck for gas turbines and Raptor turbopumps, and believes an in-house foundry could provide optionality in production scheduling, design, and cost.

• Operational AnalysisUnit Economics Framework

  Intelligence per watt per dollar per second

  The report measures competitiveness by the intelligence output delivered per unit of energy, capital, and time, emphasizing the importance of time to power and deployment speed for AI value creation.

Asset mapping & comparison

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

• SPACE EXPLORATION TECHNOLOGIES CORP（SPCX.US）

  Core Covered Company

  Strengths

  Space, connectivity services, and AI businesses can share manufacturing, capital, and engineering capabilities; vertical integration could improve the availability and deployment flexibility of critical power equipment.

  Weaknesses

  New single-crystal casting operations are technically complex, with long production-ramp and certification cycles, making it difficult to immediately relieve supply constraints in the near term.

  Comparison

  After excluding the value of Space and connectivity services, the report believes the market's implied valuation for consumer and enterprise AI businesses is approximately 1x 2028 EV/Sales, below new-cloud-service peers.

  Risks

  Uncertain power interconnection, delayed gas-turbine deliveries, casting yields below expectations, weaker-than-expected AI commercialization, and increased financing needs.

• Howmet Aerospace（HWM）

  Comparable Company in the Blade and Vane Supply Chain

  Strengths

  Positioned in the highly specialized hot-section casting supply segment, benefiting from demand growth and tight supply.

  Weaknesses

  Customer certification requirements, capacity expansion, and supply-chain constraints may limit near-term incremental output.

  Comparison

  Together with a small number of other independent foundries, it forms a concentrated market; SpaceX's in-house capacity aims to reduce reliance on this type of external supply.

  Risks

  New industry capacity, end-demand volatility, and customers internalizing manufacturing could alter the supply-demand balance.

Key data

• Investment RatingOverweightMorgan Stanley maintains a positive view on SpaceX.
• Target Price$300.00Implies approximately 112% upside relative to the $141.29 closing price.
• Current Share Price$141.29Closing price on 2026-08-13.
• Available Power by End-2027Approximately 3–4GWDerived from reported but not fully confirmed turbine-supply arrangements.
• End-2027 AI Power in Model5GWBelow Musk's 10GW target mentioned on the earnings call, reflecting uncertainty around power procurement and deployment timing.
• Compute Business Revenue Opportunity per GWMore than $50bn/yearThe report's estimate of the scaling opportunity for SpaceXAI.
• AI Infrastructure Turbine CommitmentsApproximately $2.8bnIncludes $805mn in purchase agreements through 2029 and an approximately $2.0bn acquisition of mobile gas turbines.
• Time to Initial Volume Production for Single-Crystal CastingPotentially more than 2–3 yearsConstrained by process complexity, equipment lead times, yields, certification, and skilled labor.
• Global Data-Center Power DemandApproximately 55GW rising to 210GWThe report expects roughly fourfold growth from 2023 to 2028.
• Gas-Turbine Orders100GW in 2025; 117GW annualized based on Q1 2026The report believes AI-driven power demand is extending industry order and delivery cycles.

Impact & implications

For SpaceX, if blade-casting capability scales as planned, it will represent more than component self-sufficiency: it will be a critical determinant of AI infrastructure go-live timing, capital efficiency, and competitive barriers. For the supply chain, AI data-center demand will continue to concentrate in gas-turbine hot-section castings, benefiting specialized foundries and equipment manufacturers with in-house casting capabilities; however, new entrants will find it difficult to replicate mature capacity in the short term.

Risks

• The Bastrop foundry could be delayed by equipment lead times, process yields, talent recruitment, and certification requirements, with initial qualified volume production potentially requiring more than 2–3 years.
• SpaceX's path to obtaining off-grid or dispatchable power is unclear, potentially causing AI data-center power scale to fall below management targets.
• Order backlogs in the gas-turbine and hot-section-casting industries could persist until around 2030, constraining the pace of capacity deployment.
• Slower Starship reusability progress or weaker-than-expected Starlink capacity and subscriber growth could depress valuations for the Space and connectivity businesses.
• Weaker-than-expected enterprise AI monetization, X & Grok business performance, or new-cloud customer expansion could erode the value of the AI segment.
• Capital expenditures or cost per watt of compute exceeding expectations, increased financing needs, and potential dilution could affect shareholder returns.
• Regulatory delays and investment-banking conflicts of interest may affect the independence of the report's views; investors should not use it as a sole decision factor.

What to watch

• Construction progress, equipment installation, hiring, and the timeline for first qualified parts at the Bastrop blade and vane foundry.
• SpaceX's additional gas-turbine purchases, mobile-power asset arrangements, and the scale of verifiable power-supply contracts.
• Whether actual available AI data-center power by the end of 2027 approaches Morgan Stanley's 5GW model estimate or Musk's 10GW target.
• Volume-production yields and certification for single-crystal/directionally solidified superalloy castings, and whether internal supply partnerships are established with existing turbine manufacturers.
• Validation of segment valuation assumptions through Starship reusability, Starlink capacity expansion, and enterprise AI customer growth.
• Whether time to power for AI infrastructure, unit capital investment, and cost per watt of compute continue to improve.