Louisiana's $100 Billion Launch Complex Highlights SpaceX's Long-Term Expansion Ambitions Far Beyond Existing Forecasts

Overview

The report examines SpaceX's announcement of a new Starbase in Louisiana involving cumulative investment of $100 billion, analyzing the complex's scale, location selection, Starship launch capacity, orbital computing demand, funding sources, and equity valuation. Morgan Stanley believes the project's implied long-term launch scale could exceed its current forecasts and that SpaceX remains attractive at the current price. However, the decisive factors for the thesis are Starship reusability and turnaround speed, the engineering path from hundreds to thousands of annual launches, and whether growth and debt markets can continue supporting capital expenditures.

Core views

SpaceX announced that it will build a new Starbase on a former Exxon site in Vermilion Parish in southern Louisiana. The announced cumulative investment is $100 billion over 10 years; construction is scheduled to begin in 2027, with the first launch targeted for 2029. The site is planned to include 5 launch complexes with 2 launch pads each, for a total of up to 10, as well as propellant production facilities, a power plant, vehicle processing facilities, and housing for employees and their families. It is expected to create 3,000 direct jobs. The site will become SpaceX's fourth and potentially largest launch facility, intended to support thousands of launches annually and serve as a central node for expanding Starship missions. The scale of the complex indicates that SpaceX's long-term launch ambitions may be more aggressive than Morgan Stanley's current model. Falcon 9 currently operates from 3 launch pads and launches approximately once every two days. By the end of 2027, SpaceX expects to have at least 3 operational Starship launch pads, including 2 in Texas and at least 1 in Florida. The firm conservatively assumes each launch pad can support only 2 launches per day. On this basis, its forecast of approximately 5,800 Starship launches in 2040 would require only 8 launch pads and would correspond to $3.5 trillion in SpaceX revenue in the model, without relying on completion of the entire Louisiana complex. The report identifies a total of 15 planned launch pads: 10 in Louisiana, Pad 1 and Pad 2 at Starbase in Texas, LC-39A at Kennedy Space Center, and SLC-37A and SLC-37B at Cape Canaveral. The Louisiana complex also strengthens the long-term demand thesis for orbital computing. Morgan Stanley expects orbital computing to account for more than 80% of its launch forecast beginning in 2032. The traditional global connectivity market may already be approaching saturation before all 10 launch pads are fully utilized, so capacity on this scale is more likely to correspond to computing infrastructure deployment. Louisiana enables southbound flight over the Gulf of Mexico into dawn-dusk sun-synchronous polar orbit while minimizing overflight of densely populated areas. Such orbits are directly relevant to orbital computing, whereas achieving similar trajectories from Florida, Texas, or California is more challenging. Local resources, policy support, and risk diversification also explain the site selection. Louisiana ranks third in US natural gas production, behind only Texas and California, while each Starship launch requires more than 1,000 metric tons of liquid methane. Locating facilities across multiple states can diversify weather, operational, regulatory, and political risks and reduce the impact of resistance to data center and AI projects within any single jurisdiction. State support includes refunds of state and local sales taxes for major aerospace facilities exceeding $1 billion, expansion of industrial tax exemptions to aerospace manufacturing and infrastructure, early dismissal and legal-cost protection for certain lawsuits, and state-facilitated land transfers. The $100 billion project also raises financing questions. The announcement's wording indicates that the amount represents cumulative investment over 10 years, roughly equivalent to all capital expenditures for the Space business in Morgan Stanley's model from 2026 to 2035. However, given the scale of AI spending, it represents only 2% of total capital expenditures during the same period. The model assumes that funding comes partly from average annual net debt issuance of approximately $80 billion. The firm notes that this creates a mutually dependent loop between growth and financing: projects can be announced first, but only if business growth materializes will SpaceX and the debt markets have reason to continue supporting the corresponding capital investment. Starship is the technical foundation of the entire scaling thesis. Its maximum payload exceeds Falcon 9's by more than 5 times, and both its first and second stages are intended to be reusable, whereas Falcon 9 reuses only the first stage. NASA data show that the original Falcon 9 reduced the historical cost of approximately $18,500/kg by 85% to $2,700/kg in 2010. Morgan Stanley estimates that Falcon 9's internal variable launch cost declined further to approximately $1,000/kg in 2025. SpaceX's long-term goal for Starship is to reduce costs by more than 99% relative to the historical average, reaching below $200/kg. The firm believes reaching the several-hundred-dollar-per-kilogram range is more a question of "when" than "if," although cost improvements will ramp over many years. Starship has recovered the Super Heavy booster three times and reused it twice; the next explicit milestone is recovery of the spacecraft itself. SpaceX has demonstrated controlled ocean landings during test flights, and Morgan Stanley expects spacecraft recovery could occur by the end of 2026, potentially as early as Flight 14. However, the firm considers the turnaround time required for reuse more important than the first landing because turnaround frequency is a better indicator of near-term cost-reduction capacity. Falcon 9 took 15 months from the first booster recovery in December 2015 to the first reuse in March 2017. The maximum number of flights by a single booster gradually increased from 10 in May 2021 to 15 in December 2022, 20 in April 2024, 30 in August 2025, and 35 in June 2026, while the average number of reuses among operational boosters has generally been significantly lower. Reusing the spacecraft is more difficult than reusing the booster because it must withstand greater thermal and mechanical stress during reentry, and its silicon-based ceramic thermal protection system may require periodic replacement even if it is reusable. On the other hand, the spacecraft has only 6 Raptor engines, compared with 33 on Super Heavy, and the model estimates that the spacecraft accounts for approximately one-third of the entire vehicle's cost. Therefore, even if the spacecraft cannot be reused initially, Starship's enormous payload scale relative to Falcon 9 can still significantly reduce costs. SpaceX may also begin transporting internal Starlink payloads in a partially reusable configuration from late 2026 to 2027. The company has also completed the last scheduled Starlink mission from Florida using Falcon 9, and future local Starlink missions will be performed by Starship. Because a single Starship can carry up to approximately 25 times Falcon 9's downlink capacity, a small number of Starship launches could offset a decline in Falcon 9 launch frequency. Morgan Stanley's cost model projects that Starship's unit launch cost will decline to approximately $500/kg in 2030 at 341 launches, below $200/kg in 2035 at approximately 2,600 launches, and below $150/kg in 2040 at approximately 6,000 launches. Approximately 6,000 launches in 2040 imply about 16 launches per day, which could be supported by 4 launch complexes and 8 launch towers, assuming 2 towers per complex and up to 2 launches per tower per day. The model assumes that each spacecraft will be reused approximately 40 times and each booster approximately 130 times by then. Including a 20% hardware buffer, the active fleet would require approximately 160 spacecraft and 60 boosters, with the principal consumables being liquid methane, liquid oxygen, and some thermal protection tiles that may require replacement. The underlying cost model divides Starship into three versions: V3 operates from 2027 to 2029 with an effective mass to orbit of 80 metric tons; V4 operates from 2030 to 2034 with an effective mass to orbit of 120 metric tons; and V5 operates from 2035 through 2040 with an effective mass to orbit of 160 metric tons. The model divides costs into propellant, refurbishment, launch operations and mission assurance, and other variable cash costs, pricing liquid oxygen at $0.15/kg and liquid methane at $0.30/kg. The difference between total launch cost and cash cost is treated as implied depreciation per flight. Initial construction costs for V3, V4, and V5 are assumed to be $125 million, $155 million, and $175 million, respectively. The model is cautious regarding early reuse: each spacecraft has a lifespan of fewer than 2 flights from 2027 to 2029, rising to 3 flights in 2030, 4 in 2031, 17 in 2035, and 43 in 2040. For boosters, it assumes an 8-year ramp to more than 30 reuses by 2034, broadly consistent with Falcon 9's historical ramp. For orbital computing, the decisive constraint is not simply reaching cost parity with terrestrial computing but whether sufficient power can be deployed quickly enough. Each launch may deploy only a single-digit number of megawatts of computing capacity. To make the upfront investment in several gigawatts of orbital computing feasible, the company must show investors a path to 1,000 or even several thousand annual launches. The report notes that two recent neocloud transactions by SpaceX were priced at significant premiums to the industry average, potentially reflecting the scarcity of immediately available, large-scale, high-end GPU clusters. Therefore, even if orbital computing costs more than terrestrial computing, SpaceX could still charge a premium if it can provide the most scalable AI infrastructure, similar to its existing launch business, where prices have risen even as costs have continued to decline. At the equity level, the report believes investors can reassess SPCX near its IPO price because fundamental momentum in Space, Connectivity, and Enterprise AI is stronger, the outlook following second-quarter results is more optimistic, and share unlocks have not weighed on the stock price as much as the market feared. At a share price of approximately $137, the implied valuation multiple for the Enterprise AI business is only in the very low single digits, while the orbital AI option is effectively valued at zero. Each additional 1GW of nominal computing load could add $27 per share, based on $50/watt, a 70% incremental margin, and a 10x EBITDA multiple, equivalent to approximately 20% of the share price at the time. Morgan Stanley expects AI compute capacity to reach 4.9GW by the end of FY27, compared with the company's target of nearly 10GW. The firm states that SpaceX trades at approximately 10 times sales and 25 times EBIT based on FY28 forecasts, corresponding to 70% sales growth and 113% EBIT growth. Earnings per share are expected to improve from -$1.69 in FY25 to $0.54 in FY26, $2.45 in FY27, and $6.36 in FY28. The $300 price target uses a sum-of-the-parts valuation: Space contributes $8, Connectivity $118, X & Grok $8, and Enterprise AI $165. The forecast period extends through 2040, with a valuation date of June 30, 2027, an 11.1% WACC, and an 11.9% cost of equity. The Enterprise AI valuation is discounted by 50% for execution risk. The perpetual growth rates for the respective segments are 4.0% for Space, 4.5% for Connectivity, 3.0% for X & Grok, and 5.0% for Enterprise AI.

Analysis framework

The report first assesses SpaceX's long-term capacity planning based on the scale, timeline, and number of launch pads at the new Louisiana complex, then incorporates geographic trajectories, natural gas supply, local incentives, and interstate risk diversification into the site-selection analysis. It subsequently tests whether the 2040 launch volume is achievable using daily launch capacity per pad and constructs a bottom-up unit launch cost model based on Starship versions, payload, construction costs, propellant, depreciation, and annual reuse counts. Finally, it links launch capacity to the speed of orbital computing deployment and derives the $300 price target through computing-load sensitivity analysis and sum-of-the-parts valuations for Space, Connectivity, X & Grok, and Enterprise AI.

Methodology notes

• Industry/Sector Analysis FrameworkSupply-demand framework

  Matching Launch Capacity with Orbital Computing Demand

  The report measures supply using the number of launch pads, daily launch capacity per pad, and annual launch frequency, and measures demand using the payload required for connectivity services and orbital computing, in order to assess the scale and future utilization of the new complex.

• (Out-of-Vocabulary Method)

  Bottom-Up Starship Launch Cost Model

  The model separately estimates the payload, construction costs, propellant, refurbishment, operations, mission assurance, depreciation, and spacecraft and booster reuse counts for each Starship generation, thereby deriving per-launch and per-kilogram-to-orbit costs.

• Valuation MethodSOTP Valuation

  SpaceX Sum-of-the-Parts Valuation

  The report separately values Space, Connectivity, X & Grok, and Enterprise AI, then sums the per-share values of each segment to derive the $300 price target.

• Valuation MethodDCF

  Long-Term Cash Flow and Perpetual Growth Valuation

  The segment valuations use forecasts through 2040, an 11.1% WACC, an 11.9% cost of equity, and different perpetual growth rates for each business, with a 50% valuation discount applied to Enterprise AI for execution risk.

• (Out-of-Vocabulary Method)

  Incremental AI Compute Capacity Sensitivity Analysis

  The report calculates the change in per-share value for each additional 1GW of computing load based on $50/watt, a 70% incremental margin, and a 10x EBITDA multiple, illustrating the valuation impact of expanding AI compute capacity.

Asset mapping & comparison

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

• SpaceX (SPCX.US)

  The Louisiana complex expands Starship's potential launch capacity and provides long-term infrastructure for orbital computing, Starlink, and other Space businesses.

  Strengths

  Experience with reusable rockets, Starship's large payload capacity, a geographically diversified launch network, the foundation of the Connectivity business, and potential expansion opportunities in orbital and enterprise AI.

  Weaknesses

  Starship spacecraft reuse and rapid turnaround still require engineering iteration, annual launch capacity in the thousands has not yet been validated, and the long-term plan depends on substantial capital expenditures and financing.

  Comparison

  Starship's maximum payload exceeds Falcon 9's by more than 5 times, and it is designed for full reuse of both stages; a single Starship can carry up to approximately 25 times Falcon 9's downlink capacity.

  Risks

  Reuse progress, Starlink user growth, Enterprise AI monetization, capital expenditures, financing dilution, regulatory delays, and compute capacity commissioning timelines may all deviate from the model.

Key data

• Cumulative Investment in Louisiana Starbase$100bnInvested over 10 years according to the announcement's wording
• Construction and First-Launch TimelineConstruction begins in 2027; first launch in 2029Planned timeline for the new complex
• Complex Launch Facilities5 launch complexes and 10 launch pads in totalEach complex includes 2 launch pads
• Direct New Employment3,000Direct jobs expected to be created by the project
• SpaceX Planned Launch Pads1510 in Louisiana, 2 in Texas, and 3 in Florida
• 2040 Starship Launch ForecastApproximately 5,800Only 8 launch pads are required at 2 launches per pad per day
• 2040 SpaceX Revenue Forecast$3.5trMorgan Stanley model forecast
• Share of Launches for Orbital Computing>80%Share of forecast launches beginning in 2032
• Liquid Methane Required per Starship Launch>1,000 metric tonsLouisiana's natural gas resources provide a site-selection advantage
• Average Annual Net Debt IssuanceApproximately $80bnA partial source of capital expenditure funding in the model
• Starship Unit Launch Cost ForecastApproximately $500/kg in 2030; below $200/kg in 2035; below $150/kg in 2040Corresponding to 341, approximately 2,600, and approximately 6,000 annual launches, respectively
• 2040 Active Fleet RequirementApproximately 160 spacecraft and 60 boostersIncluding a 20% hardware buffer
• Effective Mass to Orbit by Starship VersionV3 80 metric tons; V4 120 metric tons; V5 160 metric tonsCorresponding to 2027—2029, 2030—2034, and 2035—2040
• Initial Construction Cost by Starship Generation$125mn/$155mn/$175mnCorresponding to V3, V4, and V5, respectively
• AI Compute Capacity Estimate and Company Target4.9GW by the end of FY27, versus the company's target of nearly 10GWComparison between Morgan Stanley's estimate and the company's target
• Valuation Sensitivity per Additional 1GW of Compute Capacity+$27 per shareBased on $50/watt, a 70% incremental margin, and a 10x EBITDA multiple, equivalent to approximately 20% of the share price at the time
• FY28 Trading Multiples and Growth10x sales, 25x EBITCorresponding to 70% sales growth and 113% EBIT growth
• Price Target$300Space $8, Connectivity $118, X & Grok $8, Enterprise AI $165

Impact & implications

The report believes the new complex is not necessary to achieve its existing 2040 forecast but instead signals that SpaceX may be pursuing higher launch frequency, particularly for large-scale orbital computing deployment. If Starship can increase reuse and turnaround frequency, it will support expansion across the Space, Connectivity, and AI businesses by reducing unit costs to orbit and shortening compute deployment timelines. However, the $100 billion plan also makes growth execution, debt financing, and capital expenditure discipline critical constraints within the same logical chain. From an equity valuation perspective, the firm believes the current price assigns limited value to Enterprise AI and the orbital AI option and therefore maintains a positive stance.

Risks

• Slower-than-expected Starship reuse and turnaround could delay reductions in unit launch costs and increases in launch frequency.
• Slower Starlink user growth could weaken capacity utilization and valuation for the Connectivity business.
• Weaker-than-expected Enterprise AI monetization could prevent the modeled compute-capacity expansion and segment value from materializing.
• AI infrastructure capital expenditures or computing costs per watt may exceed expectations.
• A longer period from investment to operational compute capacity could limit the pace of orbital computing expansion.
• Greater financing requirements could result in additional debt or equity dilution.
• Delays in regulatory approvals could affect complex construction and launch plans.

What to watch

• Monitor whether the Starship spacecraft can be recovered by the end of 2026 and the actual refurbishment and turnaround time following the first recovery.
• Monitor whether Starship can establish a clear path from hundreds of annual launches to 1,000 or even several thousand launches per year.
• Monitor whether construction of the Louisiana complex begins as planned in 2027 and whether it achieves its first launch in 2029.
• Monitor whether Starship reuse progresses faster than assumed in the model and whether Starlink capacity can grow more rapidly.
• Monitor DTC and enterprise customer adoption, neocloud orders, and Cursor ARR growth.
• Monitor whether AI compute capacity can move closer to the company's target of nearly 10GW by the end of FY27.
• Monitor whether the per-watt cost and commissioning timeline of AI infrastructure decline.
• Monitor whether growth and debt markets can continue supporting the $100 billion complex and the broader capital expenditure plan.