Second-quarter results materially beat expectations, AI compute expansion supports growth, reiterate $300 price target

Overview

The report believes SpaceX delivered strong second-quarter results, with revenue, adjusted EBITDA, and margins across business segments all better than expected. The timing of revenue recognition for the AI neocloud business, Starlink government and enterprise businesses, and the Space business together drove approximately $1.4 billion of EBIT improvement. The company's targets of more than $100 billion in annualized revenue and rapid AI compute expansion reinforce medium-term growth expectations, but compute buildout also significantly increases funding needs. Morgan Stanley therefore raised revenue and EBIT forecasts while also increasing capex, ultimately maintaining its $300 price target.

Core views

In the near term, the second-quarter beat is expected to drive moderate upward revisions to consensus expectations over the next 12 to 18 months. In the medium term, AI infrastructure, global connectivity, low-cost launch, and vertical integration are expected to form a self-reinforcing infrastructure flywheel and improve unit compute cost and deployment speed. The report takes a conservative approach to the company's 2027 compute target of close to 10 GW, forecasting only about 5 GW by year-end, while assuming revenue per watt declines as compute supply increases. Long-term valuation is highly dependent on Starship entering service on schedule, Starlink capacity and user expansion, and enterprise AI commercialization, with execution risk and capital intensity both significantly higher than for most large AI infrastructure companies.

Analysis framework

The report first compares second-quarter actual results against consensus expectations and Morgan Stanley forecasts, then separately re-estimates revenue, profit, compute capacity, and capex for the Space, Connectivity, and AI businesses. Valuation uses a sum-of-the-parts approach combined with long-term cash flow assumptions, applies an additional execution risk discount to Enterprise AI, and tests the impact on valuation of changes in Starship timing, compute deployment, revenue per watt, and Starlink growth through bull, base, and bear cases.

Methodology notes

• Performance analysisActual results versus consensus expectations variance analysis

  Compares quarterly actual revenue, margins, EBIT, and adjusted EBITDA item by item against consensus expectations and analyst forecasts.

  This method is used to identify sources of outperformance. The report believes the main contributions came from the timing of revenue recognition in the AI neocloud business, Starlink government and enterprise businesses, the Space business, and margin improvement across segments.

• Forecasting modelMorgan Stanley ModelWare

  Updates revenue, profit, compute capacity, user count, and capex forecasts based on business segments and key operating metrics.

  The report raised 2026 to 2027 AI revenue and overall revenue, and increased capex and depreciation-related assumptions in line with a faster compute buildout pace.

• Valuation methodSum-of-the-parts valuation

  Values Space, Connectivity, X & Grok, and Enterprise AI separately, then aggregates them to derive the company's equity value.

  The $300 price target consists of approximately $8 from Space, approximately $118 from Connectivity, approximately $8 from X & Grok, and approximately $165 from Enterprise AI; minor differences between the segment total and the price target are due to rounding.

• Cash flow valuationRisk-adjusted long-term cash flow valuation

  Assesses the forward value of each business using long-term growth rates, cost of capital, and execution risk discounts.

  The forecast period runs through 2040, the valuation date is June 30, 2027, and the analysis uses an 11.1% weighted average cost of capital and an 11.9% cost of equity; Enterprise AI is assigned a 50% valuation discount due to execution risk.

• Scenario analysisBull, base, and bear cases

  Assesses the valuation range by varying Starship operating timing, compute deployment growth, AI commercialization, and Starlink penetration.

  The bull, base, and bear cases correspond to $600, $300, and $75, respectively; the bear case assumes Starship is delayed until 2029 before entering service and that orbital compute struggles to reach scale.

Asset mapping & comparison

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

• SpaceX (SPCX.O) common stock

  The report directly covers the security and maintains an Overweight rating and $300 price target.

  Strengths

  Has advantages in leading launch costs, Starlink global connectivity capability, AI infrastructure deployment, and vertical integration, with the potential to create a cross-business infrastructure flywheel; second-quarter revenue and profit performance significantly exceeded expectations.

  Weaknesses

  The AI and Space businesses remain highly capital-intensive, and current earnings and cash flow are sensitive to buildout pace, revenue recognition, and scaling progress.

  Comparison

  Compared with traditional AI infrastructure companies, SpaceX may have lower cost per watt and a faster path to power deployment, but its technology, regulatory, and financing execution risks are also materially higher.

  Risks

  Delays in Starship, slower Starlink user growth, enterprise AI commercialization falling short of expectations, rising unit compute costs, financing dilution, and regulatory delays could all pressure valuation.

Key data

• Second-quarter total revenue$7.814 billionThe report says revenue was approximately 16% above expectations; the table shows it was $1.067 billion above Morgan Stanley's forecast.
• Second-quarter adjusted EBITDA$3.538 billionApproximately 75% above Morgan Stanley's forecast, with an adjusted EBITDA margin of 45.3%.
• Second-quarter EBIT variance$1.427 billion above consensus expectationsActual EBIT was a loss of $143 million, versus consensus expectations for a loss of $1.570 billion.
• Annualized revenue target by end-2026More than $100 billionA company-wide target proposed by the company; the report notes that it is not formal financial guidance.
• Nominal compute target by end-2027Closer to 10 GW rather than 5 GWMorgan Stanley forecasts only about 5 GW based on execution risk.
• 2026 to 2027 revenue forecasts$48 billion, $102 billionPrior forecasts were $45 billion and $91 billion, respectively, with the main upward revision coming from the AI business.
• 2026 to 2027 EBIT forecasts$7.6 billion, $34.0 billionPrior forecasts were $5.4 billion and $29.0 billion, respectively; corresponding margins are 15.9% and 33.2%.
• 2026 to 2027 capex forecasts$64 billion, $163 billionPrior forecasts were $48 billion and $129 billion, with AI capex expected to be $53 billion and $130 billion, respectively.
• Starlink consumer user forecasts16.9 million in 2026, 28.8 million in 2027Broadly consistent with prior forecasts, with only a slight upward revision to Connectivity revenue.
• Target price and current price$300, $125.33Based on the closing price on August 4, 2026, potential upside is approximately 139.4%.
• Scenario valuationsBull $600, base $300, bear $75Valuation differences are primarily driven by Starship timing, compute expansion, and AI commercialization performance.

Impact & implications

The earnings beat and more ambitious compute and revenue targets are expected to drive moderate upward revisions to 2026 to 2027 consensus expectations and reinforce the long-term investment thesis for SpaceX as an integrated AI, connectivity, and space infrastructure platform. However, faster compute buildout puts pressure on free cash flow and financing needs, making valuation more sensitive to cost of capital, revenue per watt, and execution progress. The report also views the approximately $100 billion SPCX share unlock on August 6, 2026 as a potential opportunity to build positions, though near-term price volatility from increased supply should also be monitored.

Risks

• Starship reuse progresses more slowly than expected, which could delay connectivity capacity expansion and orbital compute deployment.
• Starlink user growth, V3 satellite deployment, or adoption by enterprise and direct-to-device terminals falls short of expectations.
• Enterprise AI commercialization, neocloud customer acquisition, and revenue per watt are weaker than forecast.
• Capex or cost per watt of buildout rises further, leading to deterioration in free cash flow.
• Compute projects take longer to power up, delaying revenue generation and investment returns.
• Increased financing needs could lead to equity dilution or higher funding costs.
• Launches, spectrum, satellite deployment, and AI infrastructure face regulatory approval delays.
• Consensus expectations for 2029 to 2031 may be too high, creating risk of medium- to long-term downward revisions.
• Morgan Stanley has investment banking and other business relationships with SpaceX, and investors should be aware of potential conflicts of interest.

What to watch

• Whether company-wide annualized revenue can reach or exceed $100 billion by the end of 2026.
• Whether compute capacity can exceed 2 GW by the end of 2026, and the actual progress toward expanding to 5 to 10 GW by the end of 2027.
• Starship's 14th flight and subsequent reuse tests, especially heat shield reliability.
• The pace of V3 broadband satellite deployment and the target of deploying 1,000 satellites by the second quarter of 2027.
• Starlink consumer user count, monthly average revenue per user, and growth in government and enterprise customers.
• Whether AI revenue per watt can avoid declining from approximately $31 in the fourth quarter of 2026 to approximately $17 in the fourth quarter of 2027.
• 2026 to 2027 capex, cost per watt of buildout, and external financing needs.
• Supply-demand and price performance after the new share unlock on August 6, 2026.
• The magnitude of upward revisions to consensus expectations over the next 12 to 18 months and forecast adjustments for 2029 to 2031.