Space Exploration Technologies Corp (SPCX): UBS sees SpaceX entering 3Q with momentum across AI, connectivity and space operations

Overview

This 3Q preview argues that SpaceX is executing across its three segments—Space, Connectivity and AI—and that near-term earnings momentum is being driven most strongly by rapidly scaling compute services. UBS sees Flight 14, Starlink adoption and additional cloud-compute demand as supporting its Buy view, while recognizing unusually heavy capital requirements and execution risk.

Core views

UBS expects a strong third quarter after what it characterizes as a solid 2Q. Its 3Q revenue estimate is US$13.8B, versus Street expectations of US$12.9B, and its adjusted EBITDA forecast is US$7.5B, versus US$6.9B for the Street. The report attributes the expected beat chiefly to AI, forecasting US$7.6B of quarterly AI revenue as the Anthropic arrangement scales and Google begins to contribute. Space revenue is expected to rise to US$1.3B from about US$1B in 2Q, supported by 35 launches, while Connectivity revenue is projected to grow 59% year on year to US$4.9B. This growth comes with substantial investment: UBS estimates US$19.3B of capex and US$13.2B of quarterly cash burn as the company expands across several business lines. For AI, UBS argues that compute demand remains exceptionally strong. The company disclosed roughly 1M GPUs across Colossus 1 and Colossus 2, plus another 220K GB300 units expected to be operational around quarter-end; UBS estimates this represents about 1.8GW of compute. Management targets a further 220K–440K GB300 units by year-end 2026, implying 2.3–2.7GW in total, while UBS models 2.3GW at year-end 2026 and 6GW by year-end 2027. An additional compute deal supports estimated cloud ARR of about US$54B by December 2026. UBS forecasts AI revenue to reach US$15B in 4Q and US$27B for 2026 as Google contributes for a full quarter, alongside Cursor and new compute deals. The report views the supply-demand imbalance as acute, citing Microsoft's plans for approximately 26GW of incremental capacity by 2032. It also highlights progress by Grok and demand from GrokBot's agentic platform, while noting that Terafab in Grimes County, Texas is expected to begin producing the company's own memory and logic chips by end-2028. In Space, Flight 14 is treated as a meaningful, though incomplete, operational milestone. Starship achieved its first orbital mission on 28 September, reached a 275km orbit, deployed all 26 V3 satellites and subsequently executed controlled deorbit, re-entry and splashdown after an in-flight vehicle health assessment. The mission lasted about three hours rather than the originally planned six-orbit, 10-hour profile, and one Raptor engine was lost during ascent; nevertheless, UBS says primary objectives were achieved and views the mission as a partial de-risking event for Starship. Booster 21 also completed its main objectives with improved engine-relight performance relative to the prior flight. UBS expects Flight 15 by year-end, including an attempt to catch the ship for the first time, and models operational Starship launches rising from two in 2026 to 30 in 2027 and 141 in 2028. The longer-term launch thesis depends on achieving full and rapid reusability, with management discussing hundreds of annual launches by decade-end. The report links Starship progress directly to Connectivity expansion. Flight 14's V3 deployment added more than 25 Tbps of incremental capacity, and UBS believes this gives greater confidence that the V3 constellation can deliver a step change in capacity and speed. It estimates 3Q net subscriber additions of 2.3M, up from 1.7M in 2Q, taking total subscribers to 14.3M after the company surpassed 13M in August. UBS expects advertising to sustain subscriber gains before V3 enters operation. The firm expects Starlink to gain share in rural and less-dense suburban markets; in the US it expects the principal share loss to fall on cable operators, some competitive pressure on fixed wireless access, and limited impact on fiber. Globally, share gains versus cable and fixed-wireless providers remain dependent on pricing and local competition. UBS values the company using a sum-of-the-parts and multiples approach. Its US$210.00 target implies approximately 30x total-company EBITDA on its 2028 estimates. The published model forecasts revenue of US$49.775B for 2026, US$120.797B for 2027 and US$168.069B for 2028, alongside adjusted EBITDA of US$25.387B, US$66.386B and US$96.904B, respectively. Forecast equity free-cash-flow yield remains negative through the periods shown, reflecting the scale of capital deployment required to build launch, satellite and compute infrastructure.

Analysis framework

UBS combines segment-level revenue, EBITDA, capex and cash-flow forecasts with operational evidence from launches, V3 satellite deployment, subscriber additions and disclosed AI-compute capacity. It then uses a sum-of-the-parts and valuation-multiple framework to set its 12-month price target.

Methodology notes

• Valuation methodsSOTP (Sum-of-the-Parts) Valuation

  Sum-of-the-parts valuation

  UBS values SpaceX by considering its distinct Space, Connectivity and AI businesses, supplemented by valuation multiples; the resulting target implies about 30x total-company 2028 EBITDA.

• Industry AnalysisSupply-demand framework

  AI compute supply-demand imbalance

  The report uses disclosed GPU capacity, planned expansions and demand from cloud customers to argue that compute demand remains stronger than available capacity.

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  Starship launch capability feeding Starlink capacity and broadband growth

  UBS links successful Starship operations and V3 deployment to more satellite capacity, faster connectivity service and potential subscriber and market-share gains.

Asset mapping & comparison

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

• Space Exploration Technologies Corp (SPCX.O)

  Primary covered company; benefits in UBS's view from AI compute growth, Starship progress and Starlink capacity expansion.

  Strengths

  Large disclosed compute build-out, additional cloud deals, successful Flight 14 objectives and rising Starlink subscribers.

  Weaknesses

  Heavy capex and estimated quarterly cash burn while multiple businesses are scaled.

  Comparison

  UBS expects Starlink mainly to take US share from cable, to provide some competition to fixed wireless access and to have limited impact on fiber operators.

  Risks

  Dependence on Elon Musk, technology delays, AI monetization and scale risks, regulatory, geopolitical and competitive pressures, supply-chain execution and unproven space-based compute models.

Key data

• 3Q26 revenue estimateUS$13.8B7% above Street estimate of US$12.9B
• 3Q26 adjusted EBITDA estimateUS$7.5B9% above Street estimate of US$6.9B
• 3Q26 AI revenue forecastUS$7.6BExpected to lead segment growth
• 4Q26 AI revenue forecastUS$15BIncludes full-quarter Google contribution and new compute-deal ramp
• Cloud ARR estimate~US$54B by December 2026Supported by an additional post-earnings compute deal
• Connectivity revenueUS$4.9BUBS estimates 59% year-on-year growth in 3Q
• Connectivity subscribers14.3MUBS estimates 2.3M net additions in 3Q versus 1.7M in 2Q
• 3Q capex and cash burnUS$19.3B capex; US$13.2B cash burnReflects expansion across multiple business vectors
• Flight 14 V3 deployment26 satellites; 25+ Tbps incremental capacityStarship's first direct deployment into the operational constellation

Impact & implications

UBS believes AI-compute scaling can drive the near-term earnings upside, while Starship execution and V3 deployment underpin a longer-duration expansion in launch and connectivity capacity. The report's valuation case depends on these businesses scaling despite sustained negative free cash flow and high capital needs.

Risks

• The report identifies dependence on Elon Musk as a key decision-maker.
• Delays in technological milestones or failure to achieve rapid Starship reusability could slow launch and capacity expansion.
• AI initiatives may fail to scale or monetize as expected.
• Regulatory, geopolitical and competitive pressures could affect execution.
• High capital intensity, supply-chain execution risk and unproven space-based compute models remain material.

What to watch

• Timing and outcome of Flight 15, including the planned first ship-catch attempt.
• Progress toward operational Starship launches, modeled at 30 in 2027 and 141 in 2028.
• AI compute additions toward management's 2.3–2.7GW year-end 2026 target and UBS's 6GW year-end 2027 model.
• Conversion of cloud-compute deals into the projected ~US$54B cloud ARR by December 2026.
• V3 constellation deployment, Starlink subscriber additions, pricing and competitive response from cable and fixed-wireless providers.
• Terafab construction and the targeted end-2028 start of internal memory and logic-chip production.