• TD Cowen analyst John Blackledge initiated coverage of SpaceX (SPCX) with a Buy rating and $200 price target, implying roughly 37% upside from the September 28 close of $145.47.
  • The firm expects AI-compute leasing to become SpaceX’s largest business by Q1 2027, forecasting revenue of $14 billion in 2026, $66 billion in 2027, and $133 billion in 2028.
  • TD Cowen projects approximately 107 million Starlink subscribers and nearly 1,000 annual launches by 2031, contingent on Starship’s successful maturation.

AI Compute Takes Center Stage

TD Cowen launched coverage of SpaceX with a Buy rating and a $200 price target, framing the company less as a conventional launch-and-broadband provider and more as an emerging AI-infrastructure platform. Analyst John Blackledge argued that a persistent shortage of AI computing capacity makes SpaceX’s compute-leasing business the primary near-term catalyst, according to the initiation report.

The bullish thesis hinges on an exceptionally fast and capital-intensive buildout of terrestrial AI-compute capacity. Blackledge estimates that leased compute could generate $14 billion in revenue in 2026, $66 billion in 2027, and $133 billion in 2028. By the first quarter of 2027, AI-compute leasing is expected to become SpaceX’s largest revenue source, representing about 60% of total revenue in 2027 and 65% in 2028. Those forecasts assume capacity rises from 2.1 gigawatts in 2026 to 6 GW by the end of 2027.

Four customers are already leasing or beginning to lease compute capacity in 2026, including Alphabet/Google (GOOG) and Anthropic, according to the report. TD Cowen projects approximately $8.1 billion in fourth-quarter 2026 AI-compute revenue and $41 billion of annual recurring revenue exiting 2026. These are analyst estimates rather than independently confirmed contractual revenue disclosures.

Starlink and Starship Form the Backbone

Beyond AI, TD Cowen anticipates roughly 107 million Starlink subscribers and close to 1,000 annual low-Earth-orbit launches by 2031. Those figures reflect a highly bullish operating scenario that depends on Starship’s successful maturation, reusable-launch economics, spectrum access, global licensing, and continued demand for satellite connectivity.

SpaceX’s February 2 acquisition of xAI created the organizational basis for the AI-compute argument, but it also complicates financial reporting and capital allocation. The company’s principal operations now span launch (Falcon 9, Falcon Heavy, Dragon, Starship), connectivity (Starlink fixed broadband and direct-to-mobile), AI infrastructure (terrestrial compute leasing, xAI, and planned orbital-data-center systems), and government and national-security services (NASA crew and cargo, lunar-lander development, and defense launches).

The immediate economic driver is the global imbalance between demand for AI computing and the availability of power, chips, data centers, cooling, and grid interconnection. SpaceX’s announced Terafab project in Grimes County, Texas, illustrates the scale of the industrial response. The company says the initial phase is expected to require approximately $16.8 billion in capital investment from SpaceX and Tesla (TSLA), include more than 100 million square feet of planned manufacturing space, and employ at least 3,000 people. SpaceX also says its broader Texas presence has created more than 56,000 direct and indirect jobs and over $28 billion in estimated economic impact since 2024—figures that are company estimates rather than independent economic assessments.

Capital Intensity and Execution Risk

The compute, chip-fabrication, data-center, power-generation, cooling, and launch infrastructure implied by TD Cowen’s forecast will require extraordinary upfront spending. High revenue growth does not automatically translate into cash flow or profit. A search result from SpaceX’s investor-relations/SEC materials indicates consolidated revenue of $4.694 billion as of the referenced filing period and a reported loss in its AI segment of $7.723 billion. Because the primary SEC document was not retrievable in full here, these figures should be treated cautiously pending review of the underlying filing and period definitions.

Government permission is central to the growth model. Reaching nearly 1,000 annual launches would require a much larger regulatory, range-safety, manufacturing, and environmental-operating capacity than exists today. Starlink’s global expansion relies on spectrum rights, landing rights, and telecommunications approvals in individual countries. In September, the FCC reportedly granted SpaceX international Section 214 authority to carry Starlink Mobile traffic between the United States and other countries; this permits cross-border carrier operations but does not by itself launch service in every country or set consumer pricing.

Starship is at a consequential transition point. SpaceX says initial orbital missions beginning with Flight 14 are intended to deploy V3 Starlink satellites and test orbital operations, including an eventual propellant-transfer capability. It says each Starship launch could add more than 20 times the Starlink capacity of a Falcon launch carrying V2 satellites.

What to Watch

The critical tests for the thesis are measurable: whether announced AI-compute customers actually begin consuming capacity and whether leasing revenue appears at the forecasted scale; whether the company can build power-rich terrestrial compute sites on time and within budget; whether Starship achieves reliable orbital operation, deployment of V3 satellites, high reuse, and substantially higher flight cadence; whether Starlink continues subscriber growth while retaining spectrum and regulatory approvals across markets; and whether the combined SpaceX/xAI structure can manage large losses and capital spending without diluting public shareholders or impairing other operations.

The analyst consensus cited by CNBC is positive: 31 of 40 analysts rated the company Buy or Strong Buy, with an average target suggesting 57% upside. Such consensus, however, does not eliminate downside risk; it can also mean valuation is vulnerable if execution slips.

SpaceX reportedly began public trading in June after what CNBC described as the largest IPO in history, priced at $135 per share. The stock reportedly reached $225.64 before retreating to $145.47 by the cited September 28 close. This volatility illustrates that the valuation already reflects significant expectations for Starship, Starlink, and AI, even before the full AI-compute ramp has been demonstrated.

A slowdown in AI spending, lower compute prices, launch failures, regulatory setbacks, or cost overruns could materially weaken the forecasts. The proper interpretation of TD Cowen’s initiation is therefore not simply “SpaceX is a Buy.” It is a high-conviction bet that AI-compute scarcity persists long enough for SpaceX to scale capacity, and that Starship turns unprecedented launch cadence from a vision into a dependable industrial capability. SpaceX did not immediately respond to a request for comment.