Inside the Largest IPO Ever: Breaking Down the SpaceX S-1

· Source: Tanay’s Newsletter · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Intermediate, long

Summary

SpaceX's S-1 filing reveals a company valued at ~\$1.75T, preparing for a ~\$75B IPO, structured into Space, Connectivity (Starlink), and AI (xAI) segments. The Space segment, founded in 2002, achieved launch costs of ~\$2,700 per kilogram, 85% below the industry average, enabling internal Starlink deployments that comprised 74% of its 165 projected 2025 launches. Starlink, the Connectivity segment, is highly profitable, generating \$11.4B revenue in 2025 with a 39% operating margin, growing 50%, and serving 8.9M subscribers. The AI segment, xAI, initially showed a \$6.4B operating loss on \$3.2B revenue in 2025, but recent deals, including a \$1.25B/month compute agreement with Anthropic and a potential \$60B acquisition of Cursor, project a \$40B-\$41B run-rate by late 2026. This long-term vision, heavily reliant on Starship, includes orbital AI compute data centers by 2028 and multiplanetary colonization, reflected in Elon Musk's pay package.

Key takeaway

For investors evaluating high-growth, high-risk ventures, SpaceX's S-1 highlights a unique long-term play. You should consider how the company's profitable Starlink business funds ambitious Starship development and a rapidly scaling AI segment, now bolstered by significant compute contracts. Your assessment must weigh the execution risk of orbital compute and multiplanetary goals against the potential for unprecedented market creation.

Key insights

SpaceX aims to be the physical infrastructure for the AI and space economy, leveraging cheap launch to enable orbital compute and multiplanetary life.

Principles

Method

SpaceX's strategy involves developing fully reusable Starship rockets to enable cheap mass-to-orbit, deploying large satellite constellations for connectivity and orbital compute, and integrating AI models and applications.

In practice

Topics

Best for: CTO, VP of Engineering/Data, Executive, Investor, Director of AI/ML, Consultant

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Tanay’s Newsletter.