As the initial public offering deadline nears, SpaceX Chief Financial Officer Brett Johnsen has systematically articulated the strategic logic underpinning the company's five core business pillars. This cohesive ecosystem centers on Starship launch capabilities while integrating Starlink, orbital computing, terrestrial AI infrastructure, and self-developed chip manufacturing. In a detailed discussion with investor Gavin Baker, Johnsen disclosed that SpaceX's AI infrastructure division generates annual revenues of $3.75 billion, securing its position among the top five global AI infrastructure providers. He characterized the recent maiden flight of the Starship as a 'huge success' and projected that once second-stage vehicles achieve rapid reusability, the cost per kilogram of payload to orbit will decrease by a factor of 10 compared to the Falcon rocket. Data compiled by Woofun AI shows that this cost reduction is critical for unlocking the economics of large-scale orbital deployment.

The capital logic and timing presented by Johnsen reveal a tightly coupled financial engine: Starship launches facilitate the deployment of Starlink V3 satellites, driving business growth and generating essential cash flow. This liquidity supports the substantial capital expenditures required for orbital computing initiatives, while terrestrial data centers serve to validate the broader business model.

Concurrently, the Terafab project ensures the scalability of the entire system. Johnsen summarized this framework as 'all these elements working together in harmony,' emphasizing that each business segment remains open to external competition. Competitors can purchase SpaceX launch services, Starlink is available to all users, terrestrial computing power is sold externally, and the Grok model is accessible to anyone. This strategy enhances vertical integration efficiency while expanding operational scale to further drive down costs. Woofun AI notes that this open-market approach distinguishes SpaceX from traditional vertically integrated monopolies.

Brett Johnsen emphasized that launch capability serves as the foundational bedrock for all SpaceX operations. 'If you want to be a space company, you must first ensure that you have a reliable way to access space,' he stated. SpaceX currently maintains the lowest cost per kilogram of payload to orbit in the industry, with Starship engineered to reduce this metric by another factor of 10. The recent successful E3 flight demonstrated the capabilities of the new V3 Raptor engine, modifications to the vehicle's underside, and second-stage soft landing technology. Johnsen described this as 'a demonstration of the entire system's capabilities' and expressed confidence that it lays the groundwork for future missions. The significance of Starship extends beyond cost efficiency to its payload capacity; it can deliver 100 tons to low Earth orbit, a prerequisite for commercializing large-scale payloads essential for the orbital computing business.

To accommodate the anticipated high frequency of launches, SpaceX is constructing two new launch sites in southern Texas. The first site at Cape Canaveral is nearly complete, while the second is scheduled to become operational within 1 year. Johnsen stated that the company is preparing to expand production capacity to handle thousands of launches annually. Starlink currently serves over 10 million users and has deployed more than 10,000 satellites across more than 160 countries. Johnsen believes that 10 million users represents merely the beginning, with significant growth expected. He noted that the global internet and cellular connectivity market is valued at approximately $1.6 trillion, with Starlink's competitive advantages rooted in speed, low latency, and wide coverage. Compared to terrestrial solutions, satellite networks offer extremely low marginal costs for coverage expansion, eliminating the need for trenching or pipeline laying.

The growth trajectory of the satellite network is primarily driven by direct mobile services. Johnsen mentioned that the next generation of direct-connected devices will achieve 5G quality within 2 years, enabling true global coverage 'whether you are in the heart of the desert or on the highest mountain.' He predicted this new segment will be highly profitable and complement SpaceX's broadband services. Orbital computing is described as a 'market that truly requires Starship to realize its potential' due to its reliance on large-scale payloads and cost sensitivity. Johnsen expects to demonstrate orbital computing capabilities as early as next year. Technically, these satellites function as 'racks in space'—essentially enlarged Starlink V3 units equipped with NVIDIA GPUs and heat-dissipating metal plates, utilizing existing inter-satellite communication links and propulsion systems. Woofun AI analysis suggests that the shift to orbital computing represents a paradigm shift in data center economics.

The cost logic for orbital computing is highly favorable: solar panels in space generate approximately 5 times more energy than terrestrial counterparts with no atmospheric loss. In a vacuum environment, radiative cooling replaces liquid cooling, eliminating land leasing costs. As satellite launch scales increase, the primary cost factor—silicon—will continue to decline according to Moore's Law. 'The cost curve for terrestrial data centers is rising, while the advantages of orbital computing will only become greater,' Johnsen observed. Prior to widespread orbital deployment, SpaceX utilizes terrestrial data centers to validate the model. The partnership with Anthropic for computing power hosting has generated the aforementioned $3.75 billion in annual revenue. SpaceX's AI business also encompasses the Grok model, enterprise API services, and the recently acquired code tool Cursor. Johnsen revealed that Cursor is already used by more than half of the Fortune 500 companies, with internal models showing significant performance improvements after fine-tuning on SpaceX's AI clusters.

Johnsen highlighted that a key differentiator for SpaceX's AI business is the real-time data feed from the X platform. 'We must ensure that the AI models people rely on truly pursue truth, and the integration of real-time content from X is an important element that distinguishes our AI products,' he said. The primary motivation behind developing Terafab was supply chain security. 'Whether it's NVIDIA chips or TPUs, ultimately, they all depend on TSMC. If our goal is to launch 100 gigawatts of computing power per year, we must ensure a stable supply of chips, free from the impact of tariffs,' he explained. Terafab is a collaborative project between SpaceX and Tesla, with Intel joining to bring decades of semiconductor experience. Johnsen believes that having SpaceX and Tesla as anchor customers committing to purchase all produced chips eliminates the primary risk of building new wafer foundries, leaving only capital risks. He expressed optimism about applying first-principles methods to the semiconductor industry, noting that Elon Musk has already proven the effectiveness of challenging established manufacturing processes in other sectors. When asked about the importance of launch capability, Johnsen reiterated that it is the core enabler for Starlink, direct mobile services, and AI computing, with Starship representing the next evolutionary step in rapid reusability.