Strategic AI Shift and Quarterly Performance
During the August 4 earnings call, SpaceX CEO Elon Musk outlined an aggressive roadmap for the company’s artificial intelligence division, marking a significant pivot in the firm’s business model. While aerospace remains a core operation, AI computing has emerged as a primary growth engine, with the department generating $2.6 billion in revenue this quarter—a figure that triples year-over-year and now surpasses the company’s traditional space launch revenue of $962 million.
Despite the rapid growth in AI revenue, the division currently operates at a $1.5 billion loss as the company invests heavily in infrastructure. Total capital expenditure for the quarter reached $18.37 billion, reflecting massive investments in power and cooling systems required to support large-scale AI data centers.
Grok Development Roadmap
The company’s model development cycle is accelerating, with Musk confirming the release of Grok 4.6 scheduled for next week. This will be closely followed by Grok 4.7 in approximately three to four weeks. The most significant milestone, however, is scheduled for the end of 2026 with the release of Grok 5. Musk stated that Grok 5 will integrate the entirety of SpaceX’s proprietary data, which he characterized as essential to creating the industry’s most capable AI engineer.
Infrastructure and Space-Based Computing
SpaceX has committed to an exclusive partnership with Nvidia, utilizing their Vera Rubin architecture for all AI computing services. Musk described this system as the current market leader. To support this, the company has set ambitious targets for computing power: reaching 2 gigawatts of online capacity by the end of this year and scaling to 15 gigawatts by 2027.
Beyond terrestrial data centers, SpaceX is extending its AI footprint into low Earth orbit. Through the ‘Starmind’ satellite program, the company is co-developing payloads with Nvidia to provide enterprise-grade space-based computing. The first batch of these satellites is expected to launch next year, marking a shift toward global, space-integrated AI infrastructure.