SpaceX's upcoming initial public offering (IPO) at $135 per share, set to raise $75 billion, signals that Elon Musk is pursuing something far more ambitious than traditional space exploration. The company's regulatory filings reveal a deliberate strategy to build an end-to-end artificial intelligence (AI) platform that connects satellite internet, custom computing chips, AI models, and potentially terrestrial communications networks into a single, vertically integrated ecosystem.

This is not a merger with Tesla or a simple acquisition spree. Instead, Musk is methodically assembling the connective tissue of what could become a sovereign AI infrastructure empire, according to investment analysis. The IPO proceeds are earmarked for specific strategic purposes that paint a clear picture of where SpaceX's ambitions lie beyond rockets and satellites.

What Is SpaceX Actually Building With Its IPO Money?

SpaceX's S-1 filing, the regulatory document required for going public, reveals that some of the $75 billion in IPO proceeds will be allocated toward three critical areas:

• Nvidia GPU Procurement: SpaceX plans to purchase additional graphics processing units (GPUs) from Nvidia, the dominant supplier of AI training hardware, to power its AI infrastructure.
• Terafab Fabrication Facility: The company is building its own custom silicon manufacturing facility, called Terafab, to reduce dependence on external chip suppliers and develop proprietary computing hardware.
• Strategic Acquisitions: SpaceX's filing explicitly references appetite for acquisitions, signaling that the company intends to expand its technology portfolio beyond what it currently owns.

This three-pronged approach mirrors how successful tech companies have consolidated power in previous eras. Rather than relying on external vendors, SpaceX is attempting to control the entire stack: the chips that train AI models, the models themselves, and the networks that distribute them.

How SpaceX Is Consolidating AI Control Across Multiple Layers?

Musk has already made several strategic moves that demonstrate this consolidation strategy. SpaceX acquired xAI, the company behind the Grok generative AI model, bringing AI model development directly under SpaceX's roof. The company is also reportedly in discussions to acquire Cursor, a developer tooling platform, to extend its reach into software development infrastructure.

These acquisitions are not random. They represent a deliberate effort to build what former mergers and acquisitions analysts describe as a vertically integrated technology stack spanning compute, model development, and software tooling. Each layer serves a specific purpose in the broader ecosystem.

The missing piece, according to investment analysis, could be terrestrial communications infrastructure. Nokia, the Finnish telecommunications equipment manufacturer, operates one of the world's most dominant radio access networking (RAN) platforms. RAN refers to the distributed architecture of cell towers, antennas, and signal-processing equipment that connects networks to end devices like smartphones.

What makes Nokia particularly strategic is its AI-RAN platform, which enables real-time network optimization and on-device inference, meaning data processing happens locally on devices rather than being routed through distant data centers. This capability would be the terrestrial counterpart to what Starlink, SpaceX's satellite internet service, does from orbit.

Why Would SpaceX Want Nokia?

If SpaceX controlled both low-Earth-orbit broadband connectivity through Starlink and AI-native terrestrial radio infrastructure through Nokia, the company would not merely be competing in telecommunications. Instead, it would fundamentally reshape how data moves and where intelligence lives for enterprises and consumers at global scale.

The connection between SpaceX and Nokia already exists through Nvidia. In October 2025, Nvidia invested $1 billion in Nokia specifically to co-develop AI-RAN solutions. SpaceX, meanwhile, operates as a major customer of Nvidia's GPUs. A deal bringing Nokia into this ecosystem would create what analysts describe as a tightly coupled loop: Nvidia supplies AI chips, SpaceX owns the network layer from orbit to ground level, and Nokia provides the terrestrial radio infrastructure that ties the system together.

Such a structure would be more than an interesting telecom deal. It represents the path toward building a sovereign AI infrastructure empire that competitors would struggle to replicate. The combined stack would include Nvidia and Terafab accelerators training AI models on the ground, Nokia towers deploying inference on edge devices, and Starlink acting as a global distribution network.

How Does SpaceX's Strategy Compare to Other AI Giants?

SpaceX is not the only company pursuing an IPO in 2026. Anthropic, the AI safety-focused company behind the Claude family of AI models, and OpenAI are also expected to go public before year's end. Together, the three companies could raise approximately $240 billion at a combined valuation exceeding $4 trillion.

However, SpaceX's business model differs fundamentally from Anthropic's. SpaceX is pursuing a complex, multi-layered infrastructure play, while Anthropic is primarily a software business focused on developing and monetizing AI models. SpaceX's valuation of approximately $1.77 trillion reflects pricing in significant growth from revenue streams that do not yet exist, including space-based data centers and expanded AI services.

At its current valuation, SpaceX is trading at roughly 94 times sales, a valuation multiple that has never been seen with any trillion-dollar company in the market. By comparison, Anthropic, which recently raised capital at a $965 billion valuation, is expected to command a valuation of $1.2 trillion or more by the time it goes public, trading at approximately 25 times sales for a business whose revenue is growing exponentially.

The key difference is that Anthropic's valuation is based on exceptional revenue growth and actual business momentum, whereas SpaceX's valuation is based partly on potential future growth narratives that have not yet materialized. SpaceX's business produced approximately $18.7 billion in revenue in 2025, up 33 percent year-over-year, but the company posted a $4.9 billion net loss, though it was profitable on an adjusted earnings basis.

What Does This Mean for the Broader AI Infrastructure Market?

SpaceX's IPO filing and acquisition strategy suggest that Musk believes the future of AI is not just about building better models. Instead, it is about controlling the infrastructure that trains, deploys, and distributes those models globally. This represents a fundamental shift in how technology companies think about competitive advantage.

Rather than competing on model performance alone, SpaceX is attempting to become indispensable to the entire AI supply chain. By owning the chips, the models, the software tools, and potentially the networks that connect everything, SpaceX would create a closed-loop system that would be difficult for competitors to replicate or disrupt.

The IPO is scheduled to begin trading on June 12, marking the beginning of what could be a transformative period for how AI infrastructure is built and controlled globally.