Softbank is committing an initial $19 billion, with a total planned investment of $500 billion over four years, to build terrestrial data centers for AI, according to Fortune. An initial $19 billion, with a total planned investment of $500 billion over four years, meets the escalating demand for AI compute, contrasting sharply with Elon Musk's orbital ambitions. Skepticism already surrounds Musk's idea for space-based AI data centers, according to The Wall Street Journal, questioning the near-term feasibility of a 2026 impact for such solutions.

The immediate, massive demand for AI compute drives vast terrestrial investments. Yet, Elon Musk pushes a long-term, space-based solution. Critics argue this orbital approach is too slow and costly for the current battle, making its general applicability questionable.

Terrestrial data centers will dominate the immediate AI compute landscape. Still, Elon Musk's orbital venture is poised to secure significant future business for SpaceX, creating a bifurcated infrastructure future most companies are not yet prepared for.

Musk's Orbital Ambition Takes Shape

  • SpaceX acquired xAI in February 2026, integrating its data center operations and Grok AI models into SpaceX's orbital compute framework, according to Crypto Briefing.

SpaceX actively builds foundational components for its orbital AI compute network. Integrating xAI's operations and models establishes an in-house AI capability for space-based processing. Integrating xAI's operations and models establishes an in-house AI capability for space-based processing, aligning with Musk's broader vision for a self-sustaining space economy.

The Terrestrial Counter-Offensive

Softbank focuses on building data centers on Earth, backing OpenAI's Stargate project with an initial $19 billion commitment, according to Fortune. The total planned investment reaches $500 billion over four years. The initial $19 billion commitment and total planned investment of $500 billion over four years signals the industry's immediate priority: rapid, ground-based deployment for the urgent AI compute crunch, challenging the timeline of orbital alternatives.