Trump's AI Ownership Plan Sparks Opposition Between Musk and Vance; Cuban Questions Taxpayer Interests

On June 21, 2026, Trump proposed a new AI ownership plan. JD Vance supported a sovereign wealth fund model, while Elon Musk and Mark Cuban publicly opposed it, leading to a heated debate on X.

Core Content of the Plan and Technical Background

Trump's AI ownership plan aims to manage the large-scale data and computing resources that AI model training relies on through a government-led mechanism. The sovereign wealth fund model, proposed by JD Vance, specifically involves establishing a national fund to hold equity stakes in some AI companies or computing capacity shares, thereby distributing future returns.

This model involves technical resource scheduling issues. AI training requires tens of thousands of GPU clusters running continuously for months, with a single training cost potentially reaching hundreds of millions of dollars. If the fund directly holds computing capacity, it could reduce duplicate procurement through unified scheduling, but it would also require establishing cross-institutional data access protocols and model versioning systems.

Technical and Economic Doubts from Opponents

Elon Musk and Mark Cuban explicitly opposed this proposal. Musk pointed out that government ownership of AI assets could reduce private enterprises' incentive to update models, as the revenue distribution mechanism would dilute the return on direct investment. Cuban emphasized that if taxpayer funds flow into high-risk AI projects, the public would bear the losses from any failures.

From an architectural perspective, private companies currently maximize training efficiency through self-built data centers and dedicated networks. Introducing a sovereign wealth fund could turn computing capacity allocation into an administrative approval process, increasing latency and reducing hardware utilization. Real-world cases show that similar government funds in the semiconductor sector typically have payback periods exceeding ten years, while AI model iteration cycles have already shortened to months.

Technical Dimensions of the Debate on X

From June 21 to 23, 2026, discussions on X quickly intensified. Supporters argued that a sovereign fund could centrally purchase high-end GPUs, lowering hardware barriers for individual companies. Opponents cited specific data, noting that current leading AI labs already operate training clusters exceeding 100,000 GPU cards, and government intervention would struggle to match their customized network topologies.

Neither side provided a complete cost model comparison, but the public debate focused on a verifiable question: if AI inference service pricing is influenced by the fund, would it deviate from market supply and demand? In a reply, Musk directly stated that xAI's current training costs are already covered by private financing, and an additional government layer would only increase audit overhead.

Future Trends and System Impact

If the plan advances, AI system architecture may become a dual track: commercial models continue to pursue peak performance, while fund-supported projects focus on auditability and multi-party data fusion. This would change model release cadence and technology stack choices.

From an engineering practice perspective, the sovereign wealth fund would need to address technical challenges such as model weight hosting, gradient accumulation log retention, and cross-regional data compliance. These requirements differ significantly from existing private cloud API invocation models.