Highlights

• The Trump administration is considering the Pentagon's OPEN AI program to set "reference prices" for critical minerals like germanium, gallium, antimony, and tungsten, enforced through coordinated tariffs to counter Chinese market dominance.
• AI models can estimate cost curves but cannot create liquidity or compel transactions in thinly traded markets; without allied coordination and increased non-Chinese refining capacity, arbitrage will undermine enforcement.
• This market engineering experiment tests whether algorithmic price discovery can substitute for capital guarantees and physical supply infrastructure—investors should focus on bankable contracts, not modeling transparency.

---

The Trump administration is reportedly considering using a Pentagon-developed AI model to calculate “reference prices” for certain critical minerals — then reinforcing those prices through coordinated tariffs within a proposed trade bloc. The aim is to counter alleged Chinese market distortion and create greater pricing stability for Western producers. It is a bold idea. Execution is another matter.

Algorithmic “Fair Value”

The OPEN (Open Price Exploration for National Security) program, launched by DARPA in 2023, is designed to model mineral pricing based on production costs, processing inputs, and supply dynamics — while adjusting for perceived non-market behavior. Reuters reports (opens in a new tab) that germanium, gallium, antimony, and tungsten are early focus areas — all thinly traded markets with limited price transparency.

It is true that these markets suffer from opacity and episodic volatility. Greater pricing clarity could, in theory, improve investment confidence. But models do not create markets.

Thin Markets, Hard Constraints

Thinly traded metals are thin because volumes are small and demand is fragmented. AI can estimate cost curves. It cannot create liquidity or compel buyers to transact at modeled prices.

If tariffs raise reference prices, manufacturers may absorb higher costs—or shift sourcing. Unless allied nations adopt a coordinated framework, arbitrage pressure will undermine enforcement.

Price architecture without supply architecture is fragile.

The Structural Reality

China still dominates processing in many critical minerals. Pricing power follows physical control. Unless non-Chinese refining capacity scales materially, an AI-derived benchmark risks becoming political architecture layered over unchanged supply reality.

Industrial policy without procurement commitments becomes advisory, not foundational.

The experiment here is not simply AI pricing. It is whether market architecture alone can substitute for capital guarantees.

Investors should watch one metric above all: bankable contracts.

The Bigger Signal

This is market engineering — an attempt to redesign price discovery rather than subsidize production directly.

That is innovative. It is also unproven at scale. Investors should distinguish between analytical transparency and physical leverage. Algorithms cannot substitute for processing plants.