The Manhattan Project Needs Silver
Why the most important AI commodity isn't chips.
Software scales infinitely. The physical world doesn’t.
AI is the new Manhattan Project. Every breakthrough in the stack requires infrastructure: data centers, power generation, batteries, connectivity, military systems. All of it needs raw materials. And one material sits at the center of nearly everything being built.
Silver.
The Supply Trap
The silver market has run structural deficits for five consecutive years. Total shortfall: 678 million ounces. That’s not a temporary imbalance—it’s a trend.
Here’s the problem: you can’t just mine more silver.
70-80% of silver production is byproduct from copper, lead, and zinc mining. To increase silver supply, you’d need to massively expand base metal mining. New mines take 7-10 years from discovery to production. Supply is structurally constrained by geology and economics.
Current annual supply sits around 820 million ounces. That number cannot scale quickly. The infrastructure doesn’t exist.
The Demand Collision
Now look at what’s competing for that fixed supply.
Solid-state batteries: Samsung’s new technology requires roughly 35 ounces of silver per vehicle—20-40x more than current lithium-ion EVs. BYD is developing the same tech. Both are targeting 2027 mass production. Samsung has already prepaid a Mexican mine for two years of supply. They’re not waiting for the market to figure this out.
Military rebuild: The US has depleted 25% of all THAAD missiles ever made in recent conflicts. $20 billion is allocated for munitions rebuild. Silver is required in missiles, jets, drones, and guidance systems. The US no longer maintains a national silver stockpile.
Solar deployment: Each gigawatt of capacity requires roughly 20 metric tons of silver. No viable substitute exists. The world is deploying panels at record pace.
AI infrastructure: Every chip, every connection, every server in every data center needs silver. The explosion of compute is an explosion of demand.
Robotics: Billions of units coming—military and commercial. All needing batteries and electronics. All needing silver.
These aren’t competing trends. They’re stacking. And they’re all accelerating simultaneously.
The Impossible Equation
Projected demand trajectory: 2-4 billion ounces annually.
Current supply: 820 million ounces.
Supply growth potential: minimal for 7-10 years.
Someone doesn’t get their silver. The only question is who, and at what price.
The Giffen Good Dynamic
Here’s what makes this different from normal supply crunches.
When silver prices rise, industrial buyers don’t reduce demand—they accelerate it. Executives see the shortage coming. They panic. They hoard. Demand increases precisely because price is rising.
It’s not “we’ll find a substitute” or “we’ll use less.” It’s “get silver or shut down the factory.” It’s “get silver or can’t build the missile.” At that point, price is whatever it takes.
The higher it goes, the more urgent the buying. Self-reinforcing feedback loop.
The Positioning
Smart money has already moved.
For the first time in recorded history, US banks have flipped net long on silver. JP Morgan—the bank that ran silver suppression for decades—is now pulling physical metal out of COMEX vaults.
European banks are holding the short position. 215 million ounces short, against 127 million registered for delivery. The math doesn’t work. 88 million ounces of that position simply don’t exist.
China implemented export controls January 1, 2026. They control 60-70% of global refined silver output. They understood the leverage before anyone else did.
The US added silver to the Critical Minerals List in November 2025. Trump invoked Section 232 authority for critical minerals in January. $7.5 billion allocated to critical minerals initiatives. $2 billion for national defense stockpile rebuild.
Governments and corporations aren’t waiting for price discovery. They’re securing supply directly, at any cost. The market mechanism is being bypassed.
What Breaks
The US government will get its silver. So will China. So will Russia. Governments with military production needs don’t ask politely—they secure supply by whatever means necessary. That’s already happening.
Major corporations with enough foresight are next. Samsung didn’t prepay a mine because they enjoy tying up capital. They did it because they understand the pecking order.
Everyone else competes for what’s left.
When demand exceeds supply and supply can’t scale, something breaks. Could be European banks holding short positions that don’t exist. Could be industrial production lines going idle. Could be solar deployment targets that quietly get pushed back.
The price is going up. That’s not a prediction. It’s arithmetic.