Goldman just rewrote the uranium deficit math to include small modular reactors (SMRs), and the number they landed on flips the entire fuel cycle math for the next two decades.

The three names below sit directly in that flow of capital: one builds the reactors, one builds the turbines and grid gear that feeds them, and one owns the fleet the hyperscalers are already lining up to buy from.

Goldman finally did the math the uranium market has been dancing around for two years. By adding small modular reactors to its long-term demand model, the bank now sees a 2.3 billion pound uranium supply deficit by 2045.

That is not a rounding error. That is a structural gap mines, exploration, and enrichment capacity cannot fill on the current build schedule. If you own uranium names, that 2045 gap is the thesis.

The bank raised its long-term demand estimate by roughly 6% and tacked on 62 million pounds of additional SMR uranium demand alone.

Term contract requests from utilities are climbing every week, according to industry consultants, and the URA uranium ETF popped 4.5% Monday on the news. Track weekly utility term contract activity as your tell.

Here is what makes this real for you: SMR demand was previously treated as speculative. Now a Wall Street bank is embedding it in base case forecasts, which means every utility procurement officer just got a new floor for their price assumptions.

When the model changes, the term contracts change. When the term contracts change, the equities move, and that is where you decide whether to act before the repricing or after.

Your takeaway: You do not need spot uranium to rip to make money here. You need term price contracts to keep drifting higher and utility buyers to keep signing longer-dated agreements.

That is already happening. Position in the fuel cycle and the equipment layer before Q3 utility procurement announcements start rolling in.

Premier Li Qiang cleared eight new nuclear reactor units in a single executive meeting on July 31. Four of them use the Hualong One 2.0 platform. Total investment: roughly $25 billion across four provinces. Mark that date on your uranium thesis.

For context, that is more reactor approvals in one afternoon than the United States has permitted in a decade.

China is now on a run-rate of building more nuclear capacity than the rest of the world combined, and the fuel commitments that come with those reactors are already being written.

If you are sizing uranium exposure, that run-rate is your baseline demand signal.

The knock-on effect matters more than the headline. Every gigawatt of Chinese nuclear capacity coming online tightens the global uranium market you are trading.

It also tightens the market for reactor components, pressure vessels, and turbine sets, a supply chain that the West spent 30 years hollowing out.

The Kremlin, meanwhile, still supplies enriched fuel to about a fifth of US electricity generation, and Washington's 2028 import ban is barreling toward a shortfall the domestic enrichment base cannot cover.

Watch 2028 and US enrichment names as the pinch point.

Your takeaway: Think about this in layers. Fuel supply, enrichment, and equipment are all constrained. Any US-listed name touching those three chokepoints is going to see multiple compression flip to multiple expansion.

Do not chase the pure-play uranium miners on green days. Own the equipment and reactor developer layer where the story compounds.

The World Gold Council data landed Monday and the surprise was not the 289 tonnes central banks bought in Q2 2026, up 62% year over year. The surprise was who joined the party.

The Bank of Korea returned to gold buying for the first time since 2013. Poland scooped up another 51 tonnes on its way to a 700-tonne reserve target. The People's Bank of China added 33 tonnes, its biggest quarterly buy since Q4 2023.

Total H1 2026 official sector buying hit 345 tonnes even after a downward revision to Q1, so if you own gold, that is the bid you are trading alongside.

Here is the punchline that gets lost in the tonnage tables: gold prices tend to rise roughly 1% for every 20 to 30 tonnes central banks buy above the 600-tonne annual average.

Q2 alone came in at almost half the annual average in three months. That is the bid underneath every gold pullback you are seeing right now, and it is why gold at $4,090 is still trading well off its $5,590 peak but refusing to break down.

Your takeaway: Central bank buying is a structural bid, not a trade. When BOK, Poland, and PBOC are all accumulating in the same quarter, you have a de-dollarization signal you can actually price.

Keep gold exposure through miners with expanding margins, not through the metal ETFs alone.

Oklo is building small fast reactors, and the customer list reads like a hyperscaler roll call. The Aurora powerhouse platform is designed for behind-the-meter deployment at data centers, which is exactly the demand vector Goldman just formalized in its uranium model.

The company has more than 14 GW of customer interest in its pipeline, an INL site allocation, and a Department of Energy fuel award that lets it move faster than any competing SMR developer.

You are looking at a name where the catalyst calendar is dense: combined license application progress, additional utility offtakes, and any executive order tailwind on domestic nuclear deployment.

If you want reactor developer optionality without the balance sheet risk of a pre-revenue miner, OKLO is the cleanest way to own the SMR thesis on a US exchange.

Every reactor Goldman just modeled needs a turbine, a generator set, and grid gear to feed the electrons back into the system. GE Vernova sits at the center of that supply chain and has a $130B-plus backlog to show for it.

The nuclear services segment is expanding, the company is a partner on BWRX-300 SMR deployments, and the gas turbine side is a bonus that hyperscalers keep signing capacity contracts against.

The stock has already worked, but the setup here is about backlog conversion into margin. The order book is locked in through 2028 on power equipment, and pricing power keeps improving with every announcement of a new AI data center campus.

This is the pick-and-shovel play on the entire nuclear buildout, not just a single reactor design.

Vistra owns one of the largest nuclear fleets in the country, and it has already inked hyperscaler power deals that value those reactors materially higher than the market gives it credit for.

The Comanche Peak and Perry stations are generating capacity you cannot replicate, and the merchant power exposure means every tick higher in PJM and ERCOT capacity auctions flows straight to the bottom line.

Free cash flow is expanding, the buyback is active, and management keeps guiding higher on the long-term earnings power of the nuclear fleet.

If central bank gold buying is telling you about the monetary side of the trade, hyperscaler nuclear contracts are telling you about the physical side. You want to own both. VST gets you the physical.

• Gold: $4,090/oz, down about 5% from the $4,325 Jan 1 open and well off the $5,590 peak. Central bank buying of 289 tonnes in Q2 provides the floor.

• Silver: $62/oz, down about 22% from the $80 Jan 1 open after peaking at $121 earlier this year. Industrial demand still strong, but the correction is not over.

• Copper: $6.66/lb, up about 28% year to date from the $5.20 Jan 1 open. Trading at 52-week highs on tight scrap supply and grid buildout demand.

• Uranium: Term contract requests keep grinding higher week after week. Goldman's new 2.3B lb deficit by 2045 sets a fresh floor for utility procurement assumptions.

• Lithium: Carbonate prices holding near an $18,000/tonne floor as inventory drawdowns continue and LFP producers are fully booked through year-end. Structural deficit setup building for 2027.

• Rare Earths: Malaysia is weighing conditional easing of its raw export ban. China's October 2025 export controls remain suspended until November 2026, which is a hard deadline every Western processor is racing against.

• Cobalt: Prices reshaping the EV supply chain, with vertical integration accelerating as automakers move upstream. DRC (Democratic Republic of Congo) shipments to China still dominant.

• Platinum: Firming on hydrogen economy demand and constrained South African mine supply. The setup for a break above recent highs is building.

Metal Trend Exploration Focus

The uranium deficit math got a Wall Street stamp of approval this week, and that is the kind of signal that changes term contract behavior across the entire utility complex.

Combine it with central bank gold buying at the highest run rate in three years, and you have two structural bids under the resource sector that were not fully priced in at Q2 earnings.

Copper keeps grinding higher on grid demand, lithium is basing before the next EV inventory cycle, and the rare earth supply chain is being rewired in front of the November 2026 China export control deadline.

Every one of these threads points the same direction. Get your positioning right before the term contract cycle catches up to the models.

— Noah Zelvis

Resource Brief