🚨 Anthropic just showed a 27-minute workshop on how to actually do prompts for Claude.
Taught by the people who built it.
Free. No registration. No paywall.
I've seen $300 courses that don't cover what they teach in the first 8 minutes.
Watch the session
☢️ THE URANIUM RACE: Meta, Amazon and Microsoft are now signing for nuclear power the way they used to sign for cloud contracts. The uranium to feed all of it does not exist yet.
This is the railroads moment for the AI buildout, and the market is still pricing it like a boring utility.
Uranium is having the strangest bull market I've ever watched. The price ripped past 101 dollars in January, then fell asleep for six months while the spot tape did nothing.
Meanwhile governments kept announcing reactors and the hyperscalers kept signing for nuclear power like it was cloud compute. The demand side is screaming and the price is whispering.
URANIUM MINERS
The miners are the leverage trade. When a scarce commodity finally reprices, the companies pulling it out of the ground earn on every pound at the new price while their costs barely move. That's why this group runs harder than the metal itself in every cycle. It's also why it falls harder when sentiment turns.
The catch is that not all of them are actually mining yet. Some are producers with real pounds and signed contracts. Others are developers selling a story about pounds that show up at the end of the decade. In a market this early, I'd rather own the ones already delivering into a rising term price than the ones still permitting a hole in the ground.
Tickers: $CCJ , $UEC , $NXE , $DNN and $UUUU
🪨 PHYSICAL URANIUM
This is the purest way to own the thesis. No mines, no drills, no management team fumbling a ramp. Just pounds of uranium sitting in a vault and a share price that tracks them. If you believe the metal reprices and you don't want to bet on any single operator's execution, this is the cleanest expression there is.
The tradeoff is you give up the leverage. Physical goes up with the metal and not a dollar more, while a good miner multiplies it. So I think of this layer as the anchor, not the moonshot. It's where you park conviction when you want to be right about uranium without having to be right about a company.
Tickers: $SRUUF and $UROY
☢️ NUCLEAR UTILITIES
These are the buyers who cannot walk away. A reactor needs fuel whether the spot price is 40 dollars or 140, and the operators running the fleet have to keep the lights on for millions of people. That's the demand floor under this entire market. It does not blink.
What changed is that these operators went from nobody's favorite stock to the cleanest AI trade on the board. The data center buildout needs power that runs all day and all night, and only one carbon free source does that at scale today. Owning the fleet is owning the toll booth on that demand. Slower than the miners, far more durable.
Tickers: $CEG , $VST , $DUK and $SO
🔋 REACTOR TECHNOLOGY
This is the lottery ticket layer, and I mean that as a compliment. Small modular reactors and advanced designs are the bet that the next wave of nuclear looks nothing like the giant plants of the past. Factory built, smaller, faster to site, dropped right next to the thing that needs the power.
Be honest about what you're buying here. Most of these companies have more PowerPoint than revenue, and the timelines are long and slippery. But if even one design becomes the standard the hyperscalers order by the dozen, the winner is enormous. High risk and genuinely asymmetric. Size it small.
Tickers: $OKLO , $SMR , $GEV , $BWXT and $RYCEY
📈 URANIUM ETFS
If picking the individual layer sounds like work, this is the shortcut. One ticker buys the whole basket and you stop worrying about which miner missed guidance or which reactor startup slipped another year. For most people who just want exposure to the trend, this is the honest answer.
The difference between them matters more than people think. Some lean into the big producers, some hunt the tiny juniors, some blend in the reactor and utility names for a broader nuclear bet. Pick the one whose tilt matches your appetite. A junior heavy fund and a diversified nuclear fund are two completely different rides.
Tickers: $URA , $URNM , $URNJ , $NLR and $NUKZ
🔎 FOREIGN LISTED (flagged separately, not on US exchanges)
$KAP.L is Kazatomprom, the largest uranium producer on earth. The map had it under physical uranium, but it's a miner, the biggest one there is. Listed in London and Astana. Trades over the counter in the US as NATKY.
$YCA.L is Yellow Cake, a pure physical holder just like the Sprott trust. Listed in London. Over the counter in the US as YLLXF.
$U.UN is the Toronto listing of the exact same Sprott Physical Uranium Trust as above. Same pounds, same trust. Two tickers.
🧠 FINAL THOUGHTS
My honest take after mapping the whole thing: the crowd is loud about the miners and quiet about the two layers that matter most. The physical trusts, Sprott and Yellow Cake, are the cleanest way to be right about the metal without gambling on a single company. And the utilities, Constellation and its peers, are the demand that literally cannot quit, now wearing an AI costume the market is only starting to price.
Here's the claim I'll defend in the replies. The reactor technology names are where both the fortunes and the wipeouts live, and most people buying them have no idea which one they actually own. NuScale, Oklo, BWX Technologies and Rolls Royce are not the same bet in different wrappers. One of them might print. The rest are options that expire worthless if the timelines keep slipping.
"The scale of China's oil demand collapse has been so dramatic that Chinese policymakers are reportedly examining whether this historic slump reflects a temporary response to elevated global prices or a more structural shift in consumption patterns." - JPM
Crack spreads are behaving as if oil were still $110.
They're almost as high as the price of a barrel of oil...Truly unprecedented.
Consumers don't buy crude.
They buy gasoline, diesel, and jet fuel.
Demand for refined products remains extremely strong while inventories continue to fall.
The clock is ticking...
How fast is AI debt rising? In Q3 '25 total hyperscaler IG leverage was 0.9x. 6 months later it is 1.8x, bigger than all of Energy. It is rising at more than 0.3x turns per quarter (and that excludes SPVs). In one year, hyperscalers will have the highest leverage ex utilities
‼️THIS IS INSANE:
46 of 68 global central banks are currently overshooting their inflation targets.
The US is running +1.8 percentage points above its 2% target, and the Eurozone +1.2 percentage points above.
Meanwhile, markets are pricing a 98% probability of a 25bps rate cut from the ECB on June 11th and an 83% probability of a 25bps hike from the Bank of Japan on June 16th.
The first FOMC meeting under new Fed Chair Kevin Warsh follows on June 17th.
If the outcome is too dovish, the long end of the Treasury market heads toward 6%; too hawkish, and the S&P 500 pulls back toward 7,000.
Central banks are historically failing to fulfill their most basic mandate.
During WNFM last week, Jonathan Hinze, president of @UxC_Nuclear stated that based on the 'known' growth of nuclear capacity (~400GW today to ~500GW ~2032) he is concerned about the supply chain. When pressed on this, he stated that he is most concerned about "uranium." (1/x)
Sulfur prices have nearly doubled from here
They are at all-time highs and up 140% since the beginning of the Iran War 🇮🇷
This ought to have consequences on fertilizer and commodities such as Ags, Nickel, Lithium, Copper, Silver