🚨 AI IS USING MORE POWER THAN EVER BEFORE.
Most people think AI is just software running on the cloud.
But behind every AI answer is a massive data center burning electricity every second.
And demand is rising fast as Microsoft, Amazon, Google, Meta, OpenAI, and others scale their AI systems.
The problem is simple.
Power systems were not built for this level of demand.
And building new ones takes years, not months.
That gap is starting to matter.
Companies tied to this shift include Eaton $ETN , Quanta Services $PWR , GE Vernova $GEV , and Constellation Energy $CEG .
The AI story is not just about intelligence.
It is also about power.
🚨 NVIDIA IS NO LONGER A CHIP COMPANY.
It's an infrastructure company.
In 2000, electricity built the industrial economy.
In 2025, compute is building the AI economy.
Every AI model.
Every AI startup.
Every AI data center.
Runs through $NVDA
The company isn't selling chips.
It's selling access to intelligence.
🚨 The S&P 500 just closed at another record. Nine straight weeks of gains, the longest winning streak in over a year.
And the man who runs America’s biggest bank is waving a caution flag.
Jamie Dimon spent last week warning that the market looks exuberant and that the risks are being underpriced.
Both things are true at the same time. That’s the part people forget.
Records and warnings don’t take turns. They show up together, every single time, right up until one of them finally wins.
The real question was never whether the market is high. It’s whether you’re in it with a plan, or just because it keeps going up.
🚨 THE STOCK MARKET RUNS ON A 100-YEAR-OLD EMOTIONAL LOOP. YOU’RE STANDING IN IT RIGHT NOW.
It climbs while everyone’s scared.
They pile in at the top.
It drops, and they swear it off for good.
Then it climbs again, while everyone’s scared.
The chart changes. The feelings never do.
🚨 EVERYONE SAYS BUFFETT IS WRONG TO SIT ON $330 BILLION IN CASH. THEY'RE MISSING THE POINT ENTIRELY.
He's not confused. He's not old. He's not missing the AI trade.
Buffett doesn't sell because he's scared.
He sells when he can't find value.
- At 25x earnings on the S&P, he can't find value.
- At $90 oil with inflation elevated, he can't find value. - - At $4 trillion market caps with negative free cash flow, he can't find value.
Cash isn't a mistake when nothing is cheap.
Cash is discipline.
The question isn't why Buffett is selling. The question is what does he know that the market is ignoring?
🚨 INFLATION HIT A 3-YEAR HIGH YESTERDAY. STOCKS SET RECORDS ANYWAY.
PCE, the Fed's preferred inflation gauge, came in at 3.8%. The highest reading in close to three years.
Nasdaq and S&P 500 closed at all-time highs the same afternoon.
If that looks like a contradiction, you are paying attention. Here is why it happened.
The number was bad. But it was not worse than expected. Markets had been bracing for 3.9%.
That is the move most people miss. Markets do not trade the data. They trade the gap between the data and what was already priced in. A less-bad-than-feared print is a relief, even when the print itself looks awful on paper.
The other piece: earnings are still strong. Real revenue is still rolling in, especially from the AI buildout. That is the engine the market is pricing, and inflation alone has not been enough to stall it.
The headline tells you what happened. The reaction tells you what the market thinks actually matters.
🚨 THE AI BOOM IS STUCK WAITING FOR A METAL BOX MOST PEOPLE FORGOT EXISTED.
Every AI data center needs more power than a small town. All of that power has to move through a big grey box called a transformer.
The world cannot make them fast enough.
Wait times for the largest electrical transformers have stretched into years, with the order books at the major makers booked solid well into the late 2020s.
So when a tech giant announces a $50 billion data center, the chips are the easy part. The power gear behind them is what actually holds the timeline hostage.
The companies sitting on top of that bottleneck: Eaton $ETN , Hubbell $HUBB , GE Vernova $GEV .
The chip names grab all the headlines. The unglamorous boxes that have to ship for any of it to work quietly run the entire buildout.
🚨 EVERY HOT SECTOR GETS CALLED A BUBBLE. MOST AREN'T.
Right now it is AI chips. Micron tripled this year and the bubble calls are everywhere.
So here is the actual test that separates a real bubble from ordinary hype.
Hype is a story with no cash behind it yet. A bubble is when the price has sprinted miles past even the best-case cash.
Ask three things.
- Is real money being earned, or only promised?
- Is everyone already in, with no buyer left to arrive?
- Is the price assuming a flawless future with zero room to disappoint?
Hype can climb for years. A bubble needs a greater fool, and eventually runs out of them.
The skill is not screaming bubble at every record. It is knowing the difference.
🚨 THE HOTTEST AI TRADE RIGHT NOW ISN'T NVIDIA. IT'S MEMORY.
Micron just crossed a trillion dollars. It has tripled this year, and most people cannot tell you why.
Here is what the smart money just worked out.
An AI chip is useless without ultra-fast memory feeding it data. That memory is called HBM, and almost nobody can make it.
Three companies on earth produce it at scale: Micron ($MU ), SK Hynix, and Samsung.
Demand is exploding. Supply is sold out years ahead, with buyers signing long-term deals just to lock it in.
So the market quietly stopped chasing only the famous chip names and started buying the memory makers behind them. SK Hynix just crossed a trillion dollars too.
The bottleneck is almost never where the headlines are pointing.
🚨 THE BETTER THE EARNINGS, THE HARDER A STOCK CAN FALL.
A company smashes its numbers. Record profit. The stock drops 10% anyway.
Everyone calls it irrational. It is the opposite.
The price already had the good news baked in.
Markets do not pay for what happened. They pay for what happens next.
A great quarter that only matches the hype is already priced in. Nothing new left to buy.
The stock falls when amazing was the assumption and the future looks a little less amazing.
Good news does not move a stock. Better than expected does.
This one idea makes earnings season stop feeling random.
Similar to what we're seeing with $ZS right now...
🚨 THE AI BOOM JUST QUIETLY BECAME AN ELECTRICITY PROBLEM.
Everyone is buying the AI story at the chip. The real story is the step they are skipping.
AI runs on data centers. Data centers eat staggering amounts of power.
The four biggest tech giants are now on track to spend over $300 billion a year building them.
That power has to be clean and constant, and basically one source fits. Nuclear.
So Microsoft is restarting Three Mile Island. Meta just signed a 20-year nuclear deal. Amazon and Google are funding reactors.
Here is the catch almost nobody prices in. New nuclear cannot arrive fast enough. The reactors are years out, late this decade at the earliest.
Demand is here now. Supply is late.
And that gap lands on you. US household power bills are forecast to rise 15 to 25% by 2030.
Follow the chain and the trade was never just chips. It is power, the grid, and whoever controls the electrons.
🚨 THE WEALTHY BUY ASSETS. EVERYONE ELSE BUYS LIABILITIES.
An asset puts money in your pocket. A liability takes money out.
A portfolio that pays dividends and grows. Asset.
A car that depreciates and eats fuel, insurance and repairs. Liability.
A flat you rent out that brings in money every month. Asset.
A credit card balance charging 22%. Liability.
The wealthy collect the first kind. Everyone else collects the second and calls them possessions.
The gap between the two groups was never income.
It is the ratio of assets to liabilities in their life.
Build the asset side. Starve the liability side.
🚨 THE INVESTING MISTAKE THAT COSTS MORE THAN ANY BAD INVESTMENT DECISION.
Checking your portfolio every day.
The research is unambiguous. Investors who check daily perform significantly worse than investors who check monthly or quarterly.
Not slightly worse. Significantly.
Daily checking leads to daily reacting.
Daily reacting leads to selling on fear and buying on greed.
That emotional pattern repeated over decades destroys returns that compound would have built.
The best investment strategy is not the most sophisticated one. It's the one you can execute without sabotaging yourself.
Set it up. Review quarterly. Let it compound.
Most people never figure this out and it's simple...
8% of the S&P 500 in a single stock reporting today.
Whatever Nvidia says about demand, Blackwell ramp and forward guidance moves the entire market. Not just the stock.
Miss on revenue or guide down and AI infrastructure names get hit across the board. Beat and raise and everything catches a bid.
Biggest earnings print of the year. No question.
🚨 HERE’S WHAT NOBODY TELLS YOU ABOUT BUILDING WEALTH IN YOUR 20s.
The math is simple. The behaviour is hard.
£300 a month at 22 years old invested at 10% annual return.
By 32 — £57,000. Feels slow. Most people quit here.
By 42 — £204,000. Starting to feel real.
By 52 — £603,000. Compounding taking over.
By 62 — £1,700,000.
You put in £144,000 over 40 years.
The market added £1,556,000…
🚨 TWO PEOPLE. SAME SALARY. COMPLETELY DIFFERENT LIVES AT 50.
Person A: earns → spends → saves what’s left.
Person B: earns → invests first → lives on what’s left.
Same income. Same job. Same city.
After 30 years at 10% annual return:
Person B has a portfolio generating more than their salary. Passively.
Person A still needs every paycheque.
The gap isn’t talent.
It isn’t luck.
It isn’t income.
It’s one decision made every single month.
Pay yourself first. Everything else follows.
Most people find out too late. Follow now or regret later.
@amitisinvesting Yields spiking and the market just shrugging it off because earnings keep coming in strong.
Last time the 10Y was near 4.7% things got ugly. But that was before $750B in annual AI capex was basically guaranteed spending regardless of rates.
Down 60%+ from ATH while still generating $8B in free cash flow a year.
The bear case is AI tools like Midjourney and Firefly eventually eat into their creative suite. The bull case is Adobe has the enterprise relationships, the compliance requirements and the workflow integrations that freelance AI tools can’t replicate.