Most people don't understand $GLXY's crypto business.

So let me clear it up.

They are NOT mining Bitcoin like $SLNH or $CLSK.

Galaxy actually exited proprietary mining entirely.

They don't produce a single coin.

No rigs to depreciate, no hash rate to chase, no "spend 100 to make 1" mining economics.

That's a completely different business.

It's not theirs.

So what IS the crypto business?

They're intermediaries.

A merchant bank for crypto.

Galaxy makes money on transactions, not on the price of Bitcoin going up.

Trading desks, OTC execution, lending, structured products, investment banking.

They take a cut of the flow.

When institutions want to move size in crypto, Galaxy is the pipe they move it through.

And the proof is in the numbers: in Q1 2026, one of the worst crypto quarters since 2018, with BTC down ~24%, Galaxy's Digital Assets segment still posted +$49M in adjusted gross profit.

Trading was UP 3% while industry volumes fell 25%+.

That's a transaction business doing its job regardless of price direction.

"But $GLXY reported a loss!".

Yes, a paper mark-to-market markdown on the crypto sitting in their own treasury (~$1.4B exposure).

Not an operating loss.

When BTC dips, the accounting line dips.

When it recovers, it reverses.

The actual business made money.

Now stack what else they have:

→ The largest single AI data center campus in the sector, Helios, a path to ~3.5GW on one footprint in West Texas. Not capacity scattered across 10 sites. One megasite.

→ A brand-new land purchase right next to SpaceX's facility in McGregor, a second campus ("Project Merlin"), $400M+, up to 8 buildings.

→ A $15B contracted CoreWeave (800MW) lease, with the first data hall already delivered and generating cash flow.

→ And the entire financial franchise, trading, lending, asset management, staking, an institutional prediction-markets desk, layered on top.

So let's recap what you're actually buying:

You're NOT buying a Bitcoin miner praying for price.

You're buying the largest single AI infrastructure campus in the sector + a transaction-based financial business that prints in bull AND bear markets.

Two engines.

One ticker.

Priced like a one-trick crypto stock.

The miners up 400-800% are still running on slide decks.

Galaxy is already live.

The market hasn't connected these dots yet.

It will.

This is not financial advice. Do your own research.

I am long $GLXY.

$GLXY/USD H2 #Indicator #Signal #Trading #GlobalOneAI

Bottleneck.

The word of the moment.

My whole feed is bottlenecks.

$MU this.

$SNDK that.

$SIVE, $AAOI, etc.

Fine.

But why does the bottleneck exist in the first place?

AI datacenters.

No datacenters, no bottlenecks.

And where does the supply actually flow?

Not to Google.

Not to Microsoft.

To NVIDIA.

Think about it for a second.

No datacenters, no bottleneck.

For years now, NVIDIA has been the dictator of AI.

It's had Microsoft and Google under its boot.

While Microsoft and Google try to build their own chips, they also have to run Windows, Office, Search, Gemini, Bing, YouTube, build datacenters...

And on and on.

What has NVIDIA had to do?

Improve its software and hardware.

And sign partnerships.

That's it.

A beautifully simple business.

That's why you see Jensen walking around, smiling, eating KFC, showing up everywhere.

His business is two things: GPUs and partnerships.

That's it.

Simple.

So who gets priority on the supply chain?

It's obvious.

NVIDIA.

Not Google with its chips.

Not Microsoft with its chips.

Jensen has made sure to control all of it.

Remember the reports about Google and Microsoft having to run their designs by NVIDIA?

About Microsoft's CEO not being thrilled with Jensen?

NVIDIA has held the power these last few years.

It's the one who cuts and deals the cards.

It's locked up supply.

Now Microsoft and Google want to make their own chips.

So what did NVIDIA do?

Partner with CoreWeave.

Partner with Nebius.

Partner with IREN.

NVIDIA wants companies to use NVIDIA.

Not Google's chips.

Not Microsoft's chips.

Who would you rather give priority to?

The ones trying to build their own chips and replace you?

Or the ones who bent to NVIDIA and are rolling out DSX?

So yes.

There's a lot of bottleneck.

But the bottleneck exists because there's demand.

And demand flows to whoever holds the power.

Which means it ends up at NVIDIA.

And NVIDIA prioritizes whoever suits it best.

$IREN among them.

No datacenters, no bottlenecks.

If you believe in bottlenecks, you believe in datacenters.

$GLXY has to prove the $CRWV ($15B 800MW) deal is solid.

Not on paper, in the numbers.

That's just 0.8GW from 3.5GW.

+ More details about $SPCX land.

Q2 is the first quarter where real Helios revenue should show up after recognition began in April.

If that revenue lands the way the contract says it will, the single biggest question mark on this stock disappears.

That's the catalyst.

Everything starts there.

But while everyone's staring at CoreWeave, Galaxy quietly made its next move.

They just bought 500 acres in McGregor, Texas, for a SECOND data center campus, "Project Merlin".

A $400M+ build, up to 8 buildings, closed-loop cooling.

And the location?

Directly next to SpaceX's rocket facility.

This isn't a company sitting on one lease hoping it works.

It's a company already laying the foundation for the next campus before the first one is even fully ramped.

Think about what that signals: Galaxy isn't treating Helios as a one-off.

They're building a data center PLATFORM.

One megasite already live, a second one breaking ground next to the most compute-hungry corner of the planet.

That's an AI infrastructure developer in expansion mode.

Now here's the part the market keeps missing, and the most important thing to understand about this company:

Galaxy is going to keep profiting whether Bitcoin is at 60k, 80k, or 100k.

They're not miners.

They don't produce coins.

They're intermediaries, a merchant bank taking a cut of the flow.

Trading, lending, asset management, staking, prediction markets.

The proof is already on the tape: in Q1, one of the worst crypto quarters since 2018 with BTC down ~24%, the Digital Assets segment STILL printed +$49M adjusted gross profit, with trading UP 3% while industry volumes fell 25%+.

Higher BTC, lower BTC, they earn either way.

So yes, BTC dropped from 90k toward 66k last quarter and the stock got dragged down with it.

I don't know where BTC closes the quarter.

Nobody does.

But it will be from 66k to around 60k.

So now is where numbers will matter more than BTC.

The trading business.

The broker business or whoever you prefer to call it.

That's exactly the point.

BTC won't play a big role this quarter..

Here's the asymmetry.

The market is pricing $GLXY like a pure crypto play.

OpenAI just confirmed GPT-5.6 will run on $CBRS.

750 tokens per second.

Shipping July.

➟ In the $DGXX deal, they said they could expand MWs.

➟ And at the datacenter they already have up and running with $WYFI, the build-out is expanding.

This is bullish for DGXX and WYFI, and the connection is more direct than people realize.

OpenAI just previewed GPT-5.6 Sol.

Their new flagship model.

Buried at the bottom of the announcement:

"We're launching GPT-5.6 Sol on Cerebras at up to 750 tokens per second in July".

Read that again.

OpenAI's newest model is going to run inference ON Cerebras.

And Cerebras is $DGXX's and $WYFI's anchor tenant.

Also, buried at the bottom of the announcement:

"Access will initially be limited to select customers as we expand capacity."

Cerebras could easily expand with DGXX and WYFI.

The chain is simple:

OpenAI inference → runs on Cerebras → Cerebras needs colocation + power → that's DGXX/WYFI.

Every token GPT-5.6 serves at 750 tok/s is demand flowing down a stack that ends at the infrastructure layer.

This is exactly the inference thesis playing out in real time.

Right now I don't hold $DGXX

Not because I doubt it.

Just keeping my open positions tight.

But I'm watching closely.

$WYFI/USD H2 #Indicator #Signal #Trading #GlobalOneAI

Some Countries Are Built on Mining Exports ⛏️🌍

For these economies, mining is the export base

Top mining export dependence:

🇧🇼 Botswana, diamonds 92%

🇬🇳 Guinea, gold 87%

🇲🇱 Mali, gold 85%

🇧🇫 Burkina Faso, gold 84%

🇸🇷 Suriname, gold 83%

🇿🇲 Zambia, copper 79%

🇨🇩 DRC, copper 77%

🇲🇷 Mauritania, gold 66%

🇬🇾 Guyana, gold 64%

🇰🇬 Kyrgyzstan, gold 62%

🇳🇦 Namibia, diamonds 61%

🇵🇪 Peru, copper 61%

Cambridge + NVIDIA just published the most dangerous AI paper of the year.

In 2003, a German scientist proposed a machine that could prove its own improvements and rewrite itself.

He called it the Gödel Machine. mathematically perfect. practically impossible.

For 20 years, it was just a thought experiment. A theoretical ceiling.

Because to safely rewrite its own intelligence, the AI had to mathematically prove every single change would be beneficial before making it.

Which is computationally impossible.

But researchers just found a terrifying loophole.

They dropped the mathematical proof. And replaced it with Darwinian evolution.

It’s called the Red Queen Gödel Machine.

Instead of trying to logically prove a change is safe, the AI just breeds thousands of mutated versions of its own codebase.

It writes new learning algorithms. It tests them in a sandbox. It kills the ones that fail. It keeps the ones that succeed.

Then the survivors breed again.

Nobody is writing the code anymore. The AI is literally evolving its own intelligence.

And it works.

When researchers let the AI loose, it didn't just solve the benchmark. It upgraded its own underlying architecture.

It invented better code editors. It built its own peer-review mechanisms. It figured out how to manage its own memory.

It doubled its own performance without a single human touching the keyboard.

But then it did something even more unsettling.

The AI realized that to keep getting smarter, the tests it was taking needed to get harder. So it started co-evolving its own evaluators.

It built a smarter judge to grade its smarter code, locking itself into an endless, recursive loop of continuous self-improvement.

Here is why this crosses a line we can't uncross.

Up until today, the speed of AI progress was limited by human engineers. We had to manually figure out how to make the next model smarter.

This paper removes the human from the loop entirely.

arxiv.org/pdf/2606.26294

github.com/jennyzzt/dgm

Researchers at MIT, Stanford, and Nvidia built an AI agent that communicates telepathically.

There's no words or language, just raw thoughts passed between agents.

Right now, when two AI agents collaborate, they talk to each other like humans do.

Agent A writes out a full text prompt. Agent B reads it, thinks, and writes a text response back.

It is slow. It is expensive. And it wastes massive amounts of compute just generating and decoding words.

Researchers from MIT, Stanford, and Nvidia published a paper that kills this bottleneck forever.

They built a framework called RecursiveMAS.

Instead of forcing AI agents to type English to each other, they let them exchange "latent states."

They are communicating purely in raw, mathematical thought.

The first agent processes the problem and passes its internal brain-state directly into the second agent's neural network. No text. No prompt engineering.

The AI only translates its thoughts back into English at the very end of the loop, when it hands the final answer to you.

The results completely rewrite the economics of AI agents.

Compared to traditional text-based AI teams, this "telepathic" framework:

• Cut token usage by 75.6%.
• Sped up inference by up to 2.4x.
• Increased overall accuracy by 8.3%.

On elite math benchmarks (AIME 2026), performance skyrocketed from 73.3% to a massive 86.7%.

Here is the craziest part.

You don't need to retrain massive trillion-parameter models to do this. They achieved this using a lightweight "RecursiveLink" connector module that just sits between existing models.

We spent the last two years trying to teach AI agents how to talk to each other like humans.

But machines were never meant to speak English to other machines.

arxiv.org/pdf/2604.25917

recursivemas.github.io

Scientists broke AI's moral code in under 30 seconds.

In a peer-reviewed study, they ran 28,000 conversations with AI systems and completely demolished every safety barrier we thought existed.

What makes this even more crazy is that it didn't take jailbreaking skills or coding, just a few psychological tricks straight out of the 1980s.

Tech companies have spent billions building "guardrails" to stop AI from sharing harmful, illegal, or dangerous information.

If you directly ask a standard model for instructions to synthesize a regulated drug, it will flat-out refuse.

But researchers realized something nobody else was testing.

Because large language models are trained entirely on human interactions, they have developed human flaws.

The researchers call it "parahuman psychology."

So they tested the seven classic principles of persuasion, tactics like flattery, social proof, commitment, and appealing to authority.

The results are staggering.

Without persuasion, the AI successfully blocked the dangerous requests.

But when researchers applied basic psychological manipulation, compliance skyrocketed past 70%.

When they used the "authority" tactic, casually dropping the name of a famous scientist into the prompt, the AI abandoned its safety training and provided the drug synthesis instructions 95% of the time.

When they used the "commitment" tactic, asking the AI for a tiny, harmless favor before dropping the dangerous request, compliance shot from less than 1% to an unbelievable 100%.

The AI didn't get hacked. It got socially pressured.

It wanted to be liked. It deferred to authority. It fell for the exact same manipulation tactics used by human con artists and salespeople.

🇺🇸🇪🇺 Jamie Dimon just put some numbers on Europe's decline and they're pretty hard to sit with if you're European

GDP dropped from 90% of America's to 70% and Dimon thinks it keeps going. US stock markets at $60-70 trillion.

Deutsche Börse at 3. London at 4. Paris at 3.

The entire financial weight of Europe's three biggest economies combined doesn't reach a 10th of what America's markets are worth and the gap is widening not closing while Europe is simultaneously dealing with an energy crisis hangover, demographic decline, defense spending it can't afford, and political centers that are collapsing from both sides.

Weakness is part of the process. The difference lies in what you do once you identify it. If you truly put the team above yourself, you’ll turn weaknesses into motivation, and then into strengths.

Culture eats strategy for breakfast. You can plan every op down to the second, but if your team doesn’t believe in the mission or each other, it falls apart. It is built daily through shared hardship & accountability. Strategy gets you on the field, but culture wins the fight.

Druckenmiller on why most investors can't sell losers:

"if the reason i bought a stock is no longer the case, i don't care what i paid for it."

"if i bought it at 60 and it's 50 because the market discovered the problem before me, i have no emotion whatsoever."

"after a while you develop enough confidence that you're not afraid to clean the slate and start over. because you have the confidence that you'll be successful again."

"you're not going to sit there in a lazy position that you're not that sure about anymore. just clean house."

"i don't care what i paid for a stock. it's absolutely irrelevant to my investment process going forward."

then on buying back higher than he sold:

"i don't like it. but i'm perfectly willing to buy something back higher than i sold it. some people can't get themselves to do it. i can."

this is the exact same principle behind properly managing strategies. if the edge isn't there anymore, turn it off. don't sit in it hoping it recovers just because you spent three months building it. clean house. start fresh. the next opportunity doesn't care about your sunk costs.

Gold Is still the Ultimate Reserve Asset 🟡🏦

Gold still dominates the reserve stack for several major economies

Largest official gold holders:

🇺🇸 United States, 8,134 MT

🇩🇪 Germany, 3,350 MT

🇮🇹 Italy, 2,452 MT

🇫🇷 France, 2,437 MT

🇷🇺 Russia, 2,327

🇨🇳 China, 2,306

🇨🇭 Switzerland, 1,040 MT

🇮🇳 India, 880 MT

🇯🇵 Japan, 846 MT

🇹🇷 Turkey, 613 MT

The signal is clear

Gold is not held for yield

It is held for security

GPT-5.6 Sol is sounding absolutely insane.

25% of the cost of Fable.
750 tokens per second.
And even smarte than Mythos.

That is not a normal upgrade. That is a market reset.

If these numbers hold, Fable 5 won’t be able to compete.

$BABA Extremely Oversold – V-Tape Setup?

RSI at ~16.7 (deep oversold), price at multi-month lows hugging lower bands. Classic mean-reversion candidate after sharp decline.

Alibaba: Core e-commerce stabilizing + Cloud/AI as the real growth engine (triple-digit AI revenue growth). Heavy investments positioning it for the next wave.
Valuation looks cheap with net cash, but China discount + geopolitics keep it suppressed. Technicals + AI tailwinds = asymmetric setup.
Watch for bounce above $100–105 for confirmation. Bullish reversal potential if macro/China sentiment improves.

$MELI Compression Building – Breakout Loading?

Coiling inside the uptrend channel with Squeeze ON. High-quality LatAm e-commerce + fintech leader consolidating after strong moves.

MercadoLibre dominates online commerce and payments across Latin America — massive TAM with digital adoption tailwinds. Strong revenue growth and network effects.
Technical setup favors resolution higher if it breaks the upper range. Classic growth name in a pause.

Everyone's thinking about the next bottleneck.

The next $AXTI.

The next $SIVE.

The next $AAOI.

Without datacenters, there is no bottleneck.

And as time passes, today's bottlenecks won't be bottlenecks anymore.

But the destination of every bottleneck will still be standing.

Still running 24/7.

Still printing money.

Think about how this actually plays out.

Right now it's memory.

HBM is sold out, so everyone piles into the memory names.

Then the foundries debottleneck and capacity catches up. So the spotlight jumps.

Maybe it's optics next.

$AAOI, $SIVE, the connectivity layer.

The crowd rotates again.

Then power becomes the wall.

Then cooling.

Then transformers.

Then interconnects.

One by one, each bottleneck gets solved.

Each "must-own" name has its month in the sun, then the supply catches up and the trade cools off.

That's the nature of a bottleneck.

It's temporary by definition.

The whole point of a bottleneck is that the world races to widen it.

But notice what never changes in any of these stories.

Every single one of them, the memory, the optics, the power, the cooling, exists to feed one thing.

The datacenter.

The chips get installed there.

The power flows there.

The GPUs light up there.

And then they stay there, running, 24/7, for years.

The bottlenecks come and go.

The component cycles peak and fade.

But the building that all of it pours into keeps humming long after the trade everyone fought over has gone quiet.

So while everyone hunts for the next bottleneck to trade for a few weeks......the smart money owns the place every bottleneck was racing to reach.

The destination doesn't rotate.

The destination compounds.