Numbskully

[Numbskully @NUMBHTF]

@Chef_PewPew LaFerrari, only $3-8 million, chump change for one of the greatest traders I have had the privilege of knowing. 🫶

[Dana Perino @DanaPerino]

The WWII vets featured in the President’s speech watched the fireworks from their memorial. ❤️🇺🇸🙏🏻🎆

[Numbskully @NUMBHTF]

WATCH THIS!!!
You Won’t BELIEVE What Just Happened in CUBA!!! https://t.co/QKHU8zlpng via @YouTube

[Ena Gill @JaiShreeRam90]

@catturd2 Rather than circulating the disturbing video of Charlie Kirk being shot, I encourage you to share photos of his loving family, now living their worst nightmare.

Every child deserves a father.
Every wife deserves a husband.
And America needs Charlie Kirk.

💔

[Chief Steven Sund @ChiefSund]

Ma’am, it is long past time to be honest with the American people. On January 3, I requested National Guard assistance, but your Sergeant at Arms denied it. Under federal law (2 U.S.C. §1970), I was prohibited from calling them in without specific approval. That same day, Carol Corbin at the Pentagon offered National Guard support, but I was forced to decline because I lacked the legal authority.

On January 6, while the Capitol was under attack and despite my repeated calls, your Sergeant at Arms again denied my urgent requests for over 70 agonizing minutes, “running it up the chain” for your approval.

When I needed assistance, it was denied. Yet when it suited you, you ordered fencing topped with concertina wire and surrounded the Capitol with thousands of armed National Guard troops.

[Johnny Midnight ⚡️ @its_The_Dr]

He is absolutely 💯 correct! He says let them keep on doing what they are doing in California!

[Shay Boloor @StockSavvyShay]

THIS IS THE REAL STORY HEADING INTO $NVDA EARNINGS NEXT WEEK

The most powerful companies don’t just build products -- they build dependencies. And this earnings report isn’t just about what the company sold. It’s about how deeply it’s embedded in the modern enterprise operating system.

CUDA isn’t an add-on. It’s not the feature that wins deals. It’s the dependency that makes deals inevitable. For nearly two decades, NVIDIA has been expanding CUDA from a GPU programming language into the foundation layer of the accelerated computing economy. The performance delta is no longer the reason enterprises adopt NVIDIA. It’s the ecosystem lock-in. Once a Fortune 500 builds a model in CUDA, that model is optimized for NVIDIA hardware, dependent on NVIDIA drivers, and scalable only through NVIDIA-compatible architecture.

And unlike traditional enterprise software moats, CUDA isn’t bloated. It abstracts away the complexity of massively parallel workloads, so researchers don’t have to understand GPU architecture to use it -- and that convenience at scale is everything. Across scientific computing, national defense, bioinformatics, AV simulation, and AI training, CUDA is the interface. It is the assumption. And because no serious challenger has matched its developer tooling or performance tuning across generations, CUDA is also the trap door. You don’t switch out of it. You build deeper into it.

That’s the part Wall Street keeps underpricing. This is no longer about selling chips into demand surges. It’s about selling access to the infrastructure that everything intelligent runs on. CUDA is the monopoly operating layer for accelerated compute -- and the frictionless glue for the world’s largest and most critical workloads.

But compute alone doesn’t scale AI. Data has to move. Memory has to sync. And GPUs have to act as one. That’s where NVLink enters -- not as a bus, but as NVIDIA’s second moat. CUDA controls the code. NVLink controls the flow.

For years, PCIe was the bottleneck for distributed AI training. NVLink changed that. And now, NVLink Fusion -- NVIDIA’s next evolution in interconnect -- has redefined what’s possible. At Computex, NVIDIA opened its architecture to allow hyperscalers to plug their custom chips directly into NVLink-connected systems. That wasn’t a gesture of openness. It was a power move. Because even if $AMZN uses Trainium or $GOOGL leans into TPUs, they still need NVLink to operate at hyperscale speed and efficiency.

This is the shift the market is still catching up to. Hyperscalers didn’t escape the NVIDIA tax by building their own silicon. They just changed where they pay it -- from GPU purchases to fabric dependence. And that dependence is now baked into roadmaps at $MSFT, $META, $ORCL and even sovereign AI buildouts. You don’t have to buy NVIDIA’s GPU to be in its orbit. You just have to need scale.

NVLink isn’t just interconnect. It’s an API for physical infrastructure. It defines how components talk to each other in-memory. It determines latency at the rack level. It decides how quickly a model can go from training to inference. And just like CUDA, NVLink isn’t modular. It’s systemic.

That’s why NVIDIA’s earnings aren’t just a look at market share. They’re a signal of who’s still inside the system. $43B in revenue, up 65% YoY after doubling last year, isn’t normal growth. It’s compounding dominance. It’s the compounding of dependence.

And when Jensen pulled Rubin forward at Computex -- positioning it not just as the next chip but as a full-stack replacement for entire AI clusters -- the message was unmistakable: NVIDIA is not playing catch-up to demand. It’s front-running it. Rubin systems aren’t being built for next-gen use cases. They’re being built because the current infrastructure is already saturated.

Hyperscalers are out of headroom. $CRWV, $NBIS, and the sovereign buildouts in Saudi Arabia and India all point to the same truth: this is not a cyclical peak. This is a bandwidth shortage. AI infrastructure is now a constrained commodity -- and NVIDIA controls the gate.

So when you watch this earnings report, the gross margin delta from H20 restrictions isn’t the story. The story is whether Blackwell pricing holds. Whether Rubin is shipping early. Whether CUDA penetration continues to spread through enterprise AI workloads like Kubernetes did through DevOps. Whether NVLink becomes mandatory infrastructure for hyperscale growth.

Because if the answer to all of that is yes -- and signs point that way -- then this isn’t a chip company at all.

It’s the base layer of the Fourth Industrial Revolution.

And in a world that will be priced by intelligence, measured by compute, and scaled by memory coherence, NVIDIA is no longer a vendor. It’s the referee. The compiler. The backbone.

So when you see the stock print move -- don’t just ask if revenue beat. Ask what can’t be built without NVIDIA anymore.

Because CUDA is the language. NVLink is the bridge. And together, they’re not just defending a moat. They’ve become the map.

[Erica Knight @_EricaKnight]

🔥NEXT WEEK 🔥

[Numbskully @NUMBHTF]

I'm so happy that President Obama was able to set next to "A" President today! 🫳🎤

[FAN TRUMP ARMY @TRUMP_ARMY_]

@elonmusk Absolutely

[Ernesto Méndez @ernesto_mendez]

Viaje muy movido hacia el huracán #Milton a bordo del @NOAA WP-3D Orion #NOAA43 "Miss Piggy" para recopilar datos que ayuden a mejorar el pronóstico y apoyar la investigación sobre huracanes.
🎥 Crédito: Nick Underwood, NOAA

[Numbskully @NUMBHTF]

@flabbergank THE Best indicator I have ever used for futures, hands down. And the group of people in https://t.co/aTBxWgmrui are amazing.

[Flabbergank @flabbergank]

Tired of trying to trade the charts on #futures and getting rekt? I built Market Copilot to help flip the script. Copilot digests price action and order flow data to give you immediate, actionable insights for high probability setups. Now only 45$/mo for a limited time.

[Flabbergank @flabbergank]

New feature I'm pretty hyped for on Copilot #futures - order flow clustering! I built a model that clusters order flow behavior into categories for greater insight into market imbalance.

[AAPL Bottom Jeans 🍎🩲👖 @aaplbottomjeans]

Mag 7 + $COIN majority bullish with the exception of $GOOGL. Big orders of ITM puts came through for this stock... interesting... Wellp, let us see how we navigate OpEx! See you all tomorrow, @quantdata team!

[Jack Posobiec @JackPosobiec]

Omg

[Flabbergank @flabbergank]

Daily $NQ #futures trade using copilot @scourge_bot !

CPOC - candle point of control - showed three consecutive candles where volume was in the top wick before price retraced. Bearish.

CVD increased while price didn't - indicates passive seller.

Entered short - 230$ and done.

[Flabbergank @flabbergank]

Daily #NQ #futures trade with Copilot @scourge_bot

I spot the CVD diverging away from price action. This formation indicates the presence of a passive seller.

[Numbskully @NUMBHTF]

@flabbergank Thank you for bringing @scourge_bot to @htf666 What a slayer of an indicator. Live futures data makes for a very profitable day. Much love my friend. Join in on the excitement @ https://t.co/KGLhb1HMIy

[Flabbergank @flabbergank]

Hey everyone!

As you know I've been dabbling in futures trading and have come to find my personal edge in it. With a bit of experience under my belt, I'm happy to share that now #HellsTradingFloor and Market Copilot have access to futures data!