Most people don't realize what just happened to the entire AI trade.
Jensen Huang just announced that Artificial General Intelligence has officially arrived.
When the CEO of a $5.58 trillion company makes that declaration, the market-wide capital flows shift immediately.
Here is the financial and market structure breakdown of what happens next:
1. The Revenue Runway
Nvidia guided Q3 revenue to $108 billion (+12.3% quarter-over-quarter) after posting $96.2 billion in Q2. That puts their annualized run-rate at $432 billion, and that number assumes zero China data center revenue. At a $5.58 trillion market cap, $NVDA trades at roughly 12.9x forward sales.
2. The Capex Acceleration
Hyperscaler 2026 capital expenditure guidance revisions increased by another 22% through the latest earnings calls. Capex growth across major cloud providers is expanding faster than chip delivery guidance.
3. The Downstream Rotation
Friday's tape revealed where institutional capital is actually positioning for this next phase. Nvidia reached $234.76, sitting 98 cents below its $235.74 all-time high, but volume contracted to 128.7 million shares compared to 298.9 million on August 27. Instead of chasing the bellwether at resistance, liquidity rotated downstream into the physical stack: memory surged 6.1%, networking gained 7.1%, optical hardware rose 4.7%, and power infrastructure jumped 4.4%.
4. The Physical Bottleneck
Declaring that AGI is here means compute demand has outgrown pure silicon supply. The constraint shifts from chip design to utility power interconnection, thermal cooling capacity, and high-bandwidth memory.
If compute requirements scale at this rate, which infrastructure layer hits a supply deficit first: power generation or high-bandwidth memory?
SpaceX $SPCX is applying its rocket manufacturing playbook to break the single biggest bottleneck in the entire gas turbine industry.
Hot section blades and vanes in large industrial gas turbines must operate at temperatures far above the melting point of the alloys themselves (roughly 1,650–1,980 °C). That requires precision casting of single crystal (SX) or directionally solidified (DS) nickel-based superalloys. Only about three foundries in the world (PCC, Howmet and a couple of others) can reliably produce these parts at scale, and their order books are already full through 2030. This has become one of the most severe constraints on powering the AI data-center build-out.
SpaceX’s advantages are:
• It already performs similar high-temperature superalloy casting for Raptor turbopumps (and spent years solving blade-cracking problems).
• It is now building its own blades-and-vanes foundry in Bastrop, Texas, so it can apply that capability directly to industrial gas turbines.
• Elon says in house casting can bring complete turbines online up to 18 months sooner—a genuine game-changer given current lead times.
A more precise way to describe it is this:
SpaceX is not immediately reinventing the gas turbine. It is injecting aerospace-style rapid iteration, vertical integration and extreme manufacturing discipline into a conservative industry that has long been dominated by a handful of suppliers and plagued by multi-year deliveries. If they succeed, the usual outcomes follow: more capacity, shorter lead times, lower cost, and pressure on the incumbent foundries and OEMs to move faster.
This is classic SpaceX: first solve its own urgent needs ( SpaceXAI data centers + Starship ) and, in the process, raise the ceiling of the entire supply chain. In the near term the work is mostly for internal use; over the longer term there is a real chance the benefits spill over to the broader energy industry.
This Leopold / Situational Awareness story is crazy.
> Go from $1.5B to $20B in under two years.
> Over 1,000% since launch, 270% ytd to May alone.
> Jane Street which almost never allocates to external managers becomes an investor.
> $24B near the highs, levered as much as 4x. Up 439% for H1'26.
> July 2026: AI names go down as much as 50%.
> Compounded with up to 4x margin...
> Leopold sends letter to LPs for additional capital reported by FT yesterday, says that this is a buying opp.
> Today: CNBC reports Situational Awareness has exited all public equity positions.
> One buyer takes everything, both L/S books, in one enormous block.
> Jane Street?? Who buys a whole levered spread book in one go? Isn't that is literally Jane Street's business?
> As an LP, Jane Street had the inside view of exactly his book. Which would make this even more ironic...
> If it was JS, the arc completes: Leopold's rarest backer would be the buyer of his whole book lol.
> What's left of Leopold? Whatever's left in the private book after selling Anthropic stake and the capital raise, if any.
I can not believe how bad all the X takes about Leopold are?!
1. He's a billionaire.
2. After getting liquidated, he's still up more than 99% of other funds.
3. His investors, obviously, are very happy with him.
4. He could start a new, maybe bigger, fund in no time.
5. It seems so incredibly obvious there was massive manipulation behind the scenes to push him over the edge via misinformation, hidden as an "opinion".
6. He very accurately and eloquently explained what would happen and nailed most of what happened years in advance.
There's no world in which he's an idiot, retard or looser.
What is X smoking?
You all seem like such haters sometimes. Why does it make you so happy he's having a shit time rn? I love you all, but why does seeing others fail excite people so much more than winning together? Also the winner here is prolly one of the WORST players in this game...
Only @ParadisLabs nailed what is happening. Well done. Factual and accurate interpretation imho! 👌🏻
Global Crude Inventories Forecast by End of August 2026
The MoU collapsed. Iranian missiles hit Kuwait 48 hours ago. Two US service members killed in Jordan. One missing.
Here’s what markets aren’t pricing: we have exactly 42 days until end of August. Global crude inventories are reaching levels that will redefine “critical.”
Let me show you the math.
The recent narrative has been all about the semiconductor rally and the MAG-7 pullback.
But when you zoom out, a different picture emerges.
Semiconductors $SOX have only recently begun outperforming the MAG-7 $MAGS since ChatGPT launched in November 2022.
This really puts things into perspective: