H100 GPU rental prices just surged 40% in 6 months.

From $1.70/hour to $2.35/hour. Capacity is effectively sold out across the board. Blackwell lead times are stretching into mid-2026.

Here’s what that means in plain English:

The companies that ALREADY SECURED power, land, and grid connections for AI compute are sitting on scarce infrastructure.

$IREN. $NBIS. $CIFR.

The four largest hyperscalers plan to spend $700 billion on AI infrastructure in 2026.

That’s not a forecast. That’s disclosed capital.

GPU prices going up is the neocloud business model working exactly as designed.

Short contracts. Rising rates. Locked-in capacity.

Still holding all three.

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

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

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

$META Meta is about to release its first AI model built under a man who was paid $14.3 billion to fix the company's biggest failure.

The failure was Llama 4 and Meta spent years telling the world that open-source AI was the future.

Zuckerberg said it repeatedly on earnings calls, in interviews, in blog posts then the model dropped and almost nobody cared.

Chinese AI companies took Meta's open-source code, used it to train their own models, and made Meta look slow in its own race.

So Zuckerberg blew up the entire strategy and he paid $14.3 billion for a 49% stake in Scale AI and brought its 28-year-old founder, Alexandr Wang, in as Chief AI Officer to rebuild from scratch.

Wang is now leading a lab inside Meta working on two models simultaneously, one for text, one for images and video.

The text model is codenamed Avocado and it was supposed to launch in March but it didn't.

Internal tests showed it performing worse than Google's Gemini 3 and the release got pushed to May, minimum.

Now comes today's news.

According to Axios, Meta is preparing to release the first model built under Wang, very soon.

And there is a new wrinkle, the company plans to eventually open-source versions of it, but not at launch.

This matters more than it sounds because Meta built its entire developer reputation on open source.

Llama was free for everyone, researchers built on it, startups ran on it but that is over now.

The biggest models in the new family will stay proprietary.

That is the hybrid strategy Wang is steering, some models closed, some eventually released, but only after Meta captures whatever advantage it can first.

Meta knows it cannot beat OpenAI or Anthropic head to head right now so it is not trying to.

Instead it is betting on specific consumer use cases, its platforms have more than 3 billion users and hoping that scale compensates for where the model falls short.

The company has committed between $115 billion and $135 billion in capital expenditure this year alone, chasing a future it is not yet sure it can reach.

If Avocado launches and impresses, it validates the entire bet, the $14.3 billion, the layoffs, the cultural overhaul, the open-source reversal.

If it doesn't, Meta will have spent more money than most countries' GDP to produce a model that still trails Google and OpenAI, with no open-source goodwill left to fall back on.

🚨 "We’re building one of the world’s largest AI cloud platforms."

$IREN 2026: 4.5GW
$NBIS 2030 Target: 5GW

The potential is beyond imagination for those who read the big picture. 👇

"The next cloud wave on the ground behaves more like an industrial buildout, where the scarce input is not GPU or code or even money, but (1) synchronized access to land, (2) substations, (3) transformers, (4) transmission capacity, and (5) utility approvals."

The gap between those who started earlier and those who followed the trend is staggering.

$IREN is expected this month to announce energizing one of the largest data centers on the planet: SW1 (1.4GW).

Based on the CEO and CCO most recent interviews, the deadline is unchanged.

“The entire country can be taken out in one night, and that night might be tomorrow night.” 🤣

-Donald Trump

🇮🇷 MEANWHILE…

Iranian’s IRGC FM Araghchi:

“We do not want a ceasefire, not because of the war.”

“This time, this battle must be completed so that our enemies never even think of repeating these attacks and incursions again.”

“And I believe they have learned a good lesson so far.”

“Take this battle to the furthest extent necessary.”

What ‘lesson’ do you think he’s learned?

Space stocks continuing to push higher as we get closer to the SpaceX IPO and receive more news.

NASA’s Artemis II mission also giving the entire sector more exposure & more eyes.

$SATL $ASTS $RKLB $SIDU $FLY

$IREN has Dell purchase agreements in place for 150k GPUs at fixed pricing. $5.8B for the Microsoft deployment and $3.5B for the B300 expansion. Payments are on delivery. The supply is secured.

- Iran didn't control the Strait of Hormuz, now it does
- Iran's oil was sanctioned, now it isn't
- Iran was not building a nuclear weapon, now it will
- US bases in the Gulf were assets, now they're liabilities
- Inflation was declining, now it's rising
- No more wars became WW3

🚨 JAPAN TO INVEST UP TO $40B IN U.S. NUCLEAR

Japan plans to fund small modular reactors across Tennessee and Alabama through GE Vernova Hitachi Energy.

Power is now the biggest bottleneck for data centers, with grid constraints slowing new AI capacity.

SMRs could be a key solution as AI demand explodes.

Peter Beck’s annual letter makes it clear that $RKLB is evolving into a broader space infrastructure platform with growing backlog, deeper defense exposure & multiple growth engines.

Rocket Lab is entering 2026 with real momentum behind Neutron, deeper national security relevance & a stronger vertically integrated stack after the Mynaric deal.

$AVGO signed a long-term deal with $GOOGL to develop future TPUs and supply networking and other components for Google’s next-gen AI racks through 2031.

Broadcom also said Anthropic will gain access to about 3.5 GW of TPU compute starting in 2027.

$UNH/USDT H2 #Indicator #Signal #Trading #GlobalOneAI

US company XDOWN has developed the STUD: 44.5 cm long, 2.7 kg MTOW, 0.77 kg payload, 305 km/h, 64 km range, 17 min flight. One soldier can reportedly carry 8–12 in a backpack. Modular design reportedly supports ISR, EW, recon, counter-drone tasks.

#UAV #Drones #STUD #XDOWN

🚨There are 4 human beings in space flying around and observing the far side of the moon right now