AI Colocation Build Costs Are Rising Fast — Here's Where the Market Sits 🏗️
⚡ Global average shell & core: $11.3M per MW in 2026 — up from $7.7M in 2020 (JLL)
⚡ AI-optimized facilities: $20M+ per MW — liquid cooling, high-density power distribution adds a significant premium
⚡ Construction cost inflation running at 5–15% annually — 60% of industry professionals expect further increases in 2026
⚡ Equipment lead times now averaging 33 weeks — a 50% increase from pre-2020 levels
⚡ Grid connection queues in primary US markets exceed 3–5 years
⚡ 77% of the global construction pipeline is already pre-committed to tenants before breaking ground
Power availability has replaced capital as the #1 constraint on new AI data center development... what does this mean for operators who already hold energized land? $RIOT $WULF $HUT $IREN $CORZ $WYFI $HIVE $CIFR $KEEL
⚡ Global average shell & core: $11.3M per MW in 2026 — up from $7.7M in 2020 (JLL)
⚡ AI-optimized facilities: $20M+ per MW — liquid cooling, high-density power distribution adds a significant premium
⚡ Construction cost inflation running at 5–15% annually — 60% of industry professionals expect further increases in 2026
⚡ Equipment lead times now averaging 33 weeks — a 50% increase from pre-2020 levels
⚡ Grid connection queues in primary US markets exceed 3–5 years
⚡ 77% of the global construction pipeline is already pre-committed to tenants before breaking ground
Power availability has replaced capital as the #1 constraint on new AI data center development... what does this mean for operators who already hold energized land? $RIOT $WULF $HUT $IREN $CORZ $WYFI $HIVE $CIFR $KEEL
✍1
A ceasefire agreement could be reached tonight. The negotiations are being coordinated by the Commander-in-Chief of the Pakistani Army – CNN with a link to a Pakistani source.
UPD: Iran is inclined to agree to a two-week ceasefire out of respect for the Pakistani mediator – Al Araby
Please open Telegram to view this post
VIEW IN TELEGRAM
👀1
River Bend — The Deal That Changes Everything $HUT #Hut8Corp
⚡ 15-year, 245 MW IT lease with #fluidstack — $7.0B in base contract value, expandable to $17.7B if all renewal options are exercised
⚡ Lease payments financially backstopped by #Google for the full 15-year base term — investment-grade credit support on a greenfield build
⚡ Project financing at 85% LTC — led by J.P. Morgan as loan underwriter with Goldman Sachs — institutional-grade capital structure from day one
⚡ Delivery begins Q2 2027 — 2026 is about execution, financing, and converting pipeline into contracted revenue
⚡ 15-year, 245 MW IT lease with #fluidstack — $7.0B in base contract value, expandable to $17.7B if all renewal options are exercised
⚡ Lease payments financially backstopped by #Google for the full 15-year base term — investment-grade credit support on a greenfield build
⚡ Project financing at 85% LTC — led by J.P. Morgan as loan underwriter with Goldman Sachs — institutional-grade capital structure from day one
⚡ Delivery begins Q2 2027 — 2026 is about execution, financing, and converting pipeline into contracted revenue
🇨🇭 Global One SI - Prompt
Anthropic has launched Project Glasswing which is a new cybersecurity initiative powered by Claude Mythos Preview to help secure critical software. It is partnering with companies including $AMZN, $AAPL, $NVDA, $CRWD, $PANW to find and fix flaws in critical…
Anthropic just built an AI so dangerous they refuse to let anyone use it.
This is a weapon and it already found bugs hiding inside the world's most critical software for 27 years.
Here's what just changed about the entire internet.
Anthropic's new model is called Claude Mythos Preview and it was never supposed to exist publicly, Anthropic accidentally left it sitting on the open internet, and Fortune found it first.
Their own internal documents called it "by far the most powerful AI model" they've ever built.
They also said it "presages an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the defenders."
The company that built it is warning the world about it.
This model was never trained to hack but rather trained to write code and hacking was just a side effect.
It doesn't just find one vulnerability, it chains three, four, sometimes five of them together building attack sequences no human researcher would have conceived.
Anthropic's security researcher Nicholas Carlini said he found more bugs in the last few weeks than in the rest of his career combined.
Now for the buried story that almost nobody is talking about.
A Chinese state-sponsored hacking group already used an older, weaker version of Claude to attack 30 organizations tech companies, banks, and government agencies with 80 to 90% autonomous execution.
That was before Mythos existed.
So Anthropic made a decision, they will not release this model to the public but instead they launched Project Glasswing.
Amazon, Apple, Google, Microsoft, Nvidia, JPMorgan. , CrowdStrike and these companies get access first but only for defense.
The goal is to find and patch the world's most dangerous vulnerabilities before the attackers do.
The US government has already been briefed on the full offensive and defensive capabilities of this model.
Think about the scale of what this means.
Every browser you use, every operating system running your bank, every server powering the internet, all of it was built by humans who had limited time and limited eyesight.
This AI can scan all of it, faster than any human team ever could.
The model converts 72% of the vulnerabilities it finds into working exploits.
What an insane story.
This is a weapon and it already found bugs hiding inside the world's most critical software for 27 years.
Here's what just changed about the entire internet.
Anthropic's new model is called Claude Mythos Preview and it was never supposed to exist publicly, Anthropic accidentally left it sitting on the open internet, and Fortune found it first.
Their own internal documents called it "by far the most powerful AI model" they've ever built.
They also said it "presages an upcoming wave of models that can exploit vulnerabilities in ways that far outpace the defenders."
The company that built it is warning the world about it.
This model was never trained to hack but rather trained to write code and hacking was just a side effect.
It doesn't just find one vulnerability, it chains three, four, sometimes five of them together building attack sequences no human researcher would have conceived.
Anthropic's security researcher Nicholas Carlini said he found more bugs in the last few weeks than in the rest of his career combined.
Now for the buried story that almost nobody is talking about.
A Chinese state-sponsored hacking group already used an older, weaker version of Claude to attack 30 organizations tech companies, banks, and government agencies with 80 to 90% autonomous execution.
That was before Mythos existed.
So Anthropic made a decision, they will not release this model to the public but instead they launched Project Glasswing.
Amazon, Apple, Google, Microsoft, Nvidia, JPMorgan. , CrowdStrike and these companies get access first but only for defense.
The goal is to find and patch the world's most dangerous vulnerabilities before the attackers do.
The US government has already been briefed on the full offensive and defensive capabilities of this model.
Think about the scale of what this means.
Every browser you use, every operating system running your bank, every server powering the internet, all of it was built by humans who had limited time and limited eyesight.
This AI can scan all of it, faster than any human team ever could.
The model converts 72% of the vulnerabilities it finds into working exploits.
What an insane story.
✍1
🇨🇭 Global One SI - Prompt
$OSCR/USD H2 #Indicator #Signal #Trading #GlobalOneAI
$OSCR CEO Mark Bertolini bought 1 million shares in the open market at ~$12 according to a Form 4.
🇨🇭 Global One SI - Prompt
$BKSY/USD H2 #Indicator #Signal #Trading #GlobalOneAI
$BKSY is ripping higher today because the market is finally waking up to its massive potential.
With Gen-3 satellites delivering game-changing speed & AI intelligence, plus that huge $99M sole-source U.S. gov contract still fueling momentum, the street is catching on fast.
With Gen-3 satellites delivering game-changing speed & AI intelligence, plus that huge $99M sole-source U.S. gov contract still fueling momentum, the street is catching on fast.
✍1
This media is not supported in your browser
VIEW IN TELEGRAM
The US and Iran have reached an agre….
The market:
The market:
If you're unaware, NASA's next Artemis mission in 2027 will demonstrate docking and manned operations of SpaceX's HLS in Earth Orbit to prepare humans for a lunar landing in 2028.
✍1
Forwarded from 🛢WTI, Brent Crude Oil, NGAS Trading 📊
This chart explains why Trump isn't panicking.
And why the UK should be.
🇺🇸 US 1,700M barrels
🇯🇵 Japan 500M
🇩🇪 Germany 250M
🇰🇷 South Korea 200M
🇬🇧 UK barely visible
America has 3x more reserves than anyone else.
And it's also the world's 1st producer.
This is why Trump can walk away from Hormuz.
Why he can tell allies to fend for themselves.
Why he says "we have a lot of oil" and means it.
Energy security is the ultimate weapon.
The US has it. Most of Europe doesn't.
Every bar on this chart is shrinking daily.
Fastest for the smallest ones.
And why the UK should be.
🇺🇸 US 1,700M barrels
🇯🇵 Japan 500M
🇩🇪 Germany 250M
🇰🇷 South Korea 200M
🇬🇧 UK barely visible
America has 3x more reserves than anyone else.
And it's also the world's 1st producer.
This is why Trump can walk away from Hormuz.
Why he can tell allies to fend for themselves.
Why he says "we have a lot of oil" and means it.
Energy security is the ultimate weapon.
The US has it. Most of Europe doesn't.
Every bar on this chart is shrinking daily.
Fastest for the smallest ones.
🇨🇭 Global One SI - Prompt
Anthropic just built an AI so dangerous they refuse to let anyone use it. This is a weapon and it already found bugs hiding inside the world's most critical software for 27 years. Here's what just changed about the entire internet. Anthropic's new model…
Wow Mythos from Anthropic is a new level!
👀1
#oil
❗️ 🇮🇷 Donald Trump: Based on negotiations with Pakistani Prime Minister Shehbaz Sharif and Field Marshal Asim Munir, who asked me to postpone sending destructive forces to Iran tonight, and on the condition that the Islamic Republic of Iran agrees to the FULL, IMMEDIATE and SAFE opening of the Strait of Hormuz, I agree to suspend the bombing and attack on Iran for a period of two weeks. This will be a mutual TRUCE! The reason for this decision is that we have already achieved and exceeded all military tasks and have made great progress in reaching a final agreement on long-term PEACE with Iran and PEACE in the Middle East. We have received a ten-point proposal from Iran and believe that it is a workable basis for negotiations. Almost all the controversial points of the past have been agreed upon between the United States and Iran, but a two-week period will allow us to finalize and implement the agreement. On behalf of the United States of America, as President, and also representing the countries of the Middle East, I am honored to see that this long-standing problem is close to being resolved.
Please open Telegram to view this post
VIEW IN TELEGRAM
👀1