It's a tremendous matrix. The best matrix. It has agents like you have not seen before. You would not believe it.

Central banks just bought 15x more gold than they officially admitted.

The buying is real,the reporting is a smokescreen.

The World Gold Council estimates central banks added 244 tonnes of gold in Q1 2026.

But the official IMF data shows only about 16 tonnes of reported purchases.

That gap is enormous.

Roughly 228 tonnes, about 93% of the total, went unreported.

WGC infers the rest from trade data, vault flows, and balance of payments gaps, not official disclosures.

The "15x more than reported" headline is sensational, but directionally it's accurate.

Here's why it matters.

The hidden buyers are mostly emerging market central banks, led by China and Asia.

They're stockpiling gold as a de-dollarization and sanctions hedge, and they prefer to disclose late or not at all.

The pattern is structural, not a blip.
Over 200 tonnes added in 10 of the last 11 quarters.

And it continues even with gold near record highs.

That tells you the buying is strategic, not price sensitive.

When the biggest buyers don't care about price and won't show their hand, that's a powerful floor under gold.

The extinction threat isn’t public AI. It’s AI hidden behind closed doors.

Global Humanoid Robot Suppliers Map 2026

The humanoid race is becoming a full supply chain war.

• 100+ suppliers worldwide
• 50+ humanoid robot companies
• $15B+ projected humanoid market by 2030

➤ North America
Tesla
Figure
Agility Robotics
Apptronik
Sanctuary AI
Boston Dynamics

➤ China
Unitree
UBTECH
AGIBOT
Fourier
XPeng
Leju Robot
Keenon

➤ Europe
NEURA Robotics
PAL Robotics
Aldebaran
Ericsson
iRonCub3

➤ Japan / Korea
Toyota
Honda
Sony
Samsung
Rainbow Robotics

➤ Rest of Asia
Xiaomi
LimX Dynamics
PaXini
Hyundai Robotics

The winners will need more than a robot body.

They need:

• Core components
➝ Harmonic Drive
➝ Nabtesco
➝ MAXON

• Actuators and motors
➝ Siemens
➝ Yaskawa
➝ Inovance

• Sensors and perception
➝ Ouster
➝ Hesai
➝ RoboSense
➝ Sony IMX

• Batteries and power
➝ CATL
➝ BYD
➝ Panasonic
➝ LG Energy Solution

• AI and software
➝ NVIDIA
➝ Microsoft
➝ Google DeepMind
➝ OpenAI
➝ Huawei MindSpore

• Manufacturing and ODM
➝ Foxconn
➝ Flex
➝ Jabil
➝ Pegatron

• Testing and QA
➝ TÜV SÜD
➝ SGS
➝ Intertek
➝ UL Solutions

• System integration
➝ Accenture
➝ Deloitte
➝ Capgemini
➝ SoftBank Robotics

Humanoid robots are not one product category anymore.

They are becoming a global industrial stack.
Hardware. Actuation. Perception. Batteries. AI. Manufacturing. Deployment.

Trump and the US government just handed China the easiest win in the AI race.

60% of companies are already rerouting their AI spend toward cheaper, open-source models - and most of the best ones are Chinese (Qwen, DeepSeek, GLM, Kimi).

US frontier models are now being gatekept by the government AND priced at a premium for only marginal gains in intelligence.

To be fair, this is a genuinely hard spot for the US:

On one hand, you want America to lead AI innovation (especially so that you, as the government, can have access).

But you also want the most powerful models secured and controlled for national security reasons.

The problem is that those two goals are now directly at odds.

Every restriction placed on US frontier models makes the open-source alternative more attractive.

And for now, the best open-source alternatives aren't American.

Going to be very interesting to watch things play out from here.

🇺🇸 Fed's Kashkari says ‘I am concerned about inflation, especially in services.’

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

I keep wondering what happens to money in a post labor world.

If AI and robots can produce almost everything cheaper, faster and better than humans ever could, does money slowly lose its power?

At some point, scarcity is the only thing keeping the whole game alive.

$MU $SNDK $WDC $INTC $ASML | Apple lobbies the Trump administration for political clearance to purchase consumer memory chips from Chinese manufacturer CXMT.

The move seeks to ease supply chain constraints after soaring component costs forced recent MacBook and iPad price hikes.

$GEV $CEG $OKLO $LEU $UEC | Microsoft partners with Chevron on a twenty year agreement to power a massive West Texas data center using natural gas.

The colocated Project Kilby plant secures seven GE Vernova turbines to deliver 2.67 gigawatts (GW) of independent electricity.

"The share of tokens used for US models on OpenRouter has collapsed" Bloomberg

On OpenRouter, US model token share fell from around 70% to around 30% in a year.

A clean warning signal.

open-weight Chinese models are becoming capable enough for routine work, cheaper to run, easier to customize, and less dependent on permission from a frontier lab.

AI revenue has crossed its first serious accounting test: $25B in quarterly sales now exceeds $21B in estimated chip and data-center depreciation.

i.e. AI infrastructure is starting to pay for itself before power, labor, financing, and leases are counted.

I don’t think AI takes all jobs overnight.

That’s the cartoon version.

The real transition will be slower and much weirder: AI takes tasks first, then teams shrink, hiring slows, entry-level jobs disappear, and companies quietly realize they need fewer humans every year.

McKinsey already estimates that up to 30% of US work hours could be automated by 2030. The World Economic Forum expects 92 million jobs to be displaced by 2030, even while new jobs are created.

So no, everyone won’t be jobless next Monday.

But over the next 15 years, the pressure keeps building.

The obvious transition isn’t “no work.”

It’s 5 days becoming 4.

Then maybe 3.

Then the real question begins.

$IBM unveils world's first publicly announced sub-1nm chip technology 👀

They just unveiled NanoStack, a breakthrough 3D transistor architecture designed for the 7 Å (~0.7 nm) technology node, making it the world's first publicly announced chip technology below 1 nanometer.

Instead of placing transistors side by side, NanoStack stacks them vertically, allowing nearly 100 billion transistors to fit on a silicon chip about the size of a fingernail while dramatically increasing transistor density.

IBM says its prototype delivers up to 50% higher performance or up to 70% lower power consumption compared with its 2 nm research chip introduced in 2021.

Acceleration is everywhere!

$IBM/USD H2 #Indicator #Signal #Trading #GlobalOneAI
Globalone.ch #Prompt #AI

I knew the government was playing 4D chess when they restricted Mythos/Fable...

Trump Administration just partially lifted the ban.

Two weeks of standoff, now released to a controlled set of ~100 trusted US companies and agencies.

The government wasn't perma restricting AI capabilities... they were establishing control first (and possibly private placement deals).

Now we wait for the full Fable 5 release, equity stake announcements, and clearer plans for their IPO.

You can turn Hermes & Claude Code into your personal crypto trading assistants.

It can execute real market orders, manage live positions, and more.

Most traders have no idea this is even possible.

Send this prompt to Claude Code, and you'll be able to automate your trading

🔥 INSANE: Robotics Investments hit an ATH of $16B, per a16z.

How to Build Agent Skills for Claude (Step-by-Step)

When you come back to check on your old job and it's even worse than you expected

The smartest thing you can do with Hermes right now:

Create a self-evolving loop that trains Hermes on your data - so every session is smarter than the last.

Session 1 → learns → logs it → Session 2 is smarter → logs that → repeat

Here's exactly how to set it up:

1. Create a Memory.md file on your desktop

Structure it like this:

## Preferences
## Corrections
## Patterns
## Lessons learned

2. Attach it to Hermes

Paste this prompt once:

"At the start of every session, read Memory.md and apply everything in it.

After every task, do three things:

• Log what worked and why
• Log what failed and why
• Analyse and write rules you'd apply next time

Never duplicate entries. Rewrite existing rules when you learn something better."

3. Once a week, run this prompt:

"Review everything in Memory.md. Identify patterns across all your logged lessons. Distill them into sharper, more general rules. Delete anything that's been superseded. The goal: fewer, better rules every week."

4. Archive before every cleanup

Copy Memory.md into a dated backup before your agent rewrites anything.

(In case of any mistakes, this is how you secure previous data)

That's the self-evolution loop that everyone should be running with Hermes right now.

Super simple, but it makes Hermes 10x more powerful.

Make sure to bookmark this so you actually maximise your agentic workflows.