Anthropic is reportedly moving from renting cloud compute to leasing and managing data centers itself, with more than 1GW of planned U.S. capacity and Google potentially backing the lease payments.
The old model is simple: Anthropic pays cloud providers for GPUs or custom chips, but the cloud company controls the building, power, networking, cooling, and much of the hardware schedule.
The new model puts Anthropic closer to the factory floor of AI, where data centers are not office buildings but power-heavy machine rooms built to feed training and inference systems around the clock.
To give context of its current situation, until now, Anthropic’s compute strategy has mostly run through cloud providers, with over 10GW of server rentals committed, including a $200 billion Google agreement.
The company has also locked in large cloud deals with Akamai, AWS, CoreWeave, and Fluidstack, covering Amazon’s Trainium hardware and a $50 billion Fluidstack partnership.
It has also expanded its data center team and signed a SpaceX/xAI lease for the whole Colossus 1 data center at $1.25 billion a month, plus Colossus II space.
The old model is simple: Anthropic pays cloud providers for GPUs or custom chips, but the cloud company controls the building, power, networking, cooling, and much of the hardware schedule.
The new model puts Anthropic closer to the factory floor of AI, where data centers are not office buildings but power-heavy machine rooms built to feed training and inference systems around the clock.
To give context of its current situation, until now, Anthropic’s compute strategy has mostly run through cloud providers, with over 10GW of server rentals committed, including a $200 billion Google agreement.
The company has also locked in large cloud deals with Akamai, AWS, CoreWeave, and Fluidstack, covering Amazon’s Trainium hardware and a $50 billion Fluidstack partnership.
It has also expanded its data center team and signed a SpaceX/xAI lease for the whole Colossus 1 data center at $1.25 billion a month, plus Colossus II space.
🇨🇭 Global One SI - Prompt
AI doomers are not “cautious.” They are trying to freeze the one technology that could cure diseases, end scarcity, accelerate science, and give humanity a real shot at surviving this century. Blocking AI progress while billions still suffer is pure evil.
AI doomers will lose within the next 10 years.
Not because the risks are fake.
Some of them are real.
They’ll lose because people will see what AI actually does when it gets good enough.
Cures get discovered faster.
Small businesses get superpowers.
Robots take over dangerous work.
Science speeds up.
Education gets personal.
Millions of boring tasks just disappear.
The public mood won’t stay frozen in fear once the benefits become obvious in daily life.
Nobody wants to ban electricity after seeing the lightbulb.
And nobody will want to stop AI once it starts giving people more health, more time, more freedom and more abundance.
The doom narrative only works while the future is abstract.
Reality is going to beat it.
Not because the risks are fake.
Some of them are real.
They’ll lose because people will see what AI actually does when it gets good enough.
Cures get discovered faster.
Small businesses get superpowers.
Robots take over dangerous work.
Science speeds up.
Education gets personal.
Millions of boring tasks just disappear.
The public mood won’t stay frozen in fear once the benefits become obvious in daily life.
Nobody wants to ban electricity after seeing the lightbulb.
And nobody will want to stop AI once it starts giving people more health, more time, more freedom and more abundance.
The doom narrative only works while the future is abstract.
Reality is going to beat it.
🗿1
This media is not supported in your browser
VIEW IN TELEGRAM
A positive timeline for the next 10 years:
2026:
AI agents become normal for coding, research, scheduling, customer support and admin work.
2027:
Companies stop hiring humans for repetitive digital tasks first. Not all work. Just the boring stuff nobody dreams about as a child.
2028:
Humanoid robots move from demos into real deployment: warehouses, factories, cleaning, logistics, elder care.
2029:
AI agents become full digital coworkers. They use apps, make calls, book things, negotiate, plan and finish tasks end to end.
2030:
The job market starts to feel the real shock. At the same time, goods and services begin getting cheaper because automation scales.
2031:
The first serious AI Dividend proposals arrive. If machines create the value, people will demand a share of that productivity.
2032:
Robot labor becomes cheap enough to change construction, farming, manufacturing and delivery.
2033:
Some countries and cities launch early AI Dividend systems funded by automation taxes, sovereign AI funds or public ownership stakes.
2034:
Working less becomes normal in wealthy countries. Four day weeks feel outdated. The debate shifts toward income without jobs.
2035:
The first real post labor economy begins to appear.
Not utopia.
But a world where survival is no longer tied to selling your time.
That is the timeline worth building.
2026:
AI agents become normal for coding, research, scheduling, customer support and admin work.
2027:
Companies stop hiring humans for repetitive digital tasks first. Not all work. Just the boring stuff nobody dreams about as a child.
2028:
Humanoid robots move from demos into real deployment: warehouses, factories, cleaning, logistics, elder care.
2029:
AI agents become full digital coworkers. They use apps, make calls, book things, negotiate, plan and finish tasks end to end.
2030:
The job market starts to feel the real shock. At the same time, goods and services begin getting cheaper because automation scales.
2031:
The first serious AI Dividend proposals arrive. If machines create the value, people will demand a share of that productivity.
2032:
Robot labor becomes cheap enough to change construction, farming, manufacturing and delivery.
2033:
Some countries and cities launch early AI Dividend systems funded by automation taxes, sovereign AI funds or public ownership stakes.
2034:
Working less becomes normal in wealthy countries. Four day weeks feel outdated. The debate shifts toward income without jobs.
2035:
The first real post labor economy begins to appear.
Not utopia.
But a world where survival is no longer tied to selling your time.
That is the timeline worth building.
✍1
🇨🇭 Global One SI - Prompt
And this is exactly why I bought the dip ago. Conviction Pays 🫡
Today marks the beginning of a new chapter.
$SPCX vs other space stocks today.
Clear liquidity vacuum effect. Halo effect follows which makes companies as $PL, $ASTS, $RKLB buying opportunities at these levels.
Have a great weekend.
$SPCX vs other space stocks today.
Clear liquidity vacuum effect. Halo effect follows which makes companies as $PL, $ASTS, $RKLB buying opportunities at these levels.
Have a great weekend.
🇨🇭 Global One SI - Prompt
Photo
This media is not supported in your browser
VIEW IN TELEGRAM
When finance class taught you to value companies based on future cash flows, but then a no cash flow company ipos at a $1 trillion valuation.
This media is not supported in your browser
VIEW IN TELEGRAM
Picture the moment everything goes dark.
Not a blackout.
Worse.
The networks still hum...
... but they no longer answer to you.
The self-driving cars freeze mid-intersection.
The factory robots stall.
The drones over the border stop reporting.
The hospital's monitors flatline their data.
And somewhere far away, in a country that is not yours, someone's hand is on the switch.
This is the scenario that keeps national-security planners up at night. And almost nobody is pricing it in.
👇
Here's the chain that gets us there.
AI is becoming everything.
And the next phase of it doesn't live in a data center.
It lives at the EDGE.
On the towers.
Wired into the cars, the grids, the factories, the weapons.
Real-time intelligence, processed locally, in milliseconds.
That's the future everyone's racing toward:
AI-RAN.
The cell network turned into the nervous system of an entire civilization.
Now ask the question no one wants to ask: who builds that nervous system?
🇨🇳 Huawei & ZTE — China.
🇫🇮 Nokia — Finland.
🇸🇪 Ericsson — Sweden.
🇰🇷 Samsung — South Korea.
Not one of them is American.
China's hardware the US already ripped out, espionage, backdoors, an adversary wired into the grid.
The others are allies, but they're still foreign.
And here's the catastrophic part: when that network was just carrying phone calls, foreign hands were a tolerable risk.
When it becomes the real-time brain running your economy and your military, foreign control of that layer becomes a loaded gun pointed at your own head.
A single compromised radio layer, in that world, isn't a data breach. It's a kill switch on a nation.
You would be running the most powerful AI economy in history on hardware you didn't build, can't fully audit, and don't control.
That's not a risk.
That's a fault line under the entire house.
This is why Washington is moving fast.
Funding Open RAN to shatter the foreign monopoly.
Standing up Open6G, the DoD-funded hub, to build a trusted, domestic foundation before the AI era locks in.
But there's one terrifying gap in the plan: you cannot build an American radio supply chain if no American company makes the radio.
And at the spec that matters most, the 64T64R Massive MIMO radio that AI-RAN runs on, there is exactly one.
$AMPG.
The only American company to design and commercialize it. Inside the DoD-funded hub.
Defense-qualified:
Lockheed, Northrop, L3Harris, Boeing.
Already deployed at a Tier-1 carrier.
Already tested running AI-RAN with NVIDIA on top.
So this was never about a stock beating Nokia or Ericsson. It's about whether the most powerful nation on Earth will own the foundation of its own AI future, or rent it from someone who can switch it off.
When AI becomes everything, the edge becomes everything. And the edge runs on the radio.
There's only one American one.
And the clock is running.
Not financial advice. I'm long $AMPG. DYOR. 📡
Not a blackout.
Worse.
The networks still hum...
... but they no longer answer to you.
The self-driving cars freeze mid-intersection.
The factory robots stall.
The drones over the border stop reporting.
The hospital's monitors flatline their data.
And somewhere far away, in a country that is not yours, someone's hand is on the switch.
This is the scenario that keeps national-security planners up at night. And almost nobody is pricing it in.
👇
Here's the chain that gets us there.
AI is becoming everything.
And the next phase of it doesn't live in a data center.
It lives at the EDGE.
On the towers.
Wired into the cars, the grids, the factories, the weapons.
Real-time intelligence, processed locally, in milliseconds.
That's the future everyone's racing toward:
AI-RAN.
The cell network turned into the nervous system of an entire civilization.
Now ask the question no one wants to ask: who builds that nervous system?
🇨🇳 Huawei & ZTE — China.
🇫🇮 Nokia — Finland.
🇸🇪 Ericsson — Sweden.
🇰🇷 Samsung — South Korea.
Not one of them is American.
China's hardware the US already ripped out, espionage, backdoors, an adversary wired into the grid.
The others are allies, but they're still foreign.
And here's the catastrophic part: when that network was just carrying phone calls, foreign hands were a tolerable risk.
When it becomes the real-time brain running your economy and your military, foreign control of that layer becomes a loaded gun pointed at your own head.
A single compromised radio layer, in that world, isn't a data breach. It's a kill switch on a nation.
You would be running the most powerful AI economy in history on hardware you didn't build, can't fully audit, and don't control.
That's not a risk.
That's a fault line under the entire house.
This is why Washington is moving fast.
Funding Open RAN to shatter the foreign monopoly.
Standing up Open6G, the DoD-funded hub, to build a trusted, domestic foundation before the AI era locks in.
But there's one terrifying gap in the plan: you cannot build an American radio supply chain if no American company makes the radio.
And at the spec that matters most, the 64T64R Massive MIMO radio that AI-RAN runs on, there is exactly one.
$AMPG.
The only American company to design and commercialize it. Inside the DoD-funded hub.
Defense-qualified:
Lockheed, Northrop, L3Harris, Boeing.
Already deployed at a Tier-1 carrier.
Already tested running AI-RAN with NVIDIA on top.
So this was never about a stock beating Nokia or Ericsson. It's about whether the most powerful nation on Earth will own the foundation of its own AI future, or rent it from someone who can switch it off.
When AI becomes everything, the edge becomes everything. And the edge runs on the radio.
There's only one American one.
And the clock is running.
Not financial advice. I'm long $AMPG. DYOR. 📡
✍1
🇨🇭 Global One SI - Prompt
If I had to line up humanoids for the FIFA World Cup this is what I would do. NEO on the left wing because 1X is chasing the home robot role. Optimus as striker because Tesla brings manufacturing pressure. Figure 03 on the right because Figure is pushing…
This media is not supported in your browser
VIEW IN TELEGRAM
Team Players $SPCX $TSLA
🇨🇭 Global One SI - Prompt
Elon Musk Net Worth Since 2000 💰 🇺🇸 2000 ⟶ 💵 $28M 🇺🇸 2001 ⟶ 💵 $34M 🇺🇸 2002 ⟶ 💵 $180M 🇺🇸 2003 ⟶ 💵 $160M 🇺🇸 2004 ⟶ 💵 $210M 🇺🇸 2005 ⟶ 💵 $250M 🇺🇸 2006 ⟶ 💵 $320M 🇺🇸 2007 ⟶ 💵 $400M 🇺🇸 2008 ⟶ 💵 $150M 🇺🇸 2009 ⟶ 💵 $350M 🇺🇸 2010 ⟶ 💵 $650M 🇺🇸 2011 ⟶ 💵 $680M 🇺🇸 2012…
Well it's official, Elon Musk becomes the world's first public TRILLIONAIRE!
Combining his two primary holdings ($SPCX + $TSLA ) and a closing price above IPO at $160.95, it gives Mr. Musk a combined net worth of approximately $1.32 trillion
I wonder what he'll buy next?
Combining his two primary holdings ($SPCX + $TSLA ) and a closing price above IPO at $160.95, it gives Mr. Musk a combined net worth of approximately $1.32 trillion
I wonder what he'll buy next?
🐳1
This media is not supported in your browser
VIEW IN TELEGRAM
ANTHROPIC JUST CRACKED THE ENTERPRISE WALL
Claude agents can now run their tools inside your own infrastructure, while anthropic still runs the brain.
The thing that kept banks, hospitals and governments from deploying ai agents in prod is gone. your data never leaves your perimeter.
- the agent loop, model calls and orchestration stay on anthropic's side
- tool execution, files and network egress move fully into your environment
- ships with cloudflare, daytona, modal, vercel and 5 new providers
- mcp tunnels let agents reach private servers with zero inbound firewall rules
The model wars get the headlines. this is the part that actually ships agents to prod.
Claude agents can now run their tools inside your own infrastructure, while anthropic still runs the brain.
The thing that kept banks, hospitals and governments from deploying ai agents in prod is gone. your data never leaves your perimeter.
- the agent loop, model calls and orchestration stay on anthropic's side
- tool execution, files and network egress move fully into your environment
- ships with cloudflare, daytona, modal, vercel and 5 new providers
- mcp tunnels let agents reach private servers with zero inbound firewall rules
The model wars get the headlines. this is the part that actually ships agents to prod.
👀1
$DGXX The multi-billion dollar Cerebras partnership completely de-risks Digi Power X's operational pipeline, turning a speculative grid operator into a highly tangible AI proxy.
Furthermore, institutional confidence is accelerating rapidly, as evidenced by Citadel’s massive 4.9% stake accumulation in the company.
With the recent $35 million commitment to next-gen Nvidia Vera Rubin systems, the company proves it is not just selling space, but aggressively scaling high-margin GPU-as-a-Service capabilities.
Furthermore, institutional confidence is accelerating rapidly, as evidenced by Citadel’s massive 4.9% stake accumulation in the company.
With the recent $35 million commitment to next-gen Nvidia Vera Rubin systems, the company proves it is not just selling space, but aggressively scaling high-margin GPU-as-a-Service capabilities.
$DGXX scores a massive win by securing $NVDA highly anticipated, next-gen Vera Rubin platform early.
This $35 million deal is a phenomenal move for three major reasons:
Future-Proof Monopoly: They bypass current Blackwell limits to offer unmatched 288GB HBM4 power by Q1 2027.
Massive Margin Growth: Booking this ultra-premium capacity via NeoCloudz will drive skyrocketing, high-margin AI revenue.
Flawless Financial Health: Funding this entire next-gen revolution purely from cash on hand proves incredible liquidity and zero debt risk.
This $35 million deal is a phenomenal move for three major reasons:
Future-Proof Monopoly: They bypass current Blackwell limits to offer unmatched 288GB HBM4 power by Q1 2027.
Massive Margin Growth: Booking this ultra-premium capacity via NeoCloudz will drive skyrocketing, high-margin AI revenue.
Flawless Financial Health: Funding this entire next-gen revolution purely from cash on hand proves incredible liquidity and zero debt risk.
The announcement of $HQ establishing its second quantum computer testbed in Dublin represents a major commercial and technological expansion.
By integrating a 256-qubit system from $IONQ into its European headquarters, the company addresses the two primary challenges in the current quantum computing industry: hardware modality risk and geographic market access.
This expansion diversifies Horizon Quantum’s operational infrastructure and positions its proprietary software ecosystem for broader commercial adoption.
Operating distinct testbeds across different hardware architectures directly accelerates Horizon Quantum's software development.
The company’s existing facility in Singapore utilizes superconducting quantum processors, whereas the new Dublin deployment introduces trapped-ion technology from IonQ.
Because it remains unproven which hardware architecture will ultimately dominate the market, developing software that operates seamlessly across both modalities mitigates technical risk. This dual-track approach ensures that Horizon Quantum’s developer platform remains hardware-agnostic, allowing customers to compile and run applications regardless of the underlying processor technology.
The technical specifications of the Dublin system provide immediate computing advantages for software optimization.
The IonQ system offers 256 qubits combined with a target two-qubit gate fidelity of 99.99% and all-to-all connectivity. In quantum computing, high gate fidelity reduces the error rates that typically disrupt complex algorithmic executions.
By deploying its Triple Alpha software platform on this high-fidelity hardware, $HQ can demonstrate enterprise-grade performance metrics, which are essential for attracting corporate clients in sectors like finance, cryptography, and logistics.
Establishing this facility in Ireland secures a strategic operational base within the European Union.
Through collaboration with IDA Ireland and local academic institutions, Horizon Quantum gains direct access to regional deep-tech talent and European research funding frameworks.
This geographic expansion follows the company's recent public listing on the Nasdaq, demonstrating a rapid deployment of capital into physical infrastructure.
By anchoring its technology in both Asia and Europe, Horizon Quantum scales its deployment capacity to capture early enterprise demand for quantum-accelerated computing infrastructure.
By integrating a 256-qubit system from $IONQ into its European headquarters, the company addresses the two primary challenges in the current quantum computing industry: hardware modality risk and geographic market access.
This expansion diversifies Horizon Quantum’s operational infrastructure and positions its proprietary software ecosystem for broader commercial adoption.
Operating distinct testbeds across different hardware architectures directly accelerates Horizon Quantum's software development.
The company’s existing facility in Singapore utilizes superconducting quantum processors, whereas the new Dublin deployment introduces trapped-ion technology from IonQ.
Because it remains unproven which hardware architecture will ultimately dominate the market, developing software that operates seamlessly across both modalities mitigates technical risk. This dual-track approach ensures that Horizon Quantum’s developer platform remains hardware-agnostic, allowing customers to compile and run applications regardless of the underlying processor technology.
The technical specifications of the Dublin system provide immediate computing advantages for software optimization.
The IonQ system offers 256 qubits combined with a target two-qubit gate fidelity of 99.99% and all-to-all connectivity. In quantum computing, high gate fidelity reduces the error rates that typically disrupt complex algorithmic executions.
By deploying its Triple Alpha software platform on this high-fidelity hardware, $HQ can demonstrate enterprise-grade performance metrics, which are essential for attracting corporate clients in sectors like finance, cryptography, and logistics.
Establishing this facility in Ireland secures a strategic operational base within the European Union.
Through collaboration with IDA Ireland and local academic institutions, Horizon Quantum gains direct access to regional deep-tech talent and European research funding frameworks.
This geographic expansion follows the company's recent public listing on the Nasdaq, demonstrating a rapid deployment of capital into physical infrastructure.
By anchoring its technology in both Asia and Europe, Horizon Quantum scales its deployment capacity to capture early enterprise demand for quantum-accelerated computing infrastructure.
Do you realize Edge AI is going to control everything around you?
And did you know the United States 🇺🇸 depends almost 100% on Europe 🇪🇺 and companies like $NOK?
Yes.
For the thing that will control everything.
The solution is being developed right now by $AMPG and $NVDA.
Let me explain the whole thing.
Edge AI isn't figurative.
It's literal.
The traffic lights at your intersection.
The car next to you.
The robots on the factory floor.
The drones overhead.
The power grid feeding your house.
All of it run by AI that cannot wait for a faraway data center to answer.
It needs decisions in milliseconds, made locally, at the edge of the network.
And the edge of the network is the cell tower.
That's where this entire revolution physically lives.
Now the dependence problem.
The radios those towers run on are built by foreign companies: Nokia (Finland), Ericsson (Sweden), Samsung (Korea).
The Chinese ones, Huawei and ZTE, are banned outright.
The most powerful nation on Earth, about to wire its cars, factories, grids and defense systems into an AI-driven network, doesn't make its own radios.
That's not a supply-chain detail.
That's a strategic vulnerability.
Here's how the solution fits together.
The industry calls this AI-RAN, and the AI-RAN
Alliance breaks it into 3 focus areas:
1️⃣ AI-AND-RAN: AI and the network sharing the same infrastructure. The tower runs network traffic AND AI workloads on the same hardware. Every tower becomes a mini data center. Payoff: asset utilization, monetizing billions in idle infrastructure.
2️⃣ AI-ON-RAN: AI applications enabled by the network. The network becomes a platform for real-time edge AI: autonomous vehicles, AR, industrial robots. Payoff: new applications that cloud-only AI physically can't do, because the round-trip is too slow.
3️⃣ AI-FOR-RAN: AI improving the network itself. Smarter beamforming, scheduling, self-optimizing cells. Payoff: spectral efficiency, more performance from the same scarce, expensive spectrum.
Now the part the diagram doesn't say out loud.
All three are software and compute layers.
That's NVIDIA's world: Aerial, the GPU brain.
NVIDIA is pushing so hard it's even signaling GPUs inside the radio itself will become essential as MIMO scales toward 6G.
But NVIDIA doesn't make radios.
The brain needs a body.
An open, high-capacity radio the AI layers can actually plug into, not a locked foreign box.
$AMPG is the only American company that designs and commercializes that open 64T64R Massive MIMO radio.
And the two have already met: AMPG's radio ran with NVIDIA's Aerial software in the world's first open-source AI-RAN demo, inside Open6G, the hub the DoD is funding precisely to build a trusted, domestic foundation for next-gen wireless.
Source: prnewswire.com/news-releases/northeastern-university-demonstrates-first-open-source-prototype-of-a-massive-mimo-ai-ran-system-302777855.html
America's brain (NVIDIA) and America's radio (AMPG), tested together, in the government's own lab.
Add what's already real: deployed at Telus, a Tier-1 carrier, 2 of 5 radios per sector per Telus's own VP.
Defense-qualified with Lockheed, Northrop, L3Harris, Boeing. Debt-free, 48% gross margins, sub-$1B.
So when people say the next photonics-style run is hiding in plain sight in 5G/6G, Edge AI and AI-RAN, they're right.
But ask the better question: which layer do you own?
Everyone's crowded into software and compute, where the giants already rule.
I'm in the layer with exactly one American name: the physical foundation the entire revolution has to run on, in a country that just realized it can't keep renting it from abroad.
That's why I'm in $AMPG.
Not financial advice. I'm long $AMPG. DYOR.
And did you know the United States 🇺🇸 depends almost 100% on Europe 🇪🇺 and companies like $NOK?
Yes.
For the thing that will control everything.
The solution is being developed right now by $AMPG and $NVDA.
Let me explain the whole thing.
Edge AI isn't figurative.
It's literal.
The traffic lights at your intersection.
The car next to you.
The robots on the factory floor.
The drones overhead.
The power grid feeding your house.
All of it run by AI that cannot wait for a faraway data center to answer.
It needs decisions in milliseconds, made locally, at the edge of the network.
And the edge of the network is the cell tower.
That's where this entire revolution physically lives.
Now the dependence problem.
The radios those towers run on are built by foreign companies: Nokia (Finland), Ericsson (Sweden), Samsung (Korea).
The Chinese ones, Huawei and ZTE, are banned outright.
The most powerful nation on Earth, about to wire its cars, factories, grids and defense systems into an AI-driven network, doesn't make its own radios.
That's not a supply-chain detail.
That's a strategic vulnerability.
Here's how the solution fits together.
The industry calls this AI-RAN, and the AI-RAN
Alliance breaks it into 3 focus areas:
1️⃣ AI-AND-RAN: AI and the network sharing the same infrastructure. The tower runs network traffic AND AI workloads on the same hardware. Every tower becomes a mini data center. Payoff: asset utilization, monetizing billions in idle infrastructure.
2️⃣ AI-ON-RAN: AI applications enabled by the network. The network becomes a platform for real-time edge AI: autonomous vehicles, AR, industrial robots. Payoff: new applications that cloud-only AI physically can't do, because the round-trip is too slow.
3️⃣ AI-FOR-RAN: AI improving the network itself. Smarter beamforming, scheduling, self-optimizing cells. Payoff: spectral efficiency, more performance from the same scarce, expensive spectrum.
Now the part the diagram doesn't say out loud.
All three are software and compute layers.
That's NVIDIA's world: Aerial, the GPU brain.
NVIDIA is pushing so hard it's even signaling GPUs inside the radio itself will become essential as MIMO scales toward 6G.
But NVIDIA doesn't make radios.
The brain needs a body.
An open, high-capacity radio the AI layers can actually plug into, not a locked foreign box.
$AMPG is the only American company that designs and commercializes that open 64T64R Massive MIMO radio.
And the two have already met: AMPG's radio ran with NVIDIA's Aerial software in the world's first open-source AI-RAN demo, inside Open6G, the hub the DoD is funding precisely to build a trusted, domestic foundation for next-gen wireless.
Source: prnewswire.com/news-releases/northeastern-university-demonstrates-first-open-source-prototype-of-a-massive-mimo-ai-ran-system-302777855.html
America's brain (NVIDIA) and America's radio (AMPG), tested together, in the government's own lab.
Add what's already real: deployed at Telus, a Tier-1 carrier, 2 of 5 radios per sector per Telus's own VP.
Defense-qualified with Lockheed, Northrop, L3Harris, Boeing. Debt-free, 48% gross margins, sub-$1B.
So when people say the next photonics-style run is hiding in plain sight in 5G/6G, Edge AI and AI-RAN, they're right.
But ask the better question: which layer do you own?
Everyone's crowded into software and compute, where the giants already rule.
I'm in the layer with exactly one American name: the physical foundation the entire revolution has to run on, in a country that just realized it can't keep renting it from abroad.
That's why I'm in $AMPG.
Not financial advice. I'm long $AMPG. DYOR.