
$NVDA is developing a new GPU-based chip designed to go directly inside the Radio Unit (RU) — the tower-mounted part of a base station, like $AMPG 64T64R mMIMO panel — as a major expansion of its AI-RAN strategy.
Olyth
17 posts

@olyth_terminal
Institutional-Grade Market Terminal.

$NVDA is developing a new GPU-based chip designed to go directly inside the Radio Unit (RU) — the tower-mounted part of a base station, like $AMPG 64T64R mMIMO panel — as a major expansion of its AI-RAN strategy.

I'm a simple man. Edge AI is the future. Everyone can see that coming. It's going to run on radios. The same layer $NVDA $NOK and the big telecom players are racing to build. And $AMPG is a Strategic Partner in Open6G. The government tells me what to buy. I just buy. I won't bet against Uncle Sam. Not that dumb. Like when Trump told people to buy $INTC or $DELL. Surprise. He was right. $AMPG is in the wireless hub the US Department of Defense is funding, sitting in the top tier next to NVIDIA, Dell and Qualcomm. It's also the only American company making the 64T64R radio that open AI-RAN runs on. So I keep it simple. I buy $AMPG. Because I don't believe the US wants to depend on foreign vendors for something this critical. Not when it's about to control every device we own. Not after what I saw with Claude. Simple as that. They're whispering it to me clearly. Still sub $1B market cap. Not financial advice. I'm long $AMPG. DYOR. 📡



$AMPG told you - did you listen? Can we now trigger the short squeeze ?!? Thanks 🙏 🚀 🧨






🔮 GENESIS agentic AI framework is the biggest moat for $AMPG. And they didn't even build it. It was publicly unveiled June 4th, and I think it's the reason why $AMPG is getting so much attention. GENESIS is a groundbreaking agentic AI framework developed by researchers at Northeastern University’s Institute for Intelligent Networked Systems (INSI) / Open6G OTIC. GENESIS automates the entire R&D lifecycle of Radio Access Network (RAN) software for 5G and will speed up the development of future 6G networks. What once took engineers months to achieve can now be done autonomously in hours. GENESIS gives a real speed advantage to open players like AmpliTech in rapid 6G RAN innovation. And none of the legacy companies like Nokia and Ericsson are able to take advantage of it. Their engineers would largely need to manually engineer new RAN features and the underlying code. 🧠 Here's how it works. Given a high-level intent like an O-RAN specification clause, a performance anomaly from telemetry, or a new optimization idea, it autonomously: 1. Plans and decomposes the task. 2. Writes/synthesizes code. 3. Tests it across a continuum (simulation → channel emulation → real over-the-air on production hardware). 4. Hardens, optimizes, and secures it. 5. Deploys validated code over-the-air (OTA) on live radios. 6. Feeds all results back into a persistent knowledge base called Synapse for continuous improvement. If this sounds like the future it's because it is. In head-to-head experiments on a representative 5G feature implementation task, GENESIS achieved a 100% success rate across multiple independent runs. An off-the-shelf state-of-the-art coding agent (🧐Claude) used as a baseline — given access to the same tools and testbeds — produced no working implementation on any attempt. It’s essentially AmpliTech-only for GENESIS agentic AI workflows on true open Massive MIMO hardware today. AmpliTech 64T64R is the central hardware for the first open-source Massive MIMO AI-RAN prototype with GENESIS, $NVDA AI Aerial, and OpenAirInterface (OAI). 👀 FULL OPEN SOURCE AI-RAN STACK. It is repeatedly described as the only U.S.-designed and commercialized 64T64R O-RAN Category B radio. The only radio of its configuration certified by the Open6G OTIC and participating in recent O-RAN PlugFests for this spec level. This is a significant differentiator for them in the disaggregated, Open RAN ecosystem and gives AmpliTech a strong first-mover position. Other vendors like Samsung, Ericsson, Nokia, Huawei, offer commercial 64T64R Massive MIMO radios. BUT They are proprietary and integrated, not fully disaggregated O-RAN Category B with open fronthaul suitable for GENESIS-style agentic OTA upgrades. They are not validated in the same open-source AI-RAN stacks. This is what the major mobile network operators are trying to break away from. This makes AmpliTech the de facto leader for high-capacity open Massive MIMO in research and early commercial open RAN deployments right now. Especially when it comes to AI native 6g and the AI edge networks. And the market is starting to realize that.



This is a very well researched piece by Johan. I just bought some $AMPG as it's breaking out to a new 52 week high. I'll likely cut if it loses momentum. I want to some nice continuation here.

The next phase of the AI Infrastructure buildout is at the edge. $NVDA and $AMPG are turning cell towers into intelligent edge nodes. 🧠 Here's How They're Doing It. Amplitech recently became the first to develop an open-source massive MIMO AI-RAN system. This allows the major mobile network operators to break free of legacy software/hardware stacks and supports the broader push toward AI-native 6G. Now Nvidia is developing a new GPU chip that can go directly into the radio units Amplitech makes. These intelligent edge nodes run GENESIS agents, upgrading/deploying it's own code, improving it's own software autonomously and use Nvidia AI AERIAL to optimize and send out signals more efficiently than ever before. Nvidia's approach of putting GPUs in Radio Units, like Amplitech has developed, lets base stations handle both RAN Layer 1 and edge AI workloads concurrently via Multi-Instance GPU. This creates a "robotic AI radio" or distributed AI grid. 100% True AI - from the software coding agent to the signal management. Major mobile network operators can monetize spare capacity of SLA-backed low-latency services and support 6G's AI-native vision. It turns underutilized towers into high-value edge compute nodes for external customers. Edge AI is pivotal because the future of tech is physical, distributed, and instantaneous. Together they are transforming telecom infrastructure from pure connectivity providers into distributed AI platforms, unlocking new revenue while solving core limitations of cloud-only AI. 🕵️ Why Edge AI Matters: Cloud round-trips introduce delays that are unacceptable for safety-critical or immersive uses. Edge AI processes data locally, enabling millisecond responses. This is non-negotiable for autonomous vehicles, industrial robotics, or AR/VR. Sending massive sensor/IoT data streams to the cloud is expensive and congests networks. Edge filtering and inference transmit only actionable insights, slashing bandwidth needs and costs. It also reduces energy use in power-constrained environments. Edge works offline or in remote areas, scaling to billions of devices without overwhelming central data centers. 💸 Market Momentum and Projections: Edge AI is moving from hype to mainstream deployment in 2026. The market is projected to grow from ~$26 billion in 2026 to over $245 billion by 2040 (CAGR ~17.4%). AI-RAN itself is exploding, with forecasts of $37+ billion by 2035. Telcos leveraging GPU-equipped cell sites for GPU-as-a-Service could capture $21–70 billion annually by 2030 in new revenues from low-latency inference, far beyond traditional connectivity. 🧠 Main Takeaway: Centralized cloud AI won't scale to trillions of sensors, real-world autonomy, or privacy demands. By embedding it into existing infrastructure like cell towers, developments like AI-RAN accelerate this shift, enabling breakthroughs across industries. It's a core enabler for everything from 6G to intelligent physical systems.

$AMPG maybe biggest news of today. This is old news from 2025 but now when ORAN is ramping it is very important to understand this. EVERY TELUS ORAN SITE FEATURES $AMPG ANTENNAS. 5 RADIOS PER SECTOR and TWO FDD MID-BAND RADIOS FROM AMPLITECH! This can potentially generate insane amount of revenue if ORAN ramp up continues. Credit: @Lonsdale171255 for finding this. the-mobile-network.com/2025/09/telus-…

The next phase of the AI Infrastructure buildout is at the edge. $NVDA and $AMPG are turning cell towers into intelligent edge nodes. 🧠 Here's How They're Doing It. Amplitech recently became the first to develop an open-source massive MIMO AI-RAN system. This allows the major mobile network operators to break free of legacy software/hardware stacks and supports the broader push toward AI-native 6G. Now Nvidia is developing a new GPU chip that can go directly into the radio units Amplitech makes. These intelligent edge nodes run GENESIS agents, upgrading/deploying it's own code, improving it's own software autonomously and use Nvidia AI AERIAL to optimize and send out signals more efficiently than ever before. Nvidia's approach of putting GPUs in Radio Units, like Amplitech has developed, lets base stations handle both RAN Layer 1 and edge AI workloads concurrently via Multi-Instance GPU. This creates a "robotic AI radio" or distributed AI grid. 100% True AI - from the software coding agent to the signal management. Major mobile network operators can monetize spare capacity of SLA-backed low-latency services and support 6G's AI-native vision. It turns underutilized towers into high-value edge compute nodes for external customers. Edge AI is pivotal because the future of tech is physical, distributed, and instantaneous. Together they are transforming telecom infrastructure from pure connectivity providers into distributed AI platforms, unlocking new revenue while solving core limitations of cloud-only AI. 🕵️ Why Edge AI Matters: Cloud round-trips introduce delays that are unacceptable for safety-critical or immersive uses. Edge AI processes data locally, enabling millisecond responses. This is non-negotiable for autonomous vehicles, industrial robotics, or AR/VR. Sending massive sensor/IoT data streams to the cloud is expensive and congests networks. Edge filtering and inference transmit only actionable insights, slashing bandwidth needs and costs. It also reduces energy use in power-constrained environments. Edge works offline or in remote areas, scaling to billions of devices without overwhelming central data centers. 💸 Market Momentum and Projections: Edge AI is moving from hype to mainstream deployment in 2026. The market is projected to grow from ~$26 billion in 2026 to over $245 billion by 2040 (CAGR ~17.4%). AI-RAN itself is exploding, with forecasts of $37+ billion by 2035. Telcos leveraging GPU-equipped cell sites for GPU-as-a-Service could capture $21–70 billion annually by 2030 in new revenues from low-latency inference, far beyond traditional connectivity. 🧠 Main Takeaway: Centralized cloud AI won't scale to trillions of sensors, real-world autonomy, or privacy demands. By embedding it into existing infrastructure like cell towers, developments like AI-RAN accelerate this shift, enabling breakthroughs across industries. It's a core enabler for everything from 6G to intelligent physical systems.



@aleabitoreddit Navitas, lpk, wolfspeed and xfab all collapsing


