Olyth

JUST IN: $NVDA published the blueprint for AI-RAN. $NOK is named in it. But connect it with what Washington has been funding for years and it points somewhere else: a sub-$200M American micro-cap. $AMPG. But first, one thing about NVIDIA you need to understand, because it reframes the Nokia mention entirely. NVIDIA doesn't marry anyone. It never has. Look at how it played the neoclouds: it didn't crown one winner, it backed SEVERAL at once. CoreWeave, Nebius, Lambda, all supported, invested in, or supplied simultaneously. NVIDIA's playbook has never been "pick a champion". It's "arm the entire ecosystem", because its business is selling the platform underneath ALL of them. Now apply that to AI-RAN. Nokia gets named in the blog. SoftBank runs the field trials. Samsung shows up in the KDDI digital twin project. And the world's first fully open-source Massive MIMO AI-RAN demo? That ran on AMPG's American radio, at Northeastern. See the pattern? NVIDIA is doing in RAN exactly what it did in cloud: seeding every layer, every player, every path. The Nokia mention is not a closed door. It's confirmation that NVIDIA wants Aerial running on as many radios as possible, because every radio that runs Aerial sells more GPUs. And think about which radio serves that goal best. ➟ Nokia's interest is keeping its stack proprietary, its moat intact. ➟ NVIDIA's + USA's gov interest is the opposite: open, programmable, multi-vendor, maximum adoption. The radio that fits THAT vision isn't the locked box. It's the open American 64T64R that already proved it runs on Aerial in a fully open-source stack. (That's my read of NVIDIA's incentives, based on its documented pattern, not something NVIDIA has stated about AMPG). Now, what the blog actually says. US operators spent $240 BILLION on spectrum, and Massive MIMO, the tool meant to exploit it, underdelivers in the field because the smart algorithms never fit on traditional chips. NVIDIA's fix: run the network on GPUs, and compute stops being the bottleneck. Their published math: the star scenario is a 64T64R MU-MIMO deployment, where AI beamforming delivers up to 1.62x more throughput. Field-proven: a SoftBank trial hit ~3x spectral efficiency outdoors on a GPU-based AI-RAN platform. Plus two bonuses that should sound familiar: ISAC, which "turns a radio network into a ubiquitous radar system" (drone detection, meet the $70B counter-UAS budget). And edge monetization: spare GPU compute at the tower hosting paid AI inference. Edge AI, on the RAN. Then the sentence everyone reads past: rigid ASICs "freeze algorithms in silicon," while GPUs let AI keep evolving in software. That's a shot at the incumbent model. Nokia and Ericsson built empires on locked, proprietary boxes nobody else can improve. The whole AI-native argument (open, programmable, software-first) points AWAY from the model they defend. And that's exactly why Washington built what it built: the DoD funds Open6G at Northeastern (an open American foundation for 6G) and backs OCUDU (open-source code to replace proprietary RAN software). The NTIA put $21.7M into VALOR, a lab whose stated mission is letting NEW entrants validate their tech. Open, American, programmable. That's the lane the checks keep funding. Where does AMPG sit? It's the only American company with a 64T64R Massive MIMO O-RAN radio. The exact configuration NVIDIA's blog builds its case on. And OPEN, built for the multi-vendor world Washington is funding. Not theoretical. ➟ May, inside the DoD-funded hub, AMPG's radio was the body of that world-first open-source demo, running NVIDIA's Aerial. ➟ June: the only 64T64R at that configuration at the O-RAN Global PlugFest. ➟ July 8, one day after NVIDIA's blog: the only Massive MIMO radio in live demos at VALOR, the NVIDIA-powered, NTIA-funded lab, in front of ~120 AI-RAN Alliance members. The blog published the theory. The American open radio demonstrated it live the next day. Why does this matter more for AMPG than anyone? For Nokia, AI-RAN is one initiative among many, already priced into tens of billions of market cap. For a sub-$200M company whose flagship product IS this category, the platform leader publishing the economics of your exact product class is the whole thesis. The giants got a mention. AMPG got a market thesis. Not financial advice. I'm long $AMPG. DYOR. 📡

$AMPG has a clear path to $1b in revenue by 2030 and it doesn't need much market share to get there. I've been through the O/AI RAN opportunity but let's dig into what Private 5G means to $AMPG when it comes to revenue. The Private 5G numbers are just as impressive as the AI/O-RAN numbers. In 2025 the private 5G market stood at just $5 billion. At the projected 40% CAGR the TAM is sitting at almost $30 billion by 2030. Let's say AmpliTech takes a small piece of the pie, like we assumed with the AI/O-RAN numbers. At just 1% market penetration $AMPG is looking at $300 million in revenue by 2030 from Private 5G alone. This is comparable to the opportunity in O/AI-RAN so with just 1% in each they have a $600 million opportunity. Once again, this is driven by another area of fast growing future tech that AmpliTech is positioned to be a part of, ROBOTICS. The growth in the forecast period is attributed to increasing investments in robotics and smart factories, the rising adoption of autonomous systems and a growing focus on cybersecurity for private networks. These smart factories and autonomous robots can't run on traditional WiFi, they need Private 5g. This is the deep tech of the future and $AMPG is at the center of the connectivity layer once again. Just like Edge AI. Just like Quantum Computing Just like LEO constellations. Now we find ourselves at the heart of the robotics buildout through Private 5G connectivity. Add up the numbers from O-RAN, Private 5G, LNA's for quantum computing and LEO satellites etc and there's a clear path to $1b in revenue by 2030. Just a 1.5% capture of the O-RAN and Private 5G and they can hit $900m. The best part is AmpliTech is ready to deploy these Private 5g systems TODAY.

$AMPG just announced a $10m buyback but there must be a reason why the CEO is so confident in doing this when the market cap is only $150m. It's clear the company has turned a corner on 5G deployments but there's plenty more opportunity on the horizon. While O-RAN is AmpliTech's bread and butter today the CEO was recently taking about their private 5G deployments on LinkedIn. Seems like he doesn't want the market to overlook or lose sight of this opportunity. Yes you can use private 5G for residential and office buildings but there's another use case where traditional WiFi is not an option. That's autonomous robots, the next wave of industrial automation. Traditional Wi-Fi cannot deliver the ultra-low latency, reliability, and secure connectivity that fleets of autonomous robots demand in busy factories and warehouses. Private 5G changes everything by providing dedicated spectrum, seamless mobility, and network slicing that collaborative robots operate faster with near-zero downtime. In Fawad's LinkedIn post he pointed out that "AMPG delivers complete end-to-end private 5G solutions, from O-RAN certified Massive MIMO radios and Native DAS systems to portable Network-in-a-Box deployments." Their proprietary RF components and low-noise amplifiers give these networks superior signal quality and range that robotics applications truly need. $AMPG is ready to deploy TODAY in smart manufacturing, logistics hubs, and beyond. The private 5G market is exploding with massive CAGR while robotics adoption surges creating huge tailwinds for a vertically integrated US player like Amplitech. This setup makes $AMPG a compelling way to ride both the 5G/6G infrastructure buildout and also the robotics revolution that traditional networks today can't handle. So while Fawad might have confidence in the O-RAN buildout there's always the chance there's something else coming that's not on the markets radar today.

$AMPG grabbed more today 🧑‍🚀

Yes, it's true. Northeastern University’s May 2026 demo was the first open-source Massive MIMO AI-RAN prototype. AMPG’s 64T64R O-RAN Category B radio unit was the central hardware platform, integrated with NVIDIA AI Aerial (GPU-accelerated L1/L2 on NVIDIA GPUs) and OpenAirInterface software. This validates AMPG’s radio hardware interoperability in a real AI-native Open RAN setup with NVIDIA’s tech. Their radio was successfully used alongside NVIDIA AI Aerial.