Ferbin

Haha this is me. Never applied for a job in my life. I just upskill and sell myself to people who either hire me or refer me to people hiring. Too lazy to apply for a job😆.

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. 📡

college