AMD CEO LISA SU HELD A MINI PC ON STAGE THAT RUNS A 235B MODEL AND REPLACES YOUR $440/MONTH AI STACKamd's ryzen ai max+ 395 is the first x86 chip that runs a 200 billion parameter model on one piece of silicon. cpu and gpu share 128gb of unified memory, no separate graphics card neededthe gmktec evo-x2 runs qwen3 235b fully, deepseek v3 comfortably and llama 3.3 70b with headroom. on linux you get 110gb of usable vram out of 128gbamd claimed the chip beat an nvidia rtx 5080 by more than 3x on de

💥AI is no longer confined to the cloud. At Computex, @jackhuynh showcased the AMD Ryzen AI Halo and Ryzen AI Max PRO 400 Series processors. The next wave of innovation will be built, trained, and run locally.As AI evolves from chatbots to intelligent agents, developers and enterprises need the performance, memory capacity, and flexibility to bring these experiences to life directly on the device. ⚡ Ryzen AI Halo enables local AI development with support for models up to 200B parameters, up to 1

AMD on local AI:"this is the smallest AI development system in the world, capable of running models up to 200 billion parameters locally, not connected to anything"128GB of unified memory shared by CPU, GPU, and NPU fits in your hand. Runs a model larger than GPT-3 with no internet$3,999 once. roughly $16 a month in powerCloud equivalent: up to $750 a monthThe question was never whether local AI could compete on powerIt was whether it could compete on price. Now it does

AMD just put a $3,999 AI workstation into something smaller than a hardcover book.The magic starts at 0:12.He holds it up, turns it around, and it looks almost too small to matter.It does matter.Inside is AMD's Ryzen AI Max+ chip, 128GB of unified memory, and enough local memory to run models up to 200B parameters. No cloud round trip. No API call. No private files leaving the machine.That is the part most people miss.Nvidia DGX Spark: $4,699. AMD Ryzen AI Halo: $3,999.One is Linux-only. The oth

AMD just made local 300B-parameter models a desk setup.The magic starts at 0:34.192GB unified memory. Up to 160GB for AI workloads. Multiple agents running in parallel. No cloud. No API bill. No data leaving the building.That last part is the business.One developer made the same bet early with a $2,199 AMD box. He was paying $1,700/month in cloud GPU rentals for work that billed $2,600. Then he moved the workflow local and sold one sentence:"Your data never leaves a machine you control."Month 1: