Both of those happened on the same day, in the same corner of the internet, running comparable work. Prices at the top are climbing, the used enterprise shortcut is failing in public, and the floor keeps quietly dropping. A note on the strangest hardware ladder local AI has had yet.
Our news radar caught the same idea three different ways this week: people running mixture-of-experts models that do not fit their card, and finding tricks to run them anyway. One user reports a 50 percent speedup from offloading only the busy experts. Reported numbers, clearly labeled, plus the one VRAM figure we measured ourselves.
Everyone’s arguing about Wan 2.2. So we ran it locally, honestly, the fast-and-fits 5B, the heavy two-expert 14B, the VRAM wall, and the decode that wedges. Here are the real numbers.
VRAM first, everything else second: what GPU you actually need for local AI in 2026, the memory-crisis market, the value picks, AI appliances like DGX Spark, four real builds by budget, and when renting beats buying.
Pick your GPU and RAM, get honest answers: which AI chat models, image generators, voice cloning, talking avatars, video generation, and full AI companions your machine can actually run, from people who run all of it daily.