The future of AI runs across local devices and the cloud.

@jackhuynh, AMD SVP and GM of the Computing and Graphics Group, explains how Ryzen AI Halo gives developers more room to experiment with models locally.

Watch to hear his vision for what comes next.

Oct 17, 2026 · 2:00 PM UTC

21
18
153
21,799
RepliesRelevant
Replying to @AMD
@AMD @jackhuynh more room to experiment locally is what i want, since i only run small models at home and bigger ones are out of reach. i wonder what a halo box costs compared to a normal gaming pc.
0
0
0
34
Replying to @AMD
@AMD @jackhuynh حي ذا الشوف يا القائد 😍😍
0
0
0
45
Replying to @AMD
@AMD @jackhuynh Yeni Ryzen AI Halo'da Ryzen AI Max+ PRO 495, 192 GB birleşik bellek ve 55 TOPS NPU var. AMD, 300 milyar parametreye kadar modelleri yerelde çalıştırabildiğini söylüyor. Tek sıkıntı şimdilik sadece ABD'de satılıyor, Türkiye'ye gelse bence ciddi ilgi görür.
0
0
0
86
Replying to @AMD
@AMD @jackhuynh Ryzen AI Halo sounds like something I should be painting. The chips keep getting smarter and I still mix my colors by hand.
0
0
0
66
Replying to @AMD
@AMD @jackhuynh Think bigger 🔥🏅
0
0
0
85
Replying to @AMD
@AMD @jackhuynh The desktop inference was never compute-bound, it's memory-bound, and the model either fits or it doesn't.
0
0
0
84
Replying to @AMD
@AMD @jackhuynh Building Luna, my local AI assistant, on AMD with ROCm and llama.cpp. My wall is VRAM, not software. That's why @AIatAMD's Ryzen AI Halo caught my eye. @jackhuynh I'd love to test it. #AMD #ROCm #LocalLLM #Ollama #Linux
0
0
0
2
Replying to @AMD
@AMD @jackhuynh The cloud is dying rapidly.
Adjust quickly or competition will.
0
0
0
82