ASUS V16 handles local AI, modestly

Good enough for learning and tinkering, not good enough if “local AI” means comfortably running bigger models, longer contexts, or multiple apps at once.

• –ASUS lists the V16 with an RTX 5060 Laptop GPU, 16GB DDR5, and up to 32GB max RAM, so the platform is upgradeable but the base config is still lean for AI work.
• –Ollama’s own FAQ says GPU inference models need to fit fully in VRAM, which makes 8GB a real constraint once you move past small quantized models.
• –NVIDIA’s sizing guidance pegs a 7B model at roughly 14GB in FP16, which is why this machine will be fine for smaller quantized LLMs but not for heavier local inference.
• –Compared with your M1 MacBook Air, the ASUS is the better CUDA/NVIDIA option for local AI experimentation, but the jump is about compatibility and flexibility more than raw headroom.
• –If you buy it, prioritize a RAM upgrade to 32GB and use it as an entry-level local AI box, not as a long-term workstation substitute.