… Sign in to your XDA account Nvidia and Microsoft recently unveiled the much-awaited N1X and N1 SoCs, now called RTX Spark . These Arm-based chips are meant to reinvent Windows PCs for the AI era, with Nvidia claiming that laptops powered by RTX Spark are purpose-built for agentic AI. …
… This 300-series AI Halo hasn't even launched yet, and the company has already outed what's next with Ryzen AI Max PRO 400 series CPUs. These Zen 5 chips combine RDNA 3.5 graphics with XDNA 2 NPUs to deliver some impressive results, being among the first x86 client chips to run 300B models. …
… Microsoft says the Surface RTX Spark Dev Box delivers up to 1 petaflop of AI compute and features 128GB of unified memory, enabling it to run up to 120B-parameter models locally without cloud GPU instances. …
… AMD is going about it differently Unified platform with quantization AMD has been testing its latest Ryzen AI Max+ chipset using ROCm, Ollama, and Qwen 3.5 models. …
… RDNA 2 chips, including the Steam Deck, get FSR 4.1 in early 2027. FSR 4.1 uses AI upscaling, similar to Nvidia's DLSS, offering a notable leap in visuals and frame rates over FSR 3. …
… We then saw a leak claiming that the trio was planning to pull back the curtain on the N1X and N1 chips, which would elevate Windows on Arm to a whole new level. …
… The CPU side was co-designed with MediaTek, and Nvidia rates it at 1 petaflop of AI compute, enough to run 120B-parameter models locally with up to a million tokens of context. …
… It's impressive, but read between the lines: we're several years into the LLM craze, and if that kind of run can still be newsworthy, it tells you how CUDA-first the training world remains. The broader ecosystem also moves CUDA-first. New models? CUDA first. AI software packages? …
… This means Nvidia can't just make chips bigger or hotter anymore. Instead, DLSS is a cheat code that uses AI to simulate the performance of a 1,000 W GPU using only 450 W of actual electricity. …
… It's a $3,800 AI workstation running Arm Linux with 6,144 CUDA cores the same count as an RTX 5070 and 128GB of unified memory. The machine was designed to serve language models, not render frames, so it kind of went against the entire idea behind that chip. …