If you needed a single phrase to sum up ASUS at Computex 2026, it would probably be this: “PCs built for your own AI agents, not just your apps.” With the new ProArt P16, P14, and a compact ProArt Mini PC, ASUS is betting that the next big PC upgrade cycle won’t be driven by frame rates or thinness alone, but by how well your machine can run large models, local agents, and heavy creative workloads directly on the desk in front of you.

This isn’t just another spec bump announcement. It is ASUS planting a flag in what looks like the first real generation of “personal AI workstations” – Windows PCs designed from the ground up around NVIDIA’s new RTX Spark superchip. And if you are a creator, developer, or anyone who spends their day bouncing between Premiere timelines, Blender scenes, and AI tools, this launch matters a lot more than the usual Computex noise.

At the heart of everything is NVIDIA RTX Spark: a new superchip that effectively fuses an NVIDIA Blackwell RTX GPU with 6,144 CUDA cores and fifth-generation Tensor Cores for FP4 precision to a 20-core NVIDIA Grace CPU over NVLink-C2C. In plain English, ASUS is shipping laptops and a mini PC where CPU and GPU share a unified memory pool – up to 128GB – and can push up to 1 petaflop of AI compute, all inside relatively slim or compact form factors. That unified memory model is critical because it removes one of the classic pain points for AI and 3D work: being bottlenecked by VRAM when you are trying to load big scenes or large models.

ASUS is very explicit about what that performance is for. According to the company, these ProArt machines can render 90GB-plus 3D scenes, edit 12K 4:2:2 video, generate 4K AI video, and run 120 billion parameter large language models with up to a 1 million token context window, all locally. On top of that, ASUS says you can still play AAA titles at 1440p at over 100 fps, which is a nice reminder that this much compute power doesn’t magically stop being good for games just because the marketing around it says “creator” and “AI.”

But where the messaging really stands out is the repeated phrase “purpose-built for personal agents.” NVIDIA and ASUS are both leaning into a vision where the PC is less about your direct interaction with apps and more about you orchestrating a fleet of local agents: one handling research, another helping script or storyboard, another tracking project assets, all running on-device. That is where the unified memory, petaflop-level compute, and Grace CPU actually matter – you’re not just running a single model, you are juggling multiple agents, tools, and pipelines at once.

On the laptop side, the new ProArt P16 (H7607) and P14 (H7407) are ASUS’s attempt to reconcile “AI workstation” with “machine you can actually carry.” ASUS says the new P16 is 13 percent thinner and 16 percent lighter than the previous ProArt P16 H7606, but still retains an all-day battery backed by up to a 99.9Wh pack. The chassis is CNC-machined, with a clear emphasis on premium feel, rigid build, and thermals that can cope with this much silicon in a slim shell.

The displays are classic ProArt territory, but tuned for the AI era. Both laptops use ASUS Lumina Pro OLED panels with Delta E under 1, promising color accuracy tight enough for grading and print work. The P16 can go up to a 4K 120 Hz variable refresh rate panel with NVIDIA G-Sync, while the P14 tops out at 3K, which should be more than enough for a 14-inch canvas. Peak brightness hits up to 1,600 nits, and ASUS adds anti-reflective coatings to keep the screen usable in mixed lighting and on-location shoots.

Design-wise, ASUS is pushing what it calls an “AI-era” aesthetic, shipping these in Nano Black and Neo White finishes with a smooth, tactile coating and anti-smudge treatment. It is very ProArt: understated, professional, and meant to look at home in a studio, edit suite, or a conference room rather than screaming “gaming laptop.” Underneath, ultra-thin thermal modules try to maintain the balancing act between performance and noise in an envelope creators actually want to use on the go.

Beyond raw compute, ASUS is layering in quality-of-life features that matter in daily creative work: a large haptic touchpad, a focus on responsive multitasking, and deeper integration with what Microsoft is calling “AI-enhanced Windows experiences.” Think of the P16 and P14 as ASUS’s answer to the question, “What does a future-proof creative laptop look like if you assume AI agents are just part of the workflow?”

Then there is the ProArt Mini PC, which is arguably the most interesting part of the announcement for studios and small teams. Despite a compact 150 x 150 x 51 mm chassis, ASUS claims it can deliver up to 1 petaflop of AI performance and the same 128GB unified memory configuration you see in the laptops. The idea is simple: give creators, developers, and AI enthusiasts a desktop-grade AI engine that fits under a monitor or behind a display, without the usual noise and footprint of a full tower.

The thermal design is rated for up to 140W of headroom, which suggests sustained load is a key use case: long training runs, continuous inference pipelines, or hours of rendering. Connectivity is geared toward serious work: 10GbE networking, M.2 PCIe Gen 5 x4 expansion, and high-speed storage options that can keep up with large datasets, video libraries, and massive project folders. In practice, you can imagine this box as an “edge AI node” tucked into a studio, gallery, or production office – something that quietly runs your models, powers your agents, and offloads the heavy lifting from your primary laptop.

ASUS also clearly understands that hardware alone does not sell “AI-native” workflows. On the software side, the ProArt ecosystem is leaning into local generative AI and agents through its own tools as well as partner integrations. ProArt Creator Hub acts as a central console that intelligently allocates system resources, tunes performance profiles, and gives creators a single place to monitor what the machine is doing. On top of that, ASUS is introducing MuseTree and StoryCube, two AI-assisted tools aimed at modern content production, though full details on those experiences will almost certainly evolve between announcement and launch.

The ecosystem piece matters because ASUS is trying to meet creators where they already live. These machines are optimized for more than 1,000 accelerated apps and games, including flagship names like Adobe’s suite, Blackmagic Design’s DaVinci Resolve, Blender, CapCut, ComfyUI, and OTOY’s tools. Adobe, notably, is “re-architecting” Photoshop and Premiere for RTX Spark, promising up to 2x faster AI and graphics performance on this platform compared to previous-gen hardware. Buyers also get a three-month Adobe Creative Cloud subscription out of the box, along with trials or bundles for apps like Goodnotes and GoPro Cloud, which is very on brand for ASUS’s ProArt play.

Beyond laptops and the mini PC, ASUS is using this launch to remind everyone that ProArt is now a full ecosystem. Displays, motherboards, cases, and peripherals are all being framed as part of a coherent “AI-ready” creator stack, with the new Neo White colorway extending the aesthetic across more devices. In other words, this is ASUS pitching itself as a one-stop shop for a studio that wants to build an AI-native workflow without juggling too many different vendors and design languages.

All of this is arriving at a particularly interesting moment in the PC industry. After years of incremental updates, we are watching a genuine architectural shift: PCs are transitioning from being primarily CPU and RAM-bound to AI and accelerator-centric systems driven by unified memory, specialized cores, and new software stacks. NVIDIA’s RTX Spark is one answer to that shift, just as Qualcomm’s X Elite and Apple’s M-series chips are in their own ecosystems, but the positioning here – “personal agents,” high context LLMs, large local models – feels aimed squarely at early adopters who want to own their AI workflows instead of renting them from the cloud.

For creators and developers, the big question is always going to be: “What does this unlock that I cannot do today?” The promise of running 120B parameter models locally with up to 1M tokens of context is not just marketing; it hits directly at things like long-form video analysis, large project-aware code assistants, or local research agents that can digest your entire knowledge base without shipping proprietary data off to external servers. Paired with 90GB-plus 3D scenes and 12K video editing support, this starts to look like hardware built for the next decade, not just the next minor version of your editing suite.

Of course, the usual caveats apply. Availability is pegged for fall 2026 in select regions, and ASUS is not yet talking detailed configurations, SKUs, or pricing. As with any cutting-edge hardware, the real-world experience is going to depend heavily on thermals, driver maturity, and how quickly software developers ship optimized builds for RTX Spark. You will also need to think about whether your workflow is ready to take advantage of local agents and large models, or whether you are still more comfortable leaning on cloud-based AI for the next year or two.

But taken as a whole, ASUS’s ProArt P16, P14, and ProArt Mini PC announcements at Computex 2026 feel like a clear signal of where high-end Windows PCs are heading. These are not just faster laptops and a nicer mini PC; they are reference designs for what an AI-first creative stack looks like when local compute is treated as a core requirement, not an optional accelerator. If you are a creator or developer trying to future-proof your next machine, this is the kind of hardware you will want to keep on your radar as it moves from Computex stage demos to real-world desks later this year.