ASUS opened global pre-orders for its ProArt P16 and ProArt P14 laptops on October 7, 2026, putting the first wave of NVIDIA RTX Spark-powered creator machines into buyers’ hands days before launch. A third system, the ProArt GR1X Mini PC, joins the lineup as a desktop alternative built around the same chip. All three are pitched around one idea: run serious AI agents on the machine itself, not in a data center somewhere else.
The timing matters. ASUS says the first US units are scheduled to land on October 16, 2026, just over a week after pre-orders opened. That is a tight window for a launch built around a new NVIDIA superchip platform, and it puts ASUS ahead of several rivals that have only recently announced their own RTX Spark hardware.
What ASUS Just Announced
The announcement covers three distinct products. The ProArt P16 (model H7607) is a 16-inch laptop built for larger creative workloads. The ProArt P14 (model H7407) is its 14-inch sibling, aimed at people who want the same local AI muscle in a lighter frame. The ProArt GR1X Mini PC rounds out the trio as a desktop box for creators who want a stationary workstation rather than a laptop.
All three run on the NVIDIA RTX Spark platform, the same superchip architecture that has been showing up across the PC industry over the past several months. ASUS frames the launch around what it calls local agentic AI: software agents that plan tasks, write code, generate content, and move across applications without sending every request to a cloud server. That framing lines up with how ASUS describes the GR1X on its own product page, where the company states plainly that the system natively runs NVIDIA CUDA software and features the NVIDIA RTX Spark Superchip, which ASUS says is purpose-built from the ground up to reinvent the native Windows PC experience for the era of advanced, local personal AI agents, according to the official ASUS product page.
Inside the ProArt P16, P14, and GR1X Lineup
Each machine shares the same core ceiling: up to 128GB of unified memory and up to 1 petaflop of AI performance, with ASUS specifying FP4 AI compute as the measurement basis. That shared memory pool is the detail that separates this generation of AI laptops from last year’s NPU-equipped machines, most of which topped out well under 100GB and leaned on separate memory pools for CPU, GPU, and AI workloads.
The P16 is the largest of the three, built around a 16-inch display. The P14 trims the footprint to 14 inches for a lighter travel setup. The GR1X drops the display entirely and ships as a compact mini PC, giving studios and freelancers a desktop option that still runs the full RTX Spark feature set. ASUS has not published a breakdown of exact weight, battery capacity, or port selection for any of the three models in the material reviewed for this story, so those remain open questions heading into the October 16 US arrival date.
| Model | Form Factor | Max Unified Memory | AI Performance | Model Support |
|---|---|---|---|---|
| ProArt P16 (H7607) | 16-inch laptop | Up to 128GB | Up to 1 petaflop (FP4) | Up to 120B-parameter LLMs, 1M-token context |
| ProArt P14 (H7407) | 14-inch laptop | Up to 128GB | Up to 1 petaflop (FP4) | Up to 120B-parameter LLMs, 1M-token context |
| ProArt GR1X Mini PC | Desktop mini PC | Up to 128GB | Up to 1 petaflop (FP4) | Up to 120B-parameter LLMs, 1M-token context |
Pricing for all three systems was not confirmed in official ASUS material at the time of publication. Readers should treat any dollar figure circulating online as provisional until ASUS lists official US retail pricing on its own storefront.
NVIDIA RTX Spark: The Chip Making Local AI Agents Possible
NVIDIA RTX Spark is the common thread running through this launch and through a wave of competing hardware that has shipped over the past two months. The chip pairs RTX graphics with a large unified memory pool, which lets a laptop or mini PC hold an entire large language model in memory instead of swapping data between separate CPU and GPU pools. ASUS describes the approach on its RTX Spark product page, stating that the platform is designed to run local AI agents that can plan tasks, generate content, write code, and work across applications, according to ASUS’s official RTX Spark PC page.
That same page pitches the practical upside for creators and developers: the ability to run and retry AI agents locally without token limits, turning variable cloud usage into controllable computing capacity while reserving the cloud for larger workloads, per ASUS. For anyone who has watched a cloud AI subscription bill climb month over month, that pitch is the whole sales argument in one sentence.
ASUS is not the first company to put RTX Spark into a shipping product. Microsoft’s Surface Laptop Ultra launched with the same chip and the same 128GB memory ceiling and 6,144 CUDA core count that diagnostic tools have already confirmed for RTX Spark hardware. HP briefly listed its own RTX Spark laptop at retail before pulling the page, a moment shattered.io covered when it surfaced at $2,999.99. Hardware utilities have already caught up with the platform too, with CPU-Z and HWMonitor adding RTX Spark core-count tracking ahead of this ASUS launch.
Pre-Order Timeline and US Availability
ASUS opened global pre-orders for the ProArt P16 and ProArt P14 on October 7, 2026. The company has targeted October 16, 2026 for the first units to reach the United States, a nine-day gap between order and delivery that is short by PC industry standards for a brand-new chip platform.
The GR1X Mini PC was announced alongside the two laptops, though the fact sheet behind this launch does not specify a separate pre-order date for the desktop unit. Buyers interested in the mini PC form factor should expect ASUS to clarify availability closer to the October 16 laptop ship date, rather than assume identical timing across all three products.
Nine days is also a narrow runway for reviewers. Expect the first independent benchmark numbers, battery tests, and thermal readings to land right around the US arrival date rather than before it, since RTX Spark review units have historically reached outlets close to shipping rather than weeks in advance.
What “Local Agentic AI” Actually Means for ProArt Buyers
ASUS leans hard on the phrase local agentic AI, and it is worth unpacking what that buys a working creator in practice. The company states that RTX Spark is designed to run local AI agents that can plan tasks, generate content, write code, and work across applications, per the ASUS RTX Spark page. That is a step beyond the chatbot-in-a-sidebar model most laptop buyers have gotten used to.
The practical difference comes down to memory and model size. ASUS states that the ProArt P16, with up to 128GB of unified memory, supports LLMs with up to 120 billion parameters and a 1-million-token context window, enabling AI agents to reason, plan, and orchestrate complex creative workflows entirely on device, according to the ProArt P16 product page. A 1-million-token context window is large enough to hold an entire feature-length screenplay, a full codebase, or hours of transcribed video notes in a single working session, without the model losing track of earlier material.
Cross-application agent behavior is the other half of the pitch. ASUS says the systems are designed to run local AI agents for tasks spanning planning, content generation, coding, and workflows that cross between separate creative applications, rather than confining the agent to a single app window.
How RTX Spark Stacks Up Against the Competition
ASUS is entering a field that already has several occupants. NVIDIA’s own DGX Spark desktop line has been on sale with a tiered memory approach, which shattered.io previously detailed when NVIDIA split DGX Spark pricing between a 64GB and a 128GB configuration. AMD has pushed back on the RTX Spark narrative directly, publishing Gorgon Halo benchmark numbers aimed at the same local-AI buyer, a response shattered.io covered when AMD’s own benchmarks landed against RTX Spark earlier this year.
The table below lines up what has been confirmed about ASUS ProArt against other RTX Spark-class systems already covered on this site, using only figures each outlet has published.
| System | OEM | Chip Platform | Max Memory | Status |
|---|---|---|---|---|
| ProArt P16 / P14 | ASUS | NVIDIA RTX Spark | Up to 128GB | Pre-order since Oct. 7, 2026; US units Oct. 16 |
| ProArt GR1X Mini PC | ASUS | NVIDIA RTX Spark | Up to 128GB | Announced alongside P16/P14 |
| Surface Laptop Ultra | Microsoft | NVIDIA RTX Spark | 128GB | Shipping; 6,144 CUDA cores confirmed |
| RTX Spark laptop | HP | NVIDIA RTX Spark | Not disclosed | Briefly listed, then pulled from retail |
| DGX Spark | NVIDIA | NVIDIA RTX Spark | 64GB / 128GB tiers | Shipping desktop reference design |
The pattern is consistent across every entry in that table: the memory ceiling is the number every OEM leads with, because it is the figure that decides whether a given model actually fits on the device. A 128GB system can run a 120-billion-parameter model locally. A 32GB or 64GB system generally cannot, no matter how fast its processor is on paper.
The Creative Workload Case: Video, 3D, and Local LLMs
ASUS is not pitching ProArt purely as an AI agent box. The company lists a set of creative benchmarks it says the RTX Spark platform can handle: rendering 3D scenes above 90GB in size, editing 12K video at 4:2:2 chroma sampling, generating 4K AI video, and running 120-billion-parameter language models, all locally rather than through a cloud render farm.
That combination targets a specific buyer: a video editor, 3D artist, or motion designer who currently bounces files between a local workstation and a cloud GPU rental service for the heaviest jobs. If the 90GB scene and 12K edit claims hold up under independent testing, ProArt would let that work stay on one machine instead of splitting across two systems and two subscription bills.
Software support underpins all of it. ASUS says RTX Spark provides native CUDA acceleration across more than 1,000 accelerated apps, tools, and games, which matters because CUDA support is what determines whether a given creative tool can actually use the GPU’s full throughput rather than falling back to a slower generic rendering path.
Gaming Claims and the 1440p/100fps Question
ASUS also makes a gaming claim for the ProArt lineup, stating the systems can play AAA games at 1440p resolution above 100 frames per second. That is a notable claim for machines marketed primarily at creative professionals rather than gamers, and it suggests ASUS wants ProArt positioned as a do-everything workstation rather than a narrow content-creation tool.
No independent benchmark has verified that 1440p/100fps figure yet, since review units were not broadly available at the time ASUS opened pre-orders. Readers should treat it as a manufacturer claim until outlets with hands-on test units publish their own frame-rate numbers after the October 16 US arrival date.
A Short History of ASUS and NVIDIA’s Local-AI Push
ASUS has spent the past year building out AI-branded PC lines well before RTX Spark existed, including earlier Copilot-class laptops built around NPUs rather than full discrete GPU memory pools. Those machines topped out at AI workloads measured in tens of TOPS, a very different scale from the 1 petaflop figure ASUS now quotes for RTX Spark.
NVIDIA’s unified-memory bet goes back further, through its DGX line of AI appliances built for data centers and research labs. RTX Spark essentially shrinks that unified-memory philosophy down into a laptop-sized and desktop-mini-PC-sized package, which is why ASUS, Microsoft, HP, and others have all converged on similar 128GB configurations within months of each other. The whole PC industry appears to be treating 128GB of unified memory as the threshold needed to run genuinely large local models, rather than a nice-to-have spec.
What Industry Statements Reveal About the Strategy
ASUS’s own product materials lay out the company’s thinking more clearly than any outside commentary could. On the question of whether the GR1X can natively run the full Windows experience for advanced, local personal AI agents, ASUS’s answer on its own product page is simply: “Yes,” according to the ASUS ProArt GR1X page.
That same page goes further, stating that the GR1X natively runs NVIDIA CUDA software and features the NVIDIA RTX Spark Superchip, which ASUS describes as purpose-built from the ground up to reinvent the native Windows PC experience for the era of advanced, local personal AI agents, per ASUS.
On the laptop side, ASUS frames the pitch in economic terms as much as technical ones, stating that the platform lets users run and retry AI agents locally without token limits, turning variable cloud usage into controllable computing capacity while reserving the cloud for larger workloads, according to ASUS’s RTX Spark PC page. The company adds that RTX Spark is designed to run local AI agents that can plan tasks, generate content, write code, and work across applications, per the same source. And on the ProArt P16 specifically, ASUS states that up to 128GB of unified memory supports LLMs with up to 120 billion parameters and a 1-million-token context window, enabling AI agents to reason, plan, and orchestrate complex creative workflows entirely on device, according to the ProArt P16 page.
Read together, those four statements describe a company betting that the next wave of AI PC demand is not about chatbots at all. It is about cost control: moving recurring cloud-agent spend onto a one-time hardware purchase.
Market Impact: Why OEMs Are Racing to Ship RTX Spark Hardware
The speed of this rollout says a lot about how NVIDIA and its partners view the window of opportunity. ASUS moved from global pre-orders to US units in nine days. Microsoft, HP, Lenovo, and Acer have all shipped or listed RTX Spark hardware within a few months of each other, based on shattered.io’s ongoing coverage of the platform’s rollout. That kind of OEM pile-on usually signals one of two things: either NVIDIA is allocating chip supply broadly enough that nobody wants to be last, or the AI PC category has shifted from a marketing checkbox to something buyers are actually asking for by name at the point of sale.
For ASUS specifically, the ProArt branding signals a deliberate choice to chase the professional creator market rather than the mainstream consumer laptop buyer. That is a narrower audience than a general-purpose AI PC push, but it is also an audience more willing to pay a premium for a machine that genuinely replaces a cloud GPU rental bill. If the 90GB scene rendering and 12K video editing claims hold up, ProArt competes less with other laptops and more with a monthly cloud compute invoice.
The unanswered variable is price. Until ASUS publishes official US retail numbers, it is impossible to say whether ProArt undercuts the DGX Spark desktop tiers or the Surface Laptop Ultra, or whether ASUS is positioning these systems above both. That pricing gap is the single biggest open question hanging over the October 16 launch.
What’s Still Unconfirmed
A few details are worth flagging explicitly, because they have not been confirmed in official ASUS material reviewed for this story. US retail pricing for the P16, P14, and GR1X had not been published by ASUS at the time of publication. No named ASUS spokesperson or executive has been quoted on the record about the launch in the material available. And beyond the memory, AI-performance, and model-support figures above, ASUS has not detailed additional RTX Spark clock speeds, GPU core counts, or power limits specific to these three products.
Readers should treat any price figure, spokesperson quote, or detailed chip spec circulating elsewhere about this specific launch as unverified until ASUS or NVIDIA publishes it directly.
Predictions: Where RTX Spark Hardware Goes Next
A few trends look likely to play out over the next two quarters, based on how this launch and the broader RTX Spark rollout have unfolded so far.
- Expect more OEMs to announce RTX Spark systems before the end of 2026, following the pattern set by ASUS, Microsoft, HP, Lenovo, and Acer.
- Independent reviewers will likely challenge ASUS’s 1440p/100fps gaming claim and the 90GB scene/12K video editing figures within weeks of the October 16 US arrival, since those are the easiest claims to test directly.
- Pricing pressure will build fast. With DGX Spark, Surface Laptop Ultra, and now three ASUS ProArt models all competing on the same 128GB memory ceiling, the deciding factor for buyers will shift to price per configuration rather than raw spec sheets.
- The 120-billion-parameter, 1-million-token local model benchmark is likely to become the new marketing yardstick for AI PCs through 2027, replacing the TOPS-per-NPU figures that defined the previous generation.
- Expect ASUS to eventually confirm GR1X pre-order timing and pricing separately from the P16/P14 laptops, given the fact sheet behind this launch does not yet align the three products on a single calendar date.
Related
- Surface Laptop Ultra Specs Confirm 128GB RAM, 6,144 Cores
- CPU-Z, HWMonitor Track RTX Spark’s 6,144 Cores
- HP Briefly Lists RTX Spark Laptops at $2,999.99
- RTX Spark Ships to 6 PC Makers, Erases 4.2% Drop
- Nvidia Splits DGX Spark: $4,999 vs $6,950 128GB
Frequently Asked Questions
When did ASUS open pre-orders for the ProArt RTX Spark PCs?
Global pre-orders for the ProArt P16 and ProArt P14 opened on October 7, 2026. ASUS has scheduled the first US units to arrive on October 16, 2026.
What are the three ASUS ProArt RTX Spark products?
ASUS announced the ProArt P16 (model H7607), a 16-inch laptop, the ProArt P14 (model H7407), a 14-inch laptop, and the ProArt GR1X Mini PC, a desktop system. All three run on the NVIDIA RTX Spark platform.
How much memory do the ProArt RTX Spark PCs support?
ASUS states all three systems support up to 128GB of unified memory, which the company says is enough to run large language models with up to 120 billion parameters and a 1-million-token context window locally.
How much do the ASUS ProArt RTX Spark PCs cost?
ASUS had not published official US retail pricing for the P16, P14, or GR1X at the time of this report. Any price figure circulating elsewhere should be treated as unconfirmed until ASUS lists official pricing.
Can the ProArt RTX Spark PCs play games?
ASUS claims the systems can run AAA games at 1440p resolution above 100 frames per second. That figure comes from ASUS and has not yet been independently verified by third-party reviewers.
What does “local agentic AI” mean on these machines?
ASUS uses the term to describe AI agents that run entirely on the device rather than through a cloud service, capable of planning tasks, generating content, writing code, and working across multiple applications without per-token cloud costs.
How does RTX Spark compare to NVIDIA’s DGX Spark desktop?
Both run on the same underlying NVIDIA RTX Spark platform. DGX Spark ships as a dedicated desktop appliance in 64GB and 128GB memory tiers, while the ASUS ProArt lineup packages the same chip into laptop and mini-PC form factors aimed at creative professionals.
Is the ASUS ProArt GR1X available for pre-order yet?
ASUS announced the GR1X Mini PC alongside the P16 and P14 laptops, but the available information does not specify a separate confirmed pre-order date for the desktop unit.
