ASUS told investors on August 12, 2026 that the initial production run of its RTX Spark laptops and mini PCs had been fully claimed by channel partners before a single unit reached a retail shelf, and that the company is now placing additional orders with Nvidia to cover a Q4 2026 launch. That demand builds directly on Nvidia’s own rollout: at Computex 2026 in June, Nvidia named six RTX Spark laptops as its official launch lineup, and the disclosure, made during ASUS’s quarterly investor briefing, is the clearest signal yet that Nvidia’s new Grace Blackwell laptop platform has outpaced what its first wave of manufacturing partners expected — and it comes just weeks after Nvidia’s original RTX Spark reveal rattled Intel and AMD’s stock prices.
MSI has not issued a matching public statement, but it appears alongside ASUS on Nvidia’s initial RTX Spark OEM roster, and outlets covering the platform’s rollout list an MSI Prestige N16 Flip AI+ laptop and an EdgeMesa N AI+ mini PC among the first wave of hardware. What’s changed since the original sellout headlines is the scale of the reorder signal: ASUS isn’t just describing brisk pre-launch interest anymore, it’s telling shareholders it has gone back to Nvidia for more silicon.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What ASUS Told Investors About RTX Spark Demand
During the August 12 investor conference, ASUS management said the company’s planned initial quantities of RTX Spark products were fully pre-ordered by channel partners well ahead of the fall 2026 launch window. Rather than treating that as a one-time bottleneck, ASUS said it is placing additional orders to keep pace with demand, with its own RTX Spark hardware still targeted for a Q4 2026 debut. That is a notably specific commitment: ASUS is one of the only partners to attach even a rough quarter to its launch plans, while most competing OEMs have stuck to the vaguer “this fall” language Nvidia used at the platform’s unveiling.
The company did not disclose unit volumes for either the original allocation or the new order, and it stopped short of confirming that the additional order was specifically for Nvidia’s GB10 Grace Blackwell silicon versus other RTX Spark-adjacent components. Still, the sequence — sell out, then go back for more before launch day — is unusual for a first-generation AI PC category that most of the industry expected to ramp slowly.
What Is Nvidia RTX Spark, Exactly?
RTX Spark is Nvidia’s Arm-based system-on-chip platform, built around what Nvidia has disclosed is the N1X superchip — a design packing roughly 70 billion transistors on TSMC’s 3nm process — pairing a 20-core Grace CPU with a Blackwell-generation RTX GPU carrying 6,144 CUDA cores, unified with up to 128GB of LPDDR5X memory on a single package, per specs Microsoft detailed in May 2026. Nvidia, which announced the platform that same month, is offering the chip in two top-tier configurations: a higher-end design rated at roughly 1 petaflop of AI performance and a lower-tier variant closer to 400 teraflops, giving OEMs room to build both premium creator laptops and more affordable AI-capable machines off the same underlying architecture.
The pitch is straightforward: instead of routing large-language-model or diffusion-model workloads to the cloud, RTX Spark hardware is meant to run them locally, with the unified memory pool acting as the bottleneck-breaker that lets a laptop or mini PC handle models that would otherwise choke on a conventional GPU’s limited VRAM. That framing places RTX Spark squarely in the same conversation as Apple’s unified-memory M-series chips, though Nvidia is betting its CUDA software stack gives it an edge with AI developers specifically.
RTX Spark vs. DGX Spark: Two Names, Two Different Products
Nvidia’s naming has created genuine confusion, and it’s worth untangling before assessing what actually sold out. DGX Spark is Nvidia’s own branded “mini AI supercomputer,” a GB10 Grace Blackwell desktop unit sold directly and through partners including the ASUS Ascent GX10, plus configurations from PNY, Dell, HP, Lenovo, Acer, Gigabyte, and MSI, generally priced around $4,499 for comparable GB10 specs. RTX Spark, by contrast, refers to the consumer- and creator-facing laptop and mini-PC platform that ASUS, MSI, Dell, HP, Lenovo, and Microsoft Surface are building into their own product lines for a fall 2026 launch, with Acer and Gigabyte expected to join a second wave.
Both product families trace back to the same Grace Blackwell engineering, but DGX Spark is a fixed Nvidia-defined appliance while RTX Spark is a chip platform that OEMs style into their own designs — ASUS’s ProArt P14, ProArt P16, and ProArt Mini PC among them, alongside Lenovo’s take on the platform. Lenovo and Thurrott detailed that design in September 2026 at IFA, confirming the Yoga Pro 9n will support up to 128GB of LPDDR5X-9400 memory and a 4TB SSD — among the most generously specced RTX Spark configurations disclosed so far. That distinction matters for pricing: the roughly $4,499 figure attached to DGX Spark-class GB10 systems should not be assumed to carry over to RTX Spark laptops, which remain unpriced.
Pricing: What’s Confirmed and What Isn’t
As of a laptop roundup updated August 29, 2026, no manufacturer — ASUS or MSI included — had published an official retail price or exact release date for its consumer RTX Spark laptops. That’s a meaningful gap given that ASUS is already describing its initial allocation as fully claimed and is placing follow-on orders. Buyers currently have no MSRP to weigh against a reorder headline, and outlets tracking the category are explicit that the widely cited DGX Spark-class pricing near $4,499 applies to Nvidia’s mini-supercomputer desktop configurations, not the laptop SKUs ASUS and MSI are preparing.
That pricing vacuum is part of why the reorder disclosure carries weight: channel partners were willing to commit to allocations of hardware whose street price they didn’t yet know, on the strength of the RTX Spark platform’s specs and Nvidia’s brand pull in the AI-PC category alone.
The OEM Lineup Is Widening, Not Shrinking
Nvidia’s original RTX Spark OEM list named six companies at launch: ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI. Acer and Gigabyte have since been confirmed as a second-wave addition, bringing the eventual roster to eight PC makers building around the same Grace Blackwell platform, and Reuters reported in September 2026 that the first RTX Spark AI PCs from Lenovo and Acer specifically are now slated to ship in October 2026 — ahead of ASUS’s own Q4 target. Coverage of the category also points to over 30 laptop models and more than 10 desktop or mini-PC designs expected to carry RTX Spark chips as the lineup fills out through 2026, a scale that is unusually broad for a first-generation Arm-based AI PC platform from Nvidia.
| OEM | Known RTX Spark / GB10 Product(s) | Launch Wave |
|---|---|---|
| ASUS | ProArt P14, ProArt P16, ProArt Mini PC, Ascent GX10 (DGX Spark class) | First wave, targeting Q4 2026 |
| MSI | Prestige N16 Flip AI+, EdgeMesa N AI+ Mini PC | First wave, fall 2026 |
| Dell | XPS 16 (RTX Spark), DGX Spark-class desktop config | First wave, fall 2026 |
| Lenovo | Yoga Pro 9n, DGX Spark-class desktop config | First wave, fall 2026 |
| HP | DGX Spark-class desktop config, laptop TBD | First wave, fall 2026 |
| Microsoft Surface | RTX Spark laptop, model unconfirmed | First wave, fall 2026 |
| Acer | Not yet detailed | Second wave |
| Gigabyte | DGX Spark-class desktop config | Second wave |
PNY also appears among GB10-based DGX Spark partners, though it is not on Nvidia’s separate RTX Spark laptop OEM list, underscoring again that the two product families run on parallel but distinct tracks.
Why This Reorder Signal Is Different From the Original Sellout Story
Nvidia’s RTX Spark platform has already generated one round of sellout headlines this month, with early coverage pointing to strong pre-launch demand across channel partners. What the August 12 investor conference adds is a second, more concrete data point from ASUS itself: not just that initial stock moved fast, but that the company has gone back to place additional orders rather than waiting for a standard restock cycle tied to its Q4 launch window. For a company the size of ASUS, opening a second order round before the first wave has even shipped is a stronger demand signal than pre-order enthusiasm alone, since it implies internal sales forecasting was revised upward mid-cycle.
It also lands amid a broader run of Nvidia hardware news this quarter. AWS committed to roughly 2 million additional Nvidia GPUs tied to Vera CPU deployments, and Nvidia’s Grace Blackwell push has already been cited as a factor in Intel’s Xeon 6+ launch getting overshadowed in investor and press coverage the same week. RTX Spark’s reorder news fits a pattern: every recent Nvidia hardware disclosure this quarter has come with some kind of capacity or demand surprise attached.
How RTX Spark Compares to Rival AI-PC Approaches
RTX Spark’s central bet — unified memory as the unlock for local AI workloads — isn’t a new idea in the PC industry, it’s a race Apple has been running since the M1 Ultra. Apple’s approach uses its own silicon and macOS exclusively, keeping the hardware and software stack fully vertical. Qualcomm’s Snapdragon X series took a similar unified-architecture approach for Windows on Arm but has leaned on efficiency and battery life rather than raw AI throughput as its selling point. Nvidia’s play with RTX Spark is different again: it’s supplying the compute silicon to six-to-eight competing OEMs simultaneously, effectively racing against its own customers’ branding rather than building one flagship device.
| Platform | Architecture | Unified Memory (Max) | Distribution Model |
|---|---|---|---|
| Nvidia RTX Spark | Arm (Grace CPU + Blackwell GPU) | 128GB LPDDR5X | Licensed to 6-8 competing OEMs |
| Nvidia DGX Spark (GB10) | Arm (Grace CPU + Blackwell GPU) | 128GB LPDDR5X | Nvidia-branded + partner desktops, ~$4,499 class |
| Apple M-series (Mac) | Arm (Apple silicon) | Up to 512GB on Mac Studio-class hardware | Exclusive to Apple hardware |
| Qualcomm Snapdragon X | Arm | Vendor-dependent, generally lower ceiling | Licensed to Windows OEMs |
The multi-OEM approach is a double-edged strategy. It lets RTX Spark hardware reach far more retail shelves at once than a single-vendor product ever could, which likely contributed to ASUS’s initial allocation selling through so quickly. But it also means Nvidia’s own manufacturing and packaging capacity for GB10-class Grace Blackwell silicon has to stretch across eight simultaneous product lines rather than one, which is precisely the kind of constraint that produces a reorder disclosure like ASUS’s before launch has even happened.
Market Impact: Reading the Reorder as an Investor Signal
For ASUS shareholders, a fully-claimed initial allocation paired with a follow-on order is generally a reassuring disclosure — it suggests the RTX Spark line will contribute meaningfully to revenue once it actually ships, rather than sitting on shelves. But it also raises the obvious follow-up question investors were reportedly probing on the call: whether Nvidia can supply the additional Grace Blackwell units fast enough to hit ASUS’s stated Q4 2026 target, or whether the second order pushes some inventory into 2027. ASUS did not commit to a firm delivery date for the additional units during the conference.
The disclosure also lands in a stock environment already sensitive to Nvidia’s AI-PC ambitions. Nvidia’s broader RTX lineup has already seen list prices climb through 2026 amid tight GPU and AI-chip supply, and any signal that a new consumer product category is outrunning its own manufacturing plan tends to move sentiment around Nvidia’s OEM partners as much as around Nvidia itself.
Historical Context: Nvidia’s Long Road to a Desktop AI Chip
Nvidia has spent most of the past decade selling AI compute as a data-center product, from the original DGX-1 appliance in 2016 through the H100 and Blackwell-generation data-center GPUs that now anchor its enterprise business. RTX Spark and DGX Spark represent the company’s most direct attempt yet to bring that same Grace Blackwell architecture down to desk-and-lap scale, following smaller-scale experiments like the Jetson embedded line and Nvidia’s earlier RTX-branded laptop GPUs, which never offered anything close to a unified 128GB memory pool. The jump from data-center-only Grace Blackwell to a laptop chip in barely two years is a faster consumer translation than Nvidia has managed with any prior AI architecture generation.
Supply Chain Considerations Nobody Has Confirmed Yet
Notably absent from ASUS’s investor disclosure, and from other coverage of the RTX Spark rollout so far, is any specific comment on what is constraining supply: whether it’s Grace Blackwell die packaging capacity, LPDDR5X memory allocation, or general OEM assembly capacity. Given that 2026 has already seen well-documented DRAM and memory pricing pressure across the PC industry, and that RTX Spark’s entire value proposition rests on a large unified memory pool, memory supply is a reasonable candidate — but it remains unconfirmed by ASUS, MSI, or Nvidia directly, and should be treated as an open question rather than a stated fact.
What Comes Next: Five Things to Watch Before Q4 2026
First, expect ASUS and MSI to publish official RTX Spark pricing before the Q4 2026 window ASUS has already committed to, since channel partners will need firm numbers to plan retail launches. Second, watch whether MSI issues its own investor-facing statement on order volumes now that ASUS has set a public precedent for disclosing reorder activity. Third, Acer and Gigabyte’s product specifics for the second OEM wave are likely to surface over the next one to two months as Nvidia fills out its partner roster. Fourth, keep an eye on whether Nvidia or any OEM acknowledges a specific supply constraint — memory, packaging, or otherwise — since that would materially change how investors read the reorder headline. Fifth, expect competitive responses: if RTX Spark’s laptop allocation genuinely oversold this fast, Apple, Qualcomm, and AMD all have incentive to accelerate their own unified-memory AI PC messaging before the RTX Spark launch window closes out 2026.
None of these are certainties, and ASUS’s own language on the investor call was carefully hedged around quarter-level targets rather than firm dates. But the shift from sold-out headlines to a public reorder disclosure marks a genuine change in how seriously Nvidia’s OEM partners are treating the RTX Spark category just months after its unveiling.
Frequently Asked Questions
What did ASUS actually confirm about RTX Spark orders?
During its August 12, 2026 investor conference, ASUS said its planned initial quantities of RTX Spark products were fully pre-ordered by channel partners and that it is placing additional orders, with a Q4 2026 launch still targeted. The company did not disclose specific unit numbers for either the original or the new order.
Has MSI made a similar statement about RTX Spark reorders?
No. MSI is listed among Nvidia’s RTX Spark OEM partners with products including the Prestige N16 Flip AI+ laptop and EdgeMesa N AI+ mini PC, but as of this report, no equivalent public statement from MSI about sellouts or reorders has surfaced.
What is the difference between RTX Spark and DGX Spark?
DGX Spark is Nvidia’s own branded GB10 Grace Blackwell desktop appliance, sold directly and through partner configurations (including the ASUS Ascent GX10) around $4,499. RTX Spark is the separate laptop and mini-PC chip platform that ASUS, MSI, Dell, HP, Lenovo, and Microsoft Surface are building into their own products for a fall 2026 launch.
How much will RTX Spark laptops cost?
As of late August 2026, no manufacturer had published an official retail price for consumer RTX Spark laptops. The roughly $4,499 figure often cited in coverage applies to DGX Spark-class GB10 desktop configurations, not the laptop SKUs from ASUS and MSI.
Which companies are building RTX Spark hardware?
Nvidia’s initial RTX Spark OEM list includes ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI, with Acer and Gigabyte confirmed as a second wave, bringing the eventual roster to eight PC makers.
What chip powers RTX Spark laptops?
RTX Spark laptops use an Arm-based Nvidia SoC, the N1X superchip, combining a 20-core Grace CPU with a Blackwell-generation RTX GPU featuring 6,144 CUDA cores and up to 128GB of unified LPDDR5X memory, offered in roughly 1-petaflop and 400-teraflop AI performance variants. Windows Central reported in September 2026 that a scaled-down, laptop-only configuration trims that further still, to 18 CPU cores, 5,120 GPU cores, and just 24-32GB of unified memory — a budget tier below the flagship specs quoted elsewhere.
When will RTX Spark laptops actually ship?
Nvidia has described a general fall 2026 launch window for the first wave of RTX Spark hardware. ASUS is the only partner to attach a specific quarter to its own plans, targeting Q4 2026, though it has not committed to an exact ship date for the additional units it is now ordering.
Is the RTX Spark shortage caused by a memory supply issue?
That has not been confirmed by ASUS, MSI, or Nvidia. No source has publicly attributed the sellout or the reorder to a specific supply-chain constraint such as memory, packaging, or assembly capacity, so it should be treated as an open question rather than a confirmed cause.
Related Coverage
- ASUS, MSI RTX Spark Laptops Sell Out: 100K Units [2026]
- Nvidia RTX Spark Sinks Intel, AMD Stock 4.4% [2026]
- OpenAI Buys Macs as $4,699 RTX Spark Sells Out [2026]
- AWS Orders 2M More Nvidia GPUs, Stock Gains $440B [2026]
- DLSS 5 Patched to Run on RTX 4000: 39% FPS Drop [2026]
