The NVIDIA RTX Spark laptop is a Windows-on-Arm system built around one unusual idea: combine a Grace CPU, Blackwell RTX graphics and unified memory in a single platform for local AI, creative work, development and gaming. Lenovo’s first named machines are the Yoga Pro 9n and Yoga 9n 2-in-1, with RTX Spark laptops scheduled to begin arriving in a global October 2026 window.
The Yoga 9n 2-in-1 is the more eye-catching entry point. It is a 16-inch convertible with a 2,880 × 1,800 OLED touchscreen, a 360-degree hinge and up to 64 GB of unified memory. The Yoga Pro 9n takes a more traditional clamshell approach and scales to 128 GB, two SSD slots and up to 4 TB of storage.
What RTX Spark changes in a Windows laptop
RTX Spark is NVIDIA’s Arm-based Windows PC platform. Its central component, identified publicly as N1X, combines three elements that are usually spread across separate parts of a laptop: the Grace CPU, a Blackwell RTX GPU and unified LPDDR5X memory.
Unified memory means the CPU and GPU draw from the same memory pool instead of relying on separate system RAM and graphics memory. For local AI workloads, that matters because large models can require more memory than a typical laptop GPU provides. It does not automatically guarantee a faster result, but it gives the platform a different starting point for running large local models and other memory-hungry applications.
The highest announced RTX Spark configuration reaches 20 Grace CPU cores, 6,144 Blackwell CUDA cores and up to 128 GB of unified memory. A lower configuration is described with 18 CPU cores, 5,120 CUDA cores and 24–32 GB of unified memory.
The other major piece is CUDA. NVIDIA’s software platform gives developers access to libraries and tools used across AI, rendering and scientific computing. That makes RTX Spark more than an Arm laptop with a large memory pool: it is an attempt to bring NVIDIA’s established GPU software ecosystem into thin Windows systems.
That distinction is important when comparing the concept with Apple silicon. Both approaches use Arm-based chips and unified memory, but RTX Spark’s documented differentiators are CUDA, RTX ray tracing, DLSS, Reflex and G-SYNC support. There is no supplied measured result showing that RTX Spark outperforms Apple silicon, so the comparison currently comes down to software ecosystems and intended workloads rather than a clean winner.
Lenovo’s RTX Spark laptops: Yoga Pro 9n vs. Yoga 9n 2-in-1
Lenovo’s two announced designs target different kinds of power users. The Yoga Pro 9n is the performance-oriented clamshell, while the Yoga 9n 2-in-1 adds a touch display, a 360-degree hinge and Yoga Pen Gen 2 support.
A hands-on IFA presentation shows the systems’ physical design, ports, keyboard and convertible hinge. The Yoga Pro 9n includes a full-size SD card reader, while the Yoga 9n 2-in-1 uses a hinge-based soundbar. Those details matter to creators and mobile professionals in a way that a core count alone does not.
| Model | CPU/GPU | Unified memory | Display | Battery | Storage | Weight |
| Lenovo Yoga Pro 9n | Grace CPU, up to 20 cores; integrated Blackwell RTX GPU, up to 6,144 CUDA cores | Up to 128 GB LPDDR5X-9400 | 15.3-inch OLED, 2,560 × 1,600, 48–165 Hz variable refresh, 1,100-nit HDR peak | 92.5 Wh | Two SSD slots, up to 4 TB total; Gen 4 and Gen 5 options | 3.64 lb (1.65 kg) |
| Lenovo Yoga 9n 2-in-1 | Grace CPU, up to 20 cores; integrated Blackwell RTX GPU, up to 6,144 CUDA cores | Up to 64 GB LPDDR5X-9400 | 16-inch OLED touchscreen, 2,880 × 1,800, 30–120 Hz variable refresh, 1,100-nit HDR peak | 99.9 Wh | Up to 2 TB PCIe Gen 4 QLC SSD | 4.28 lb (1.94 kg) |
Both systems use Windows 11, Wi-Fi 7, Bluetooth 5.4 and a 9.2-megapixel MIPI IR camera. The specifications describe premium laptops, but they do not tell us how the systems will perform under sustained AI or gaming workloads. Battery capacity is not battery life, either—a 99.9 Wh cell is a useful data point, not an endurance result.
Local AI, creation and gaming
NVIDIA positions RTX Spark around local AI agents, CUDA development, creative applications, video editing and 1440p gaming. The idea is to run more of the workload on the laptop instead of sending every task to a cloud service. NVIDIA and its partners also describe the platform as suitable for large language models and diffusion-model workflows.
Those are platform goals and company claims, not independent retail-unit test results. A trade-show gaming demonstration shows that the hardware can run a game locally, but it does not establish a broad gaming verdict. NVIDIA has also cited demonstrations exceeding 100 frames per second at 1440p; that figure should be read in the context of NVIDIA’s demonstration rather than as a guarantee for every game or configuration.
This is why RTX Spark laptops are not simply gaming laptops with a different GPU label. Their pitch is closer to a mobile workstation with gaming features: lots of shared memory, CUDA access, RTX acceleration and a design intended for productivity and creation. If you spend your time in CUDA-enabled tools, local AI applications, 3D work or GPU-accelerated creative software, that combination could be more relevant than raw gaming branding.
The trade-off is software maturity. Windows-on-Arm compatibility remains a practical question for buyers who depend on older applications, specialized plug-ins or particular games. The platform’s success will depend not only on the silicon, but also on how well the software stack behaves outside NVIDIA’s showcase workloads.
October launch window, pricing and availability
NVIDIA has identified October 2026 as the global launch window for RTX Spark laptops and compact desktops. That is a platform window, not a confirmed shipping date for every Lenovo configuration. The exact retail date for the Yoga Pro 9n and Yoga 9n 2-in-1 has not been published in the supplied product information.
Official laptop pricing is also undisclosed. Circulating estimates of roughly $1,799 to $2,899 or more are speculative and should not be treated as Lenovo or NVIDIA launch prices. For US buyers, the useful answer is simple: there is no confirmed dollar price to use for a purchase decision yet.
That missing number is more important than usual. RTX Spark is aimed at a premium audience, and the value proposition changes dramatically depending on whether a Yoga system lands near a high-end thin-and-light laptop or a workstation-class machine. Until pricing and retail configurations appear, it is impossible to make a reliable value judgment.
The practical questions buyers should ask
Is RTX Spark mainly for gaming?
No. RTX Spark supports RTX gaming technologies, and NVIDIA has demonstrated 1440p gaming, but the platform is positioned primarily for local AI, creation, development and productivity. A buyer choosing one for gaming should wait for game-by-game performance, thermals and noise results rather than relying on the CUDA core count.
Does unified memory make local AI better?
It gives the CPU and GPU access to a shared pool that can reach 128 GB in the highest configuration, which is useful for workloads that need more memory than a conventional laptop GPU offers. It does not, by itself, prove faster inference or better battery life. Memory capacity, bandwidth, software support and sustained power behavior all matter.
What about privacy and AI agents?
Local processing can reduce the need to send some workloads to remote services, but the exact behavior of Windows-integrated agents depends on the software implementation and permissions. The concept also raises reasonable questions about how much access an agent should have to files and applications. Those questions are part of the platform’s design challenge, not a performance specification.
Should you compare it directly with a Mac?
Compare the software you actually use. RTX Spark’s strongest documented advantages are CUDA and RTX support inside Windows, while Apple silicon represents a different hardware and software ecosystem. There is no measured comparison here that establishes one platform as universally faster for local AI or creative work.
Bottom line
The Lenovo Yoga Pro 9n and Yoga 9n 2-in-1 make RTX Spark more tangible than NVIDIA’s platform announcement alone. They show what the concept looks like in a real laptop: an Arm-based Windows machine with shared memory, serious CUDA ambitions, RTX graphics and premium convertible or clamshell designs.
The hardware story is compelling on paper. The buying story is not finished. Prices, exact Lenovo retail timing, software compatibility and independent AI, gaming and battery results are the details that will decide whether RTX Spark is a genuinely useful new laptop category or simply an impressive specification sheet.
