Summary
- Lenovo's Yoga 9n two-in-one measures 0.69 inches thick and supports up to 64GB of memory, while the thinner 15-inch Yoga Pro 9n scales memory all the way to 128GB.
- Exact pricing remains mostly unannounced, but the AMD-based rival Lenovo ThinkCentre X Ultra starts at $3,699 for its base configuration, hinting at a pricey trend ahead.
Teased three months ago, now shown as real hardware
NVIDIA's RTX Spark platform made its live debut at IFA, arriving in laptops and mini PCs built to run AI agents — ones that don't just answer questions but search, generate files, and even handle bookings — entirely on-device, with no cloud connection required.
To unpack that a bit: RTX Spark is NVIDIA's chip design that fuses CPU, GPU, and memory onto a single piece of silicon. The architecture closely mirrors the approach Apple takes with its own chips in Macs, letting laptops run AI agents locally without ever reaching out to a cloud server.
Related article: NVIDIA introduces a guide for running local AI by linking multiple DGX Spark units
NVIDIA already showed how to network multiple desktop DGX Spark units together for local AI back in August. RTX Spark looks like the natural extension of that local-AI hardware strategy into the laptop category.
What's inside RTX Spark
RTX Spark is a system-on-chip that packs CPU, GPU, and memory into one unit. It pairs a Grace-architecture CPU with up to 20 cores alongside a Blackwell RTX GPU with up to 6,144 cores. That fused CPU-GPU design has more in common with Apple's own Mac silicon than with the traditional layout of Windows laptops. Based on core count alone, the chip's graphics performance is estimated to land somewhere between today's laptop RTX 5070 Ti and RTX 5080. NVIDIA has not yet said anything official about battery life.
Yoga 9n vs. Yoga Pro 9n: what's different
Lenovo's first RTX Spark laptop is the Yoga 9n, a 16-inch two-in-one. It features a 360-degree hinge, a 2880×1800, 120Hz OLED touchscreen, a six-speaker array, and a 9-megapixel webcam. At 0.69 inches thick, it's only 0.03 inches thicker than the 16-inch MacBook Pro — a slimness made possible by fusing CPU, GPU, and memory onto one chip instead of adding a discrete GPU.
Its 15-inch sibling, the Yoga Pro 9n, is even thinner at 0.66 inches and drops resolution to 2560×1600, but adds a new Force Pad touchpad that supports stylus writing and drawing. The memory story is where things get interesting: the 16-inch Yoga 9n tops out at 64GB, while the 15-inch Yoga Pro 9n scales up to 128GB — enough of a gap that the "Pro" branding could fairly be read as shorthand for memory headroom.
| Spec | Yoga 9n 2-in-1 (16-inch) | Yoga Pro 9n (15-inch) |
|---|---|---|
| Display | 2880×1800, 120Hz OLED touch | 2560×1600 OLED |
| Thickness | 0.69 in | 0.66 in |
| Max memory | 64GB | 128GB |
| Highlights | 360° hinge, 6 speakers, 9MP webcam | Stylus-ready Force Pad |
Dell, Asus, and Microsoft line up too
Lenovo isn't the only one building RTX Spark laptops. The Dell XPS 16, Asus ProArt P16, and Microsoft Surface Laptop Ultra are all set to launch this fall, and the only 14-inch-class model announced so far is the HP OmniBook X 14. These machines appear largely aimed at creators and heavy AI users.
The mini PC battle takes aim at the Mac mini
Laptops aren't the whole story. Acer unveiled the Acer SFF RTX Spark, a mini PC roughly the size of a Mac mini. Like the ProArt mini PC that Asus showed earlier this summer, it touts up to 128GB of memory and petaflop-class AI performance. The announcement lands exactly a week after Apple released the M6 Mac mini and M5 Ultra Mac Studio, setting up a direct clash between the two companies in the small-desktop local-AI market.
AMD joins the fray, and pricing is still murky
NVIDIA isn't the only player on this stage. Lenovo also announced an AMD-based machine, the ThinkCentre X Ultra, built on the Ryzen AI Max+ Pro 495 chip and likewise supporting up to 128GB of memory. Its starting price is reportedly $3,699 — and that's for the base configuration, not a fully loaded 128GB model. An AMD-based 128GB developer kit is already available for purchase.
RTX Spark laptop pricing hasn't been announced yet, but given how unusually expensive memory has become lately, don't expect these machines to come cheap.
Editor's View
What stands out from this IFA reveal is that NVIDIA didn't stop at the desktop DGX Spark — it pulled in laptop makers all at once. Having Lenovo, Dell, Asus, Microsoft, and HP simultaneously roll out RTX Spark laptops suggests NVIDIA isn't just pushing its own branded hardware this time; it's trying to unify the entire Windows ecosystem around a local-AI standard. This marks the first time a company other than Apple has pushed a fused CPU-GPU-memory design at this scale, effectively kicking off an Apple Silicon vs. NVIDIA SoC rivalry in the laptop market too.
Working with laptops built on unified-chip designs tends to leave a consistent impression: they get noticeably thinner and run cooler, but that comes with the same tradeoff every time — memory can't be upgraded later. That's likely why the Yoga 9n caps out at 64GB while only the Yoga Pro 9n opens up to 128GB. Anyone planning to run several large language models locally at once would be wise to max out memory at purchase time to avoid regret down the road.
For now, there's no reason to rush out and buy into this trend. Most prices haven't been disclosed, and practical details like battery life — which matter a lot for everyday use — still haven't surfaced. Still, where RTX Spark's actual price tags land between the $3,699 AMD-based mini PC and a 128GB MacBook Pro north of $6,000 will likely be the first real benchmark for deciding whether local-AI hardware is worth buying. As more of these launches roll out this fall and real sale prices and battery numbers start coming in, it should become a lot clearer who's actually winning this race.
Comments