Nvidia lays out RTX Spark roadmap for laptops and desktop PCs at Computex 2026 — three generations outlined, Rubin with LPDDR6 memory, followed by Rosa Feynman

tom's Harware ^ | June 1, 2026 | Jeffrey Kampman

[fireman15]

Nvidia is fully committed to transforming Windows on Arm into an agentic AI platform.

Along with its first-generation RTX Spark platform for desktop and laptop PCs, Nvidia CEO Jensen Huang revealed the company's commitment to future generations of those platforms on its future roadmaps. The company is committed to producing at least two additional generations of Spark platforms for its partners.

Beyond the Grace Blackwell RTX Spark chips (the top-end RTX Spark Superchip and an as-yet-undetailed smaller chip), Huang promised that every future generation of the company's platforms will include a Spark chip.

That means there will be a Vera Rubin pair of Sparks powered by LPDDR6 memory, and a future Rosa Feynman Spark with a presumably even faster (but as-yet unannounced) memory generation. That multi-generational promise is an important point of trust in Nvidia's commitment to transforming Windows PCs for the agentic AI era.

Building a full product and partner ecosystem is a much larger challenge than simply building and shipping a chip. It's clear that Nvidia has a small army of OEM partners ready to take those chips to market and a deep partnership with Microsoft and ISVs to unlock the capabilities of its platforms for Windows and the applications that run on it.

---

TOPICS: Business/Economy; Computers/Internet; Hobbies
KEYWORDS: ai; linux; nvidia; windows

[fireman15]

Nvidia has been the largest recipient of funds from the huge AI Server Farm Build out. Now they are going after the entirely new level of AI PC hardware with Unified Memory that is coming down the pipe.

This type of computer using Nvidia's latest "RTX Spark" consumer laptop processor (the N1X), incorporating a Blackwell GPU, an ARM-based CPU, and 128 GB of unified memory will be capable of running 70B to 90B Models (e.g., Llama 3 70B, Llama 3.1 85B) with previously unmatchable performance and high precision for local hardware.

I had assumed that as this type of hardware became available that it would reduce the demand for the capabilities of AI server farms. And it likely will for many users, but it is interesting that Nvidia is not only embracing this type of hardware... they obviously have plans to dominate it. Although AMD and Apple currently seems to have a headstart with their Ryzen AI Max / Strix Halo chips and the Apple M5 Max.

Microsoft Windows Co-Pilot approved hardware does not hold a candle to these types of machines. People who are not enthusiasts will likely have a difficult time comprehending the major advancement that this architecture and hardware actually represent. For more than ten years we have seen nothing but incremental improvements in PC hardware. That is why so many of us able to continue using hardware designed for Windows 10 or even earlier operating systems with little difficulty.

[proxy_user]

Do consumers wish to buy these devices? That is the key question.

Remember GM and Ford going all-in on electric cars, only to discover the demand was not there.

[CatOwner]

Thanks to AI, the desktop PC I had spec’ed out back in September 2025 for around $1,400 would now cost around $2,600 for essentially the same components. Most of the cost increase is tied to RAM and solid state drive (SSD) costs. And guess what’s driving those costs through the stratosphere: AI data center.

[montag813]

This is a direct response to Apple completely selling out their entire MacMini line to OpenClaw users running local models.

[Lockbox]

Offload basic AI computing to the end user? Also creating a small localized networking AI system for the easy AI needs, thus offloading the larger AI data centers? And just to thrown in a conspiracy thought, your control over your computer is over. They will control you, your money and your life?

[Dr. Sivana]

I find the name confusing. We already have SPARC.

[ProtectOurFreedom]

“Do consumers wish to buy these devices? That is the key question.”

Are you kidding? My computer is over a year old and I’m on an iPhone 16 Pro Max! I’m ready to spend, spend, spend!! (/sarc in case you didn’t get it)

[Nervous Tick]

>> Do consumers wish to buy these devices? That is the key question.

Dunno. Good question though. Not me, I can tell you that much. I’m tired of computers and software. I’m going country. “Touching grass”, as the Zoomers say.

[Dr. Sivana]

Remember GM and Ford going all-in on electric cars, only to discover the demand was not there.

This is not completely analogous. It is true that some superior chips (e.g. Alpha) never found an audience. Apple has been able to force their M series on the client base, but Apple has changed Chip families multiple times with remarkably little interruption (680x0-->PowerPC--->x86--->ARM derivative).

If a steam deck power by the nVidia chips can play Steam games, run high-end workstation programs, or perform server activities better, there will be an audience because nVidia has the resources to convince buyers that support will be maintained. I am assuming that the nVidia product either performs faster, costs less to operate, or both.

[ProtectOurFreedom]

LOL...I thought the same thing, then I thought “I haven’t heard of SPARC in a long time.” Looked it up —> The Sun SPARCstation series is no longer in production. The last model was the SPARCstation 4, and the series was succeeded by the Sun Ultra series in 1995. After acquiring Sun, Oracle Corporation ended SPARC development in 2017.

Is it still a thing?

[fireman15]

Thanks to AI, the desktop PC I had spec’ed out back in September 2025 for around $1,400 would now cost around $2,600 for essentially the same components.

I would not be surprised if the increase was greater than $1,200. The problem for the entities that hoarded a lot of hardware which is largely responsible for this mess... is that the appropriate architecture for these facilities have already changed. A huge amount of hoarded hardware is going to be dumped for pennies on the dollar. Unfortunately, a lot of it is only going to be useful for specific uses.

[TexasGator]

“Thanks to AI, the desktop PC I had spec’ed out back in September 2025 for around $1,400 would now cost around $2,600 for essentially the same components. Most of the cost increase is tied to RAM and solid state drive (SSD) costs. And guess what’s driving those costs through the stratosphere: AI data center.”

PC prices have not almost doubled.

[fireman15]

I had an interesting conversation with Gemini 3.5 about a video that Nvidia released yesterday about this subject. Here is a link to that conversation that I found quite interesting.

https://gemini.google.com/share/2ced7468918e

[libh8er]

Amazing how far Nvidia has come. My first nvidia product was a GeGorce2 GTS I bought for playing Quake.

[bobcat62]

The Datacenter bubble will pop, since token-based billing isn’t going to work. Nobody wants to pay full-freight for AI.
Jensen needs to move product, so a return to the desktop with local models makes sense.

[napscoordinator]

I definitely bought their stock. I’m happy with it for right now. We’ll see how things go. of course, I’ve never sold the stock in my life from 18 to 57. I guess there’s always a first for everything.

[TexasGator]

“Thanks to AI, the desktop PC I had spec’ed out back in September 2025 for around $1,400 would now cost around $2,600 for essentially the same components. “

32Gb ram +200
4070 GPU +100
Sandisk 1TB +100

NET $500 increase

[fireman15]

PC prices have not almost doubled.

I purchased this 64GB DDR5 5600mhz kit for $210 in November, Now it is $860...

https://www.amazon.com/Tech-5600MHz-PC5-44800-Non-ECC-Unbuffered/dp/B0CQ3PG5D4?crid=3I5PRVOP8B7LK

The Samsung 2TB NVME that I purchased in September was $114 and now it is $409...
https://www.amazon.com/dp/B0DHLCRF91

So yes, component prices have more than doubled in many cases.

[TexasGator]

“So yes, component prices have more than doubled in many cases.”

My posted numbers include a more than doubling in RAM prices.

Please note that they are for what was a $1400 system last year.

I bought my system for $3300 at the low point before the recent price increases. Similar systems are now about $4000.

[fireman15]

32Gb ram +200
4070 GPU +100
Sandisk 1TB +100

Provide actual links please...

32GB of RAM used to be about $90, now it is $390 +300
https://www.amazon.com/Crucial-5600MHz-5200MHz-4800MHz-CT2K16G56C46S5/dp/B0BLTDRRLF

4070 GPU used to be about $250, now they are $850 +600
https://www.amazon.com/GIGABYTE-GeForce-WINDFORCE-Graphics-GV-N407SWF3OC-12GD/dp/B0CSJV61BN

Sandisk 1TB NVME used to be about $90, now they are $300 +210
https://www.amazon.com/SANDISK-1TB-Optimus-7100-NVMe/dp/B0GK1M1N4V?crid=1D5TEO778W96Q

NET $1200 increase.