AI News

After Jensen Huang Resigns as CEO, How Far Can Nvidia's RTX Spark PC Road Go?

LeitechUpdated Nov 15 22:05

AI summary

Jensen Huang resigned as Nvidia CEO on November 12, with CFO Colette Kress temporarily taking over, and the article thereby discusses whether RTX Spark—released in June this year and with its first laptops going on sale on October 16—can continue to advance. RTX Spark is an Arm-architecture PC SoC developed by Nvidia in cooperation with MediaTek, integrating a Grace CPU, Blackwell GPU, RTX graphics technology, and CUDA, and supporting up to 128GB of unified memory. The first batch of reviews shows that it has advantages in local large-model inference and large-capacity unified memory, but Windows on Arm compatibility, battery life, and heat still have shortcomings, and the local Agent experience also has barriers.

Why read this

AI score 58

The article places the shortcomings in the first batch of RTX Spark reviews alongside the share of the data center business, presenting the resource trade-offs Nvidia's consumer PC road faces after the management change.

Topics

Tags

Body · AI translation

NVIDIA Shakeup! In the Post-Jensen Huang Era, Can RTX Spark Still "Rebuild the PC"? 37/21 10 hours ago left behind a project to "rebuild the PC."

Barring the unexpected, at 10 a.m. Eastern Time on November 18, Voyager 1, which has been flying for nearly 50 years, will reach a point one light-day from Earth.

If Voyager 1 embodies humanity's ambition to explore the unknown, then Jensen Huang, through one bold bet after another on future technology, also led NVIDIA to places that were initially hard to imagine.

But on November 12, Jensen Huang, who had steered NVIDIA for 33 years, formally resigned as CEO, and long-serving CFO Colette Kress took over on an interim basis. For this chip giant that was shaped almost single-handedly by Jensen Huang, an unprecedented change has already begun.

Among the most noteworthy developments, perhaps, is RTX Spark.

Image source: NVIDIA

In June this year, Jensen Huang personally unveiled RTX Spark, announcing partnerships with Microsoft and MediaTek in an attempt to redefine the personal computer in the AI era. On October 16, the first laptops equipped with RTX Spark officially went on sale, less than a month before Jensen Huang's departure.

Unlike the GeForce gaming graphics card business, this is an important new offensive by NVIDIA into the PC chip market after many years. Now, RTX Spark, which has just gotten started, has to face a new question:

Without Jensen Huang, will NVIDIA still stick to this path? At the turning point for personal computers, making RTX Spark

RTX Spark is the most representative example this year.

It is an Arm-architecture PC SoC built by NVIDIA in partnership with MediaTek, integrating the Grace CPU, Blackwell GPU, RTX graphics technology and the CUDA computing platform, supporting up to 128GB of unified memory, and capable of running larger-scale AI models on a personal computer.

Spark means "spark." Compared with traditional GeForce graphics cards, RTX Spark's ambitions may be greater.

NVIDIA previously participated in the PC market mainly through discrete GPUs; now it hopes to enter directly into the PC's core computing platform and compete with Intel, AMD, Qualcomm and Apple.

At the launch event in June this year, Jensen Huang proposed that personal computers are undergoing a fundamental change. Over the past 40 years, people have used computers by opening applications, clicking a mouse and typing text, but the realization and iteration of Agents means users can directly make requests to the computer and let AI complete the work.

From this perspective, Jensen Huang hopes that through RTX Spark, NVIDIA can bring the technology it has accumulated in GPUs, CUDA and AI into personal computers, make local AI a new capability of the PC, and even thereby change the way people interact with computers.

Of course, unlike the "lavishness" of data centers, PC users are very sensitive to price, battery life, compatibility and everyday performance. Even with powerful GPUs and AI computing power, that does not mean one can make a personal computer with an excellent experience. Moreover, there is still no clear answer as to how much local Agents can actually change the PC.

The risks here are not small, but it is actually not hard to understand why Jensen Huang is willing to bet on this.

In NVIDIA's development over the past thirty-plus years, personal computers have always been an important starting point for its technological innovation and ecosystem expansion.

In an interview with Lex Fridman this year, Jensen Huang said that GeForce remains NVIDIA's most important marketing strategy to this day. Many people first encountered NVIDIA as teenagers through gaming, and later entered university and became developers, only then gradually coming into contact with tools such as CUDA and Blender.

Image source: Lex Fridman

A glimpse of this can also be seen in the birth of CUDA.

Jensen Huang revealed that when the decision was made to bring CUDA to GeForce gaming graphics cards, the related investment increased chip costs by about 50%, putting enormous pressure on NVIDIA, whose gross margin at the time was only around 35%.

But Jensen Huang still persisted. As GeForce gaming graphics cards entered a large number of personal computers, more and more researchers and developers began to come into contact with CUDA, ultimately driving the development of the GPU general-purpose computing ecosystem. And this investment took about ten years before it truly began to pay off.

RTX Spark naturally cannot be simply equated with CUDA back then, but there is indeed a similar line of thinking behind both: promoting new computing capabilities through personal computers, cultivating developers and the software ecosystem, and thereby striving for the opportunity to define the next-generation computing platform.

For Jensen Huang, the value of the personal computing market clearly cannot be measured only by immediate revenue.

Except this time, he can no longer personally continue to drive RTX Spark forward. NVIDIA Without Its Old Captain

Objectively speaking, RTX Spark is indeed hard to count as a successful consumer product at present.

Judging from the first batch of reviews, it has shown advantages in GPU performance, local large-model inference and large-capacity unified memory, and is especially suitable for developers and creators who need to run large models locally.

But there are also many problems. The Verge's experience with several RTX Spark laptops found that Windows on Arm still has shortcomings in software compatibility, with some games and professional applications performing poorly. In addition, battery life, heat and fan noise under high load are also less than satisfactory.

What is even more awkward is that the local Agent experience that Nvidia is heavily promoting still has barriers to entry. Generating a news briefing with Hermes Agent and a local Qwen model took about 10 minutes, consumed 10% of the battery, and the final result was still not as good as cloud-based Claude.

Even for the few geeks out there, spending thousands of dollars on an RTX Spark laptop clearly still lacks a sufficiently compelling reason.

Of course, as a first-generation product, these problems are not surprising. From chip performance and power optimization to Windows software adaptation and the local Agent ecosystem, RTX Spark all needs more time to be polished. The real question is:

How much patience does Nvidia have in the post-Jensen Huang era?

Image source: Nvidia

Nvidia's revenue in the second quarter of fiscal 2027 reached $96.2 billion, of which the data center business contributed $89 billion, accounting for about 92.5%. By comparison, revenue from the entire edge computing business was only $7.2 billion, which also includes products such as workstations.

In terms of commercial returns, the status of the two is already vastly different.

For Nvidia, continuing to invest in a new PC SoC product line requires bearing the costs of long-term R&D, ecosystem building and competing with established rivals for the market. Concentrating resources on AI data centers, by contrast, can more directly serve its current main source of revenue.

Just after Jensen Huang resigned, some people have already expressed concerns about future RTX graphics cards and Nvidia's custom chips for the next-generation Switch.

This is also why the outside world has begun to worry that after Jensen Huang leaves, Nvidia will further lower the priority of its consumer business.

In addition, Cress, who has temporarily taken over as CEO, has long been responsible for the company's financial management. To the outside world, she is clearly more pragmatic and places greater emphasis on commercial returns and financial risk.

By comparison, Jensen Huang is more enthusiastic about technological innovation and long-term strategic investment, and also has more personal feelings and judgment regarding the personal computer and gaming markets; otherwise, he would not have publicly said when facing the question of "abandoning consumer GPUs":

"RTX technology was invented for gamers... you deeply hurt my feelings."

The WSJ report also mentioned that Cress and her finance team had previously raised objections multiple times to Jensen Huang's aggressive AI investment plans, and were even criticized by Jensen Huang for only looking at financial reports, like "a deer caught in the headlights."

These disagreements do not involve RTX Spark, but they are enough to show the difference in business judgment between the two. Of course, Cress is only "interim CEO."

But what we can be sure of is that there may not be many people on this planet who can truly stand in the position of Nvidia CEO and, like Jensen Huang, be willing to take greater risks and push forward the future they believe in.

And when this question is applied to RTX Spark, it becomes even bigger. After Jensen Huang leaves, how far can RTX Spark go?

As a PC product line that requires years of investment and whose market prospects have not been fully verified, while the AI data center business is still growing rapidly, how attractive is continuing to compete in the PC market to the current CEO, who places more emphasis on financial returns, and to future Nvidia CEOs?

In fact, everyone already has quite a few concerns.

Kieren Jessop, a PC analyst at Omdia, pointed out in an interview with Tom's Guide that consumers will not feel much change in the short term, but the next two to three years are worth watching. He believes that Jensen Huang has always regarded the connection between personal computers and the developer ecosystem as a strategic asset, while an interim CEO from the finance department may focus more on data center customers, leading to a reduction in new consumer projects.

Image source: Nvidia

GeForce already has mature products, users and channels, and even if its strategic position declines, it is unlikely that Nvidia will easily give it up. By comparison, RTX Spark has only recently come to market and faces great uncertainty, making it precisely one of the businesses most vulnerable to such changes.

Although MediaTek confirmed at the end of October that the two sides' second-generation SoC is under development, this only shows that the existing cooperation is still moving forward and cannot guarantee how many resources Nvidia will invest in this route in the future.

Of course, RTX Spark also has sufficient reasons to continue to exist.

If Nvidia hopes that CUDA and RTX will continue to be an important technological foundation for the next-generation personal computing platform, then RTX Spark has its long-term value.

But this market needs time to cultivate, and one or two generations of products are clearly far from enough.

And Jensen Huang's departure has also made more people realize that some long-term investments may be inseparable from the founder's own judgment and persistence.

Back then, Nvidia was willing to spend ten years taking CUDA from GeForce to the broader computing market. Today, facing a new PC market with an unknown future, is Nvidia still willing to give RTX Spark the same patience?

Image in this article from: 123RF licensed image library. Infringement report. Return to share. Related articles. All comments.

Source

Leitech