Nvidia has clearly shown at Computex 2026 that its ambitions are no longer limited to the market of graphics cards and AI data centers. The company presented a long-term strategy for the development of the DGX Spark platform for personal computers, revealing a plan that spans the next few years and includes as many as three new generations of processors based on the Rubin, Rosa and Feynman architectures.
The announcement comes just days after Microsoft and Nvidia unveiled the Surface Laptop Ultra, the first major commercial device based on the RTX Spark platform. However, the latest development map shows that RTX Spark is just the beginning of a much more ambitious plan.
DGX Spark represents Nvidia’s vision of computers optimized for local artificial intelligence. Unlike traditional PC systems, where the processor, graphics chip and AI accelerators are separate components, Nvidia builds unique platforms that unite all key elements in one system. The goal is to enable the launch of increasingly large AI models directly on user devices, without relying on cloud infrastructure.
According to the presented plan, the next big step will be the transition to the Rubin architecture, which will inherit the current Blackwell generation. This will be followed by Rosa, while the most advanced platform will be named Feynman, after the famous American physicist Richard Feynman.
NVIDIAThe choice of name itself is not random. Nvidia has been naming its most important architectures after famous scientists and researchers for years. Thus, previous generations bore the names Hopper, Lovelace, Blackwell and Rubin, paying tribute to personalities who significantly contributed to the development of science and technology.
What makes this announcement particularly significant is the fact that Nvidia is speaking publicly for the first time about a multi-year strategy for the PC segment in a way that resembles its data center plans. Until now, long-term development roadmaps have mostly been reserved for server-side AI accelerators, while consumer products have evolved through shorter cycles.
It shows how serious the company is about making AI computers the next big category of devices.
According to data presented by Nvidia, future generations of DGX Spark systems will be focused on exponential growth of AI performance. As models become larger and more demanding, the company believes that users will need significantly more powerful local systems that can process hundreds of billions of parameters without sending data to the cloud.
It represents a significant shift in the way the industry views the future of personal computing.
During the previous decade, the prevailing idea was that most advanced AI functions would be executed in large cloud data centers. However, growing concerns about privacy, the cost of cloud services and the need for lower latency have led to a new trend known as local or on-device AI processing.
Microsoft, Apple, Qualcomm and Google are already developing their own strategies in this direction, but Nvidia is trying to take a unique position by offering hardware that can simultaneously serve for professional AI tasks, creative applications and gaming.
NVIDIAIt is particularly interesting that Nvidia no longer hides that it wants a much bigger role in the processor market. Traditionally known as a graphics card company, Nvidia is now developing complete computing platforms based on the Arm architecture, directly entering a space dominated by Intel and AMD for decades.
It’s a strategy reminiscent of the success of the Apple Silicon platform. Apple has shown that an integrated approach to processor, graphics and memory can bring significant benefits in performance and energy efficiency. Nvidia is now trying to implement a similar model in the Windows ecosystem, but with a much greater focus on artificial intelligence.
However, the path to mass success will not be easy.
Application compatibility, optimization of operating systems and willingness of developers to adapt software to new architectures remain key challenges. The history of the computer industry is full of examples of technologically impressive platforms that failed to gain the support of the wider software ecosystem.
This is precisely why cooperation with Microsoft is important, which is simultaneously adapting Windows 11 for the new generation of AI computers and trying to create a development ecosystem that will take advantage of the possibilities of local artificial intelligence.
By publishing a multi-year development map, Nvidia sends a clear message to the market: AI computers are neither a passing trend nor a short-term experiment. The company believes they will become a central category of personal computing over the next decade.
Whether that vision becomes a reality will depend on many factors, including the price of the device, the quality of software support, and the actual needs of users for local AI models, concludes Tom’s Hardware.