NVIDIA and Microsoft have jointly unveiled a new PC chip called RTX Spark that brings artificial intelligence processing directly to consumer laptops and desktops, marking what executives call the first comprehensive redesign of personal computers in four decades. The chip, announced at the GTC Taipei conference on June 1st, combines a graphics processor, central processor, and 128 gigabytes of unified memory to run AI agents locally without relying on cloud services.

What Makes the RTX Spark Different From Today's PCs?

The RTX Spark represents a fundamental shift in how computers process artificial intelligence. Built on TSMC's 3-nanometer manufacturing process, the chip contains 70 billion transistors and delivers 1 petaflop of AI computing power, meaning it can perform one quadrillion floating-point calculations per second. For context, that's enough raw processing capacity to run large language models with 120 billion parameters and process documents containing up to 1 million tokens simultaneously.

The chip integrates an NVIDIA Blackwell graphics processor with 6,144 CUDA cores, which are specialized processing units designed for parallel computing tasks, alongside a custom 20-core central processor developed in collaboration with MediaTek. This combination allows the RTX Spark to handle demanding workloads that previously required either expensive desktop workstations or cloud-based services.

Jensen Huang, NVIDIA's founder and chief executive, stated that the RTX Spark is "the most amazing chip the world has ever built" and emphasized that "100% of NVIDIA's software stack runs on it". The chip will power Windows PCs from manufacturers including ASUS, Dell, HP, Lenovo, Microsoft Surface, and MSI, with laptops and compact desktops arriving starting in September 2026.

Jensen Huang, NVIDIA's founder and chief executive

How Will AI Agents Change What Your Computer Does?

Today, using a computer typically requires deliberate actions: opening applications, clicking with a mouse, and typing commands. With RTX Spark and Windows, the interaction model shifts fundamentally. Instead of manually launching tools, users can make conversational requests, and the computer completes tasks autonomously through AI agents.

Huang described this transformation as comparable to the shift from traditional phones to smartphones. "There is no question this reinvention of the computer is as big of a deal as the reinvention of the phone into what we now know as the smartphone," he said. The vision extends beyond simple task completion; Huang suggested that over time, these systems could feel less like traditional computers and more like intelligent assistants, similar to characters from science fiction.

Huang

The practical capabilities are substantial. A PC equipped with RTX Spark can render 90-gigabyte ultra-large 3D scenes using NVIDIA's OptiX and DLSS technologies, edit 12K video at 4:2:2 color sampling, run large language models with 120 billion parameters, and play demanding video games at 1440p resolution with over 100 frames per second using ray tracing. Adobe has redesigned Photoshop and Premiere specifically for the RTX Spark, with the new versions offering up to 2 times faster processing for AI, editing, color correction, and special effects workflows.

What Are the Key Technical Specifications and Capabilities?

• Processing Power: The RTX Spark delivers 1 petaflop of AI computing performance, enabling it to run large language models with 120 billion parameters and process documents with up to 1 million tokens in context.
• Memory Configuration: The chip includes 128 gigabytes of unified memory, which allows the processor and graphics processor to access the same data without copying it between separate memory pools, improving efficiency.
• Manufacturing Process: Built on TSMC's 3-nanometer process with 70 billion transistors, making it one of the most advanced consumer chips available.
• Graphics Capability: Features an NVIDIA Blackwell graphics processor with 6,144 CUDA cores and fifth-generation Tensor Cores for accelerated AI and graphics workloads.
• Physical Design: Laptops powered by RTX Spark will be only 14 millimeters thick and weigh approximately 3 pounds, available in 14-inch and 16-inch screen sizes.

What Desktop Option Is NVIDIA Offering for Professionals?

Beyond consumer laptops, NVIDIA introduced the DGX Station for Windows, a desktop-level AI supercomputer designed for professional and enterprise use. This system uses the GB300 Grace Blackwell Ultra chip and provides 748 gigabytes of unified memory, 20 petaflops of AI computing power, and 800 gigabits-per-second network connectivity through a ConnectX-8 SuperNIC.

The DGX Station can run trillion-parameter AI models and execute hundreds of AI agents simultaneously, making it suitable for organizations deploying large-scale intelligent agent systems. It supports optional configuration with an RTX Pro 6000 workstation-level graphics processor for applications combining AI computing with ray-tracing visualization for creative design and engineering work.

How Does This Challenge Intel and AMD?

NVIDIA's move directly threatens the traditional processor market dominated by Intel and AMD. The company also announced the Vera CPU, a new processor architecture designed specifically for AI agent deployment, which NVIDIA expects to generate $20 billion in revenue by the end of the fiscal year. Michael Larabel, author of the Phoronix CPU Benchmark Suite, stated that "NVIDIA Vera is the most powerful competitor to Intel and AMD x86_64 processors in history".

The Vera CPU has already attracted major customers. Three AI giants, Anthropic, OpenAI, and SpaceX, have all been among the first to deploy the Vera CPU, according to NVIDIA's announcement. Huang noted that current order volumes suggest the Vera CPU will achieve "the fastest and most successful product launch in our company's history".

Huang

What Do Australians Think About AI Agents on Personal Devices?

Consumer adoption of AI agents on personal devices remains limited in some markets. According to a survey of more than 2,000 Australians released on June 2nd, only about one in ten Australians are currently running AI agents on a dedicated device. Among those who chose a computer for this purpose, more than 40 percent opted for Apple silicon, a share considerably above Apple's overall market share in Australian households.

Trust and security concerns represent significant barriers to broader adoption. Less than one-third of Australians aged over 16 would be comfortable with AI managing more of their everyday lives, with most respondents expressing concerns about privacy and security implications of giving AI greater control and access to personal data. Alvin Lee, a senior analyst at Telsyte, noted that "more powerful on-device processing could give consumers the control they need" to address these concerns.

Fundamental features such as battery life and performance remain the dominant purchasing considerations for Australians when selecting computers. Only one in three Australians say their next smartphone or computer must include advanced AI features, as the market is currently dominated by cloud-based AI services. However, around 40 percent of respondents expressed interest in hypothetical devices from major AI companies such as OpenAI, Anthropic, and xAI, which have not yet released consumer hardware products.

What AI Models Will Run on RTX Spark?

NVIDIA announced that its Nemotron 3 Ultra model, described as the most powerful model to date for intelligent agent deployment, will run on RTX Spark systems. The model's intelligent agent productivity test score exceeds or ties with three leading AI models: Zhipu GLM 5.1, Dark Side Kimi K2.6, and Alibaba Qwen 3.5.

NVIDIA also open-sourced the Cosmos 3 foundation model for physical AI and the Alpamayo 2 Super autonomous driving inference model, while announcing the establishment of the Cosmos Alliance to support development of AI systems for robotics and autonomous systems.

The company emphasized that NVIDIA's CUDA software framework, which has become the industry standard for AI development, runs natively on the RTX Spark. This compatibility means developers can deploy existing AI applications without significant modifications.