Nvidia has spent the last few years becoming the undisputed king of data center AI hardware. Now it wants your laptop, too.
The company unveiled the RTX Spark superchip on June 1 at its GTC keynote during Computex in Taipei. It combines up to 20 Arm CPU cores with Nvidia’s high-performance Blackwell GPU architecture, delivering roughly 1 petaflop of AI compute power in a form factor slim enough for consumer laptops and compact desktops.
The chip supports up to 128GB of unified memory, a spec that puts it in direct competition with Apple’s highest-end silicon offerings. Devices powered by the RTX Spark are expected to ship in fall 2026 from hardware partners: ASUS, Dell, HP, Lenovo, Microsoft Surface, MSI, Acer, and GIGABYTE.
From cloud to couch: why on-device AI matters
The partnership with Microsoft is central to this vision. The two companies are collaborating to build what they’re calling AI personal computers, machines designed from the ground up to run intelligent agents locally rather than relying on cloud infrastructure.
Nvidia CEO Jensen Huang called the RTX Spark “the new PC” during the announcement, emphasizing how AI agents could fundamentally replace traditional user interfaces.
Picking a fight with everyone
By building an Arm-based chip for consumer PCs, Nvidia is simultaneously challenging nearly every major player in the processor market. Intel and AMD have dominated the Windows PC chip business for decades. Apple has made its own silicon the centerpiece of its Mac strategy since 2020. Qualcomm has been pushing Arm-based chips into Windows laptops through its Snapdragon X series.
The 1 petaflop figure deserves context. A petaflop is one quadrillion floating-point operations per second. Just a decade ago, that kind of performance required room-sized supercomputers.
What this means for investors and the broader market
Wealth Club strategist Susannah Streeter characterized the launch as a bold attempt to reach consumers, while stressing that Nvidia’s core financial engine remains its data center business.