For most of its history Nvidia has sold the graphics card that sits next to someone else’s processor. At Computex 2026 in Taipei it announced that it wants to be the processor too.

RTX Spark

The headline announcement was RTX Spark, a chip for Windows PCs developed with MediaTek. Nvidia rated it at 1 petaflop of FP4 AI compute, a measure aimed squarely at running AI models on the device rather than in the cloud. According to Nvidia and StorageReview, more than 30 laptop designs and more than 10 desktops using RTX Spark are due this autumn.

That puts Nvidia in direct competition with Intel, AMD and Qualcomm for the heart of the Windows PC.

The rest of the keynote

  • DGX Station for Windows: a desktop AI workstation, aimed at developers who want to run and fine-tune large models locally.
  • Vera CPU: Nvidia’s data-centre processor, part of the Vera Rubin platform, which Nvidia said was ramping into full production for shipments starting this autumn.
  • Physical AI: Cosmos 3 world models for simulation, Alpamayo 2 Super for autonomous driving, and a GR00T reference design for humanoid robots.

Why it matters for developers

  • Local AI gets a bigger audience. If dozens of mainstream laptops ship with a petaflop-class AI engine, running capable open-weight models on a laptop stops being a niche for people with gaming GPUs. That changes what “on-device” can mean for apps.
  • More chip variety on Windows. A new processor family means another target to test against, so check that your tools, runtimes and native dependencies support it when the machines ship.
  • CUDA everywhere. An Nvidia chip in mainstream PCs extends the reach of Nvidia’s software stack from the data centre to the desk, which strengthens the case for building on it.

The proof will come in the autumn: pricing, battery life and how well existing Windows software runs on the new chips.

Sources

  1. Nvidia blog: GTC Taipei and Computex 2026 news (Jun 1, 2026)
  2. StorageReview: Nvidia Computex 2026 keynote, the RTX Spark PC family, DGX Station and physical AI (Jun 1, 2026)
  3. Nvidia: Vera Rubin in full production (May 31, 2026)

We link every claim to the reporting or official statement it comes from. Where a detail rests on one outlet’s reporting, we say so. Spotted an error? Reply to any newsletter and we’ll correct it.