Global / AI Agents
Perplexity launches Windows version of Portable Computer agent powered by NVIDIA
Local AI agents running on consumer GPUs reduce latency and keep sensitive data off cloud servers.
Perplexity has released a Windows version of its Portable Computer agent, a local execution environment accelerated by NVIDIA RTX GPUs. The agent handles multistep tasks directly on users' devices, keeping sensitive information private.
Published · significance 62 of 100 (medium) · 1 source
What happened
Perplexity has made its Portable Computer—a local agent that plans and executes multistep tasks—available on Windows, powered by NVIDIA RTX acceleration. The agent uses local models to analyse data and aggregate information, keeping user data on the device rather than in cloud infrastructure.
Why it matters
This signals growing momentum toward on-device AI agents as models become more capable. By combining consumer-grade GPUs with local execution, the approach addresses both performance and privacy concerns, creating a competitive alternative to cloud-dependent agents. For NVIDIA, it extends RTX positioning beyond gaming and creative work into AI agent infrastructure.
What changes
Developers and users can now deploy multistep AI workflows locally on Windows PCs without cloud dependency or data exfiltration risk. This expands the practical scope of consumer-grade GPU acceleration for agentic work.
Involved
Sources
Written by AI from the reports above; scored by a published formula. How we work. Found a mistake? Email lockedinshreyash@gmail.com. Up To Date summarises and links to original reporting; it never reproduces articles.
Related coverage
- Jensen Huang takes Trump call onstage, dismisses robot takeover concerns — A symbolic moment: the president and Nvidia's CEO jointly addressed anxieties about automation, signalling political alignment on AI progress. (2026-09-14)
- NVIDIA showcases real-time AI applications for broadcast, sports and streaming at IBC — NVIDIA is demonstrating AI capabilities tailored for media production and live content delivery at the industry's largest annual gathering. (2026-09-09)
- NVIDIA Vera Rubin NVL72 delivers leading performance in MLPerf Inference v6.1 — System performance gains translate directly to inference economics: more tokens per dollar, efficient scaling, continuous software optimisation. (2026-09-16)
- ASML's next-generation High NA chipmaking tools sold out through 2027 — Semiconductor supply for AI chips tightens as manufacturers commit to pricier equipment. (2026-09-14)
- Skild AI launches S1 robot foundation model that learns tasks from single video — Robots that adapt to changing factory floors without reprogramming could reshape how manufacturers handle task variability. (2026-09-10)