Perplexity agent now runs privately on Windows PCs with NVIDIA RTX

Perplexity is expanding access to its private AI workflows with Portable Computer, now available on Windows PCs powered by NVIDIA GeForce RTX and NVIDIA RTX PRO Workstations. The local agent platform allows users to analyze data and automate tasks directly on their device, keeping sensitive information secure and reducing reliance on cloud computing, NVIDIA says. “As local models become more capable, AI agents can handle more work directly on a PC while keeping sensitive information on the device,” Perplexity notes, and the app simplifies setup with the Qwen 3.8 27B model pre-trained for use with Perplexity Computer and optimized for NVIDIA RTX GPUs.

Perplexity Portable Computer Launches on Windows with NVIDIA RTX NVIDIA GeForce RTX and RTX PRO GPUs now power Perplexity Portable Computer, extending the platform’s reach beyond Linux and NVIDIA DGX Spark systems to a broader base of Windows users. This expansion allows more individuals to utilize private AI workflows directly on their personal computers, bypassing the need for cloud-based processing for certain tasks, according to NVIDIA. The system requires 24GB or more of VRAM to operate, ensuring sufficient memory for complex local model processing. Support for NVIDIA DGX Station is anticipated in the near future, broadening hardware compatibility. Portable Computer integrates local and cloud AI capabilities within a single application, enabling users to manage sensitive files locally while still accessing advanced cloud-based reasoning when needed. Before transmitting any data to cloud models, the system requests explicit user permission, prioritizing data privacy and control. This approach allows for a tiered system where routine data analysis and task automation occur on-device, while more demanding computations can be offloaded with user consent. Perplexity cites a startup use case where the agent analyzes funnel exports locally to identify signup drop-off points, then shares insights directly to a Slack channel. The core of Portable Computer’s functionality relies on the Qwen 3 model. NVIDIA’s ongoing investment in GPU technology, including the recent extension of the CUDA toolkit to Windows on Arm systems and the opening of the CUDA-Q platform for fault-tolerant quantum applications, underpins this capability, the firm reports. In September 2025, NVIDIA launched the Advanced Quantum Computing research center in Boston, focusing on hybrid quantum-classical algorithms and drug discovery, demonstrating a commitment to advancing both classical and quantum computing solutions. The company stated, signaling confidence in the future of agentic AI. Connectors extending functionality to Microsoft Outlook, OneDrive, Word, Google Drive, Gmail, Slack, and GitHub further integrate the agent into existing user workflows, enhancing productivity across multiple applications.

Connected Apps Expand Perplexity’s Local AI Capabilities This move allows Perplexity to reach a significantly larger user base already invested in NVIDIA’s graphics processing capabilities, streamlining the deployment of local AI tools. The platform integrates seamlessly with existing Windows environments, eliminating the need for users to transition operating systems or rebuild existing workflows to benefit from private AI processing. “Users can put the agent to work without having to research models or configure the complex software stack typically required to run local AI,” according to the company. A key design element prioritizes user control over data privacy; Portable Computer requests permission before transmitting data to cloud models for more complex reasoning. NVIDIA’s NVQLink architecture, demonstrated with partners like Quantinuum and QuEra Computing, provides a low-latency pathway for integrating quantum processors with GPU supercomputers, hinting at future possibilities for hybrid quantum-classical workflows within the Perplexity ecosystem. Source: https://blogs.nvidia.com/blog/local-ai-perplexity-windows-pcs/ More like thisArtificial IntelligenceNVIDIA boosts MoE AI training 10.4x with JAX and Transformer EngineArtificial IntelligenceRutgers finds most Americans now use AI daily, but distrust its controlTechnologyGoogle Research Shares 5 Visuals of Male Fruit Fly’s Neural NetworkArtificial IntelligenceQuandela and NVIDIA link quantum processors to AI with NVQLinkStay currentSee today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals. Tags: Ivy Delaney Ivy Delaney has been working with neural networks and machine learning since the mid-nineties, back when a couple of hidden layers and a long afternoon of training counted as ambitious. She has watched the field go from academic curiosity to the thing quietly running underneath everything, and she brings that long view to quantum computing.

For Quantum Zeitgeist she covers the ground where the two fields meet. That means quantum machine learning and the variational algorithms it leans on, and it also means the less glamorous but more interesting story of classical machine learning already doing real work inside quantum machines, decoding error-correcting codes, calibrating noisy hardware and learning the error models that simulators depend on. She writes about the hardware those algorithms have to run on too, and about the post-quantum cryptography scramble that the same hardware has set off. Her stories typically start with the paper, whether that is peer-reviewed work, conference proceedings or an arXiv preprint, with the source linked so you can hold a claim up against the research it came from. She is unimpressed by benchmarks that will not say what they beat, and by demonstrations that only work in the press release. Latest Posts by Ivy Delaney: UChicago PME explains quantum to Chicago teens with ice cream & cell phones September 15, 2026 Aqarios boosts IBM Qiskit library with quantum constraint tool September 15, 2026 Bjarne Stroustrup, creator of the C++ programming language, has joined Susquehanna International Group (SIG) September 15, 2026