Loading story…
Perplexity has released Portable Computer, a Windows version of its Computer agent that runs locally on compatible Nvidia GeForce RTX and RTX PRO GPUs with at least 24 GB of VRAM. Available to Perplexity Pro and Max subscribers, the agent supports multistep task planning, browser use, tool calling, connectors, scheduled recurring tasks, local MCP servers for desktop applications, and one-click downloads of local models including PPLX 27B and Qwen 3.8 27B.
Although inference can run on-device and does not consume Computer credits, the deployment is not fully air-gapped: Portable Computer can connect to local and external resources such as Microsoft and Google productivity services, Slack, and GitHub. It may also ask users to authorize sending work to Perplexity cloud models when additional reasoning capacity is required, creating a hybrid local/cloud data-processing path that organizations should assess before connecting enterprise resources.
Track how attackers are adapting to this technology.
3 events from the most recent confirmed update back to the earliest known activity.
Perplexity initially released its local Portable Computer agent for Linux-based systems and Nvidia DGX Spark.
Perplexity released its local Computer agent in the Perplexity Windows app in partnership with Nvidia. The release requires a Pro or Max subscription and a compatible GeForce RTX or RTX PRO GPU with at least 24 GB of VRAM.
Perplexity subsequently introduced hybrid local/cloud compute support for Portable Computer on Apple silicon Macs.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
Follow how adversaries are adapting to this technology, and where it touches your stack today.
2 references tracked. Mallory keeps watching after this page renders.
tomshardware.com
Open sourcethenewstack.io
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
Microsoft unveiled **Project Zenith**, a developer-focused Windows 11 configuration for high-memory PCs that can run AI models exceeding 30 billion parameters locally. The initiative targets systems with at least 64 GB of unified memory and more than 250 GB/s memory bandwidth; initial devices will use AMD Ryzen AI Halo, with additional OEM and silicon partners expected. Zenith combines Windows performance tuning and developer-oriented defaults with integrated WSL containers and Microsoft Execution Containers. Microsoft said the configuration provides containment, identity controls, and enterprise management for agentic-development workloads, enabling routine AI-assisted development to move from metered cloud inference to on-device models while retaining cloud services for more demanding tasks.
Sep 8, 2026
Nvidia is reportedly preparing an **open-source AI agent platform**—described as a direct competitor to **OpenClaw**—and has been pitching the project (reportedly called *NemoClaw*) to potential corporate partners ahead of its upcoming developer conference. Reported companies in discussions include **Salesforce, Cisco, Google, Adobe, and CrowdStrike**, though the nature of any partnerships and associated benefits remains unclear. Separately, Minisforum announced an upcoming NAS positioned to run local LLM workloads and ship with **OpenClaw pre-installed**, highlighting growing interest in “always-on” personal/edge agent frameworks and local AI execution on consumer hardware. An unrelated Intel announcement covered **Bartlett Lake** P-core-only CPUs for embedded/edge use on **LGA 1700**, but did not address OpenClaw, Nvidia’s reported platform, or AI agent ecosystems.
Jun 29, 2026
OpenAI announced **GPT-5.3-Codex**, positioning it as more than a code-generation model and expanding its availability across a new macOS desktop app plus CLI, IDE extension, and web interface, with API access described as forthcoming. OpenAI and third-party coverage said the model is intended to support broader software-lifecycle work (e.g., debugging, deployments, monitoring, tests, and documentation) and highlighted benchmark gains over prior Codex versions; reporting also pushed back on claims that “Codex built itself,” characterizing OpenAI’s statement as the model being *instrumental* in its own development rather than fully autonomous. In parallel, OpenAI unveiled **Frontier**, an enterprise framework aimed at building, deploying, and managing AI agents, including promised agent security features and an approach modeled on Palantir-style **forward-deployed engineers** working alongside customer teams. Separately, *MintMCP* launched an enterprise governance platform focused on deploying, monitoring, and securing AI agents and MCP servers, emphasizing audit trails, policy enforcement, observability/guardrails, and centralized access controls to reduce risks from privileged agents (e.g., credential exposure and data exfiltration). Other items in the set were not tied to these launches, including a general CIO column on SMB IT “quick fixes” creating long-term risk and a corporate internal AI learning program announcement from Hancom.
Jun 29, 2026