Tiiny AI Pocket Lab brings GPT-scale intelligence fully offline, letting developers and creators run and customise open-source models on a secure, pocket-sized device.
Tiiny AI Inc., a US deep-tech startup, today unveiled the world’s smallest personal AI supercomputer, the Tiiny AI Pocket Lab, officially verified by Guinness World Records under “The Smallest MiniPC (100B LLM Locally).” This marks the first global public unveiling of a pocket-sized device capable of running up to 120-billion-parameter large language models entirely offline, without cloud connectivity, servers, or high-end GPUs.
Designed for energy-efficient personal intelligence, the Tiiny AI Pocket Lab operates within a 65W power envelope, delivering large-model performance with minimal energy and carbon footprint. It supports multi-step reasoning, deep context understanding, agent workflows, content generation, and provides bank-level encrypted personal memory, ensuring data privacy without internet access.
“Cloud AI has brought remarkable progress, but it also created dependency, vulnerability, and sustainability challenges,” said Samar Bhoj, GTM Director of Tiiny AI.
“With Tiiny AI Pocket Lab, we believe intelligence shouldn’t belong to data centers, but to people. This is the first step toward making advanced AI truly accessible, private, and personal, by bringing the power of large models from the cloud to every individual device.”
The device targets developers, researchers, creators, professionals, and students, covering the 10B–100B parameter ‘golden zone’, sufficient for over 80% of real-world AI applications. Tiiny AI integrates open-source software and hardware frameworks, supporting one-click deployment of leading models such as OpenAI GPT-OSS, Llama, Qwen, DeepSeek, Mistral, Phi, and AI agents like OpenManus, ComfyUI, Flowise, Presenton, Libra, Bella, SillyTavern.
Powered by TurboSparse and PowerInfer, the device achieves server-grade performance at low power. This open-source approach democratizes access to high-performance AI, eliminating cloud dependence while enabling secure, offline, and fully customisable workflows. CES 2026 will showcase the full feature set.
