Tiiny AI Inc. has officially launched the world’s smallest personal AI supercomputer, the Tiiny AI Pocket Lab, in a groundbreaking event held in Hong Kong on December 10, 2025. This compact device has been verified by Guinness World Records as the “Smallest MiniPC (100B LLM Locally).” This innovative supercomputer is designed to run a full 120-billion-parameter large language model (LLM) entirely on-device, eliminating the need for cloud connectivity, servers, or high-end GPUs.
The Tiiny AI Pocket Lab represents a significant advancement in AI technology, focusing on energy efficiency and personal intelligence. It operates within a 65W power envelope, allowing for large-model performance at a fraction of the energy and carbon footprint associated with traditional GPU-based systems. As the AI industry faces challenges related to sustainability and privacy, Tiiny AI aims to provide a portable solution tailored to individual users.
According to Samar Bhoj, GTM Director of Tiiny AI, “Cloud AI has brought remarkable progress, but it also created dependency, vulnerability, and sustainability challenges.” The introduction of the Tiiny AI Pocket Lab is seen as a pivotal step towards making advanced AI accessible and private, shifting the power of intelligence from cloud data centers to personal devices.
This pocket-sized supercomputer is tailored for a wide range of applications, supporting developers, researchers, creators, professionals, and students. It offers capabilities such as multi-step reasoning, deep context understanding, and secure processing of sensitive information, all without requiring internet access. Notably, the device stores user data, preferences, and documents locally with bank-level encryption, ensuring a level of privacy that cloud systems cannot match.
The Tiiny AI Pocket Lab operates in the ‘golden zone’ of personal AI, catering to over 80% of real-world needs with its capacity to handle models from 10 billion to 120 billion parameters. This allows for performance levels comparable to the advanced GPT-4o, enabling PhD-level reasoning and deep contextual understanding while maintaining the security of on-device processing.
As the global market for LLMs is projected to grow significantly, reaching USD 35.4 billion by 2030 according to research from Grand View Research, Tiiny AI is well-positioned to capitalize on this trend. The rapid expansion is driven by advances in self-supervised learning and rising enterprise adoption across sectors such as retail, e-commerce, healthcare, and media.
The technology behind the Tiiny AI Pocket Lab includes two key innovations: TurboSparse and PowerInfer. TurboSparse is a neuron-level sparse activation technique that enhances inference efficiency while preserving full model intelligence. Meanwhile, PowerInfer, an open-source heterogeneous inference engine, dynamically distributes computational tasks across CPU and NPU, achieving server-grade performance with significantly lower power consumption.
Key specifications of the Tiiny AI Pocket Lab include:
– **Processor:** ARMv9.2 12-core CPU
– **AI Compute Power:** Custom heterogeneous module (SoC + dNPU), delivering approximately 190 TOPS
– **Memory & Storage:** 80GB LPDDR5X + 1TB SSD
– **Model Capacity:** Runs up to 120B-parameter LLMs fully on-device
– **Power Efficiency:** 30W TDP, 65W typical system power
– **Dimensions & Weight:** 14.2 × 8 × 2.53 cm, approximately 300g, pocket-sized
– **Ecosystem:** One-click deployment of numerous open-source LLMs and agent frameworks
– **Connectivity:** Functions fully offline, without internet or cloud requirements
The device is designed for ease of use, allowing for one-click installation of leading open-source models including OpenAI GPT-OSS, Llama, and Mistral, among others. Users will also receive continuous updates and official over-the-air hardware upgrades.
The Tiiny AI Pocket Lab is set to be showcased at CES in January 2026, marking a significant milestone for the startup, which was founded in 2024 by a team of engineers from MIT, Stanford, HKUST, SJTU, Intel, and Meta. Their expertise in systems engineering and AI inference optimization has led to advancements that enable large-parameter models to run on compact consumer devices for the first time.
With a multi-million dollar seed round secured from leading global investors in 2025, Tiiny AI is on a mission to redefine the future of personal AI by ensuring it is accessible, private, and tailored for individual use.
