Tiiny AI Pocket Lab: 120 Billion-Parameter Pocket Computer
Cloud-Independent Local AI Solution
US-based startup Tiiny AI introduced Pocket Lab, its pocket-sized computer that can run large language models (LLMs) without needing cloud servers. The device was certified by Guinness World Records as the “smallest Mini PC capable of running a 100 billion-parameter LLM.” Pocket Lab offers the ability to perform local processing while preserving data privacy without requiring an internet connection.
Technical Specifications and Performance
Tiiny AI Pocket Lab can process popular models such as Llama, Mistral, and DeepSeek on its own hardware. The device has a powerful infrastructure to run models with up to 120 billion parameters locally. At the core of the system are a 12-core ARMv9.2 processor and a dedicated neural processing unit (NPU).
On the hardware side, the most notable detail is memory capacity. The device comes with 80 GB LPDDR5X RAM and 1 TB SSD storage to ensure large models can be loaded into memory and run smoothly. Total AI processing power is around 190 TOPS. While power consumption has a 30W TDP rating, typical system power is around 65W.
The device’s key technical specifications are as follows:
- Processor: 12-Core ARMv9.2
- Memory: 80 GB LPDDR5X
- Storage: 1 TB SSD
- AI Power: ~190 TOPS (NPU + SoC)
- Dimensions: 14.2 x 8 x 2.53 cm
- Weight: 300 grams
Use Cases and Expected Release Date
This hardware ensures that data remains in the local environment, especially for developers and researchers who prioritize data privacy. Pricing for the Tiiny AI Pocket Lab has not yet been announced, and details and its release to end users are expected to become clear at CES 2026.