Discover Tiiny AI’s Tiny Supercomputer: A Powerhouse for 120B AI Models in Your Palm
In the rapidly evolving world of artificial intelligence, a new player is making waves with a groundbreaking innovation. Tiiny AI, a forward-thinking startup, has unveiled what it claims to be the world’s smallest AI supercomputer. This compact device, known as the Tiiny AI Pocket Lab, promises to democratize access to powerful AI capabilities, making it accessible to a wider range of users.
Tiiny’s Revolutionary AI Pocket Lab
As edge AI becomes increasingly vital in the computing landscape, the challenge has been to provide affordable yet powerful hardware to run local AI models. High-end devices, like NVIDIA’s DGX Spark, often prove too costly for individual consumers. Tiiny AI aims to change this narrative with its Pocket Lab, a cost-effective and ultra-compact supercomputer that packs a punch in AI processing power.
Despite its diminutive size of just 14.2 × 8 × 2.53 cm and a weight of 300g, the Pocket Lab boasts the ability to support a 120-billion-parameter model. This is a remarkable feat in the realm of AI, offering capabilities for advanced reasoning and deep analysis. The device is built for on-device operations, catering to both casual users and AI enthusiasts eager to explore local large language model (LLM) deployments.
| Category | Specification |
|---|---|
| Processor | ARMv9.2 12-core CPU |
| AI Compute Power | Custom heterogeneous module (SoC + dNPU), ≈ 190 TOPS |
| Memory & Storage | 80GB LPDDR5X RAM + 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, ~300g (pocket-sized) |
| Ecosystem | One-click deployment for dozens of open-source LLMs & agent frameworks |
| Connectivity | Fully offline operation — no internet or cloud required |
Impressive Features and Capabilities
The AI Pocket Lab supports an array of models, including those from GPT-OSS, Llama, Qwen, DeepSeek, Mistral, and Phi. With its discrete NPU, the device achieves 190 TOPS, while its 80 GB LPDDR5X RAM facilitates aggressive quantization for smooth operation of a 120B model locally. Tiiny AI has also integrated two innovative techniques:
TurboSparse, a neuron-level sparse activation technique, significantly improves inference efficiency while maintaining full model intelligence.
PowerInfer, an open-source heterogeneous inference engine with more than 8,000 GitHub stars, accelerates heavy LLM workloads dynamically distributing computation across CPU and NPU, enabling server-grade performance at a fraction of traditional power consumption. Together, these technologies allow Tiiny AI Pocket Lab to deliver capabilities that previously required professional GPUs costing thousands of dollars.
The device is set to be showcased at CES 2026, marking a notable moment for this promising technology. Though release dates and retail availability remain under wraps, anticipation is building for Tiiny AI’s industry debut.