Tether, issuer of the world’s largest stablecoin by market capitalization, USDT, has released a new AI training framework that it says helps fine-tune large language models on consumer hardware, including smartphones and non-Nvidia GPUs.
According to Tuesday announcementThe system, part of its QVAC platform, uses Microsoft’s BitNet architecture and LoRA techniques to reduce memory and compute requirements, potentially reducing costs and hardware barriers to developing AI models.
The framework supports cross-platform training and inference on a range of chips, including AMD, Intel and Apple Silicon, as well as mobile GPUs from Qualcomm and Apple.
Tether said its engineers were able to refine models with up to 1 billion parameters on smartphones in less than two hours, and smaller models in minutes, with support expanding to models of up to 13 billion parameters on mobile devices.
Built on BitNet, a 1-bit model architecture, the framework can reduce VRAM requirements by up to 77.8% compared to similar 16-bit models, according to the company, allowing larger models to run on limited hardware. It also allows fine-tuning of LoRA on non-Nvidia hardware for 1-bit models, extending support beyond GPUs typically used for AI training.
The company said the performance gains extend to inference, with mobile GPUs running BitNet models several times faster than CPUs. He also highlighted use cases such as on-device training and federated learning, where models can be updated across distributed devices without sending data to centralized servers, potentially reducing reliance on cloud infrastructure.
Related: New Messari CEO banks on AI as company downsizes
Crypto Companies Are Getting Into AI, From Mining Infrastructure to Autonomous Agents
Tether’s move into AI infrastructure comes as crypto companies expand into computing and machine learning, with activity accelerating in Bitcoin mining and the rise of AI agents.
In September, Google took a 5.4% stake in Cipher Mining as part of a 10-year, $3 billion deal tied to AI data center capacity. In December, Bitcoin miner IREN announced plans to raise approximately $3.6 billion to finance AI infrastructure.
The trend continued through 2026. In February, HIVE Digital Technologies reported record revenue of $93.1 millionfueled by the growth of its AI and high-performance computing (HPC) businesses, while Core Scientific secured a $500 million loan from Morgan Stanley in March, with the possibility of expanding it to $1 billion.
The mining industry’s shift toward AI and HPC comes as AI agents, autonomous programs capable of transacting, interacting with services, and performing tasks, are gaining momentum in the crypto industry.
In October, Coinbase introduction of wallet infrastructure enabling AI agents to perform on-chain transactions. Last month, Alchemy launched a system allowing agents to access blockchain data services using USDC on Base. Also in February, Pantera and Franklin Templeton joined Arena, a Sentient platform for testing enterprise AI agents.
On Tuesday, World, the identity network co-founded by OpenAI’s Sam Altman, Launch of AgentKita toolkit that allows AI agents to verify that they are linked to a unique human using World ID capabilities while making payments via the x402 micropayment protocol.
Review: All 21 million Bitcoins are threatened by quantum computers
