Tether Data, the technology arm of the world’s largest stablecoin issuer, is expanding into artificial intelligence by launching what it calls the world’s largest synthetic dataset for STEM-focused AI models.
The company announced today QVAC Genesis Ia 41 billion token dataset built to train science and engineering-oriented language models, and QVAC workbencha cross-platform local AI application that runs models directly on consumer devices. According to Tether, QVAC stands for “QuantumVerse Automated Computer.”
The mission statement on QVAC’s website states, “QVAC is Tethered’s answer to centralized AI. An entirely new paradigm where intelligence is executed privately, locally, and without permission.” “There are no clouds. There are no gatekeepers. Just you, your machine, and your unstoppable intelligence.”
The move marks a surprising escalation of Tether’s ambitions beyond finance. The company said the dataset has been validated against mathematical, physics, biological and medical benchmarks and is designed to “level the playing field” for open source AI by providing researchers with an alternative to proprietary data maintained by companies like OpenAI and Google.
First launch of Tether AI.
Local, on-device AI is evolving.– QVAC Workbench, a mobile/desktop app to use and experiment with many AI models locally on your device in 100% privacy
– QVAC Genesis I, the largest synthetic pre-training dataset ever for large-scale language models (LLMs). … https://t.co/79lYhsobuc— Paul Ardino y (@paololay) October 24, 2025
Although QVAC Genesis I is not itself a financial product, the broader QVAC ecosystem is being built with clear links to Tether’s crypto infrastructure. Tether said in a previous company statement that QVAC’s architecture will eventually integrate Bitcoin and its own stablecoin, USDT, allowing AI agents to autonomously trade using digital assets.
This suggests that this effort could evolve beyond data and local AI tools into a network where intelligent agents can not only learn and reason, but also pay, transact, and interact directly through blockchain rails.
“Information should not be centralized,” he said. Paolo Ardoinosaid Tether’s CEO in a statement accompanying the release. “With QVAC Workbench and Genesis I, we open the door to infinite intelligence: AI that lives, learns, and evolves locally on your device.”
QVAC Consumer App
The company also released a free consumer app called QVAC Workbench. It will also be available for smartphones (Android for now, iOS “in the next few days”), and desktop platforms (Windows, macOS, Linux). ”With QVAC Workbench, all chats and interactions with your AI models stay local, on-device, and your data remains owned by you and 100% private,” the company said.
It also introduces a peer-to-peer feature called delegated inference, which allows mobile apps to offload heavy computation to desktop workstations while keeping all data private and local.
Unlike traditional training materials collected from the public internet, QVAC’s datasets are fully synthetic, generated, filtered, and validated by models trained on educational and scientific materials. Tether claims this data allows its models to “reason, problem solve, and think critically” rather than simply mimicking text patterns. Full technical details are available on the QVAC Research Blog.
of QVAC workbench This app allows users to run large language models such as Llama, MedGemma, and Qwen all on their phones and computers.
Tether assembled this twin release as part of a larger effort to create “local intelligence,” or AI that operates independently of cloud servers. The company already dominates the stablecoin market with its USDT token and is positioning its AI division. tether dataas an advocate of decentralized infrastructure that keeps both funds and information outside of corporate control.
Who owns your AI?
The project comes amid a growing debate over the role of synthetic data in model training. Despite promises of privacy and scalability, skeptics warn that synthetic training data can amplify the biases and errors of the parent model and lock in distorted inference patterns. Tether’s announcement did not specify which generation system generated the Genesis I content or how its quality was ensured.
Still, QVAC Genesis I is one of the boldest open data experiments ever by a private cryptocurrency company. If Tether’s claims hold up under intense scrutiny, they could give independent researchers and small research institutes a new foothold in the AI race, demonstrating Tether’s determination to influence not just finance but the future of artificial intelligence itself.
