Something Big Is Coming From Tether’s AI Team, CEO Paolo Ardoino Confirms

Tether CEO Paolo Ardoino announced on March 15 that the company’s AI division will release what he called a “true breakthrough” in the coming week, signaling the next phase of its decentralized intelligence push.

The statement arrived just days after the QVAC team shipped Workbench version 0.4.1, an update that overhauled the app’s interface and expanded its local AI capabilities across desktop and mobile.

Billions in Profit Fund Tether’s AI Ambitions

Tether has spent more than a year building QVAC (QuantumVerse Automatic Computer), a platform designed to run AI models entirely on consumer devices without cloud servers or API keys.

The initiative falls under Tether Data, the company’s technology arm, and represents a sharp pivot from its origins as the issuer of USDT, the largest stablecoin by market capitalization.

That pivot is funded by significant profits. Tether reported over $10 billion in net profit for 2025, driven largely by returns on $141 billion in U.S. Treasury exposure.

The company has funneled excess capital into AI, energy, peer-to-peer communication, and other frontier technology bets.

Ardoino has framed these investments as a challenge to centralized AI providers. In a February Fortune profile, he outlined a vision in which decentralized technology replaces reliance on Big Tech platforms, extending beyond finance into areas such as cloud computing and communications.

QVAC’s Expanding Toolkit

Meanwhile, the QVAC ecosystem has grown quickly since its initial announcement in mid-2025.

In October 2025, Tether Data released QVAC Genesis I, a 41-billion-token synthetic dataset for training STEM-focused AI models, alongside the first version of QVAC Workbench for on-device AI experimentation.

By December, it expanded the dataset to 148 billion tokens with Genesis II and launched QVAC Fabric LLM, an open-source framework that enabled large language model fine-tuning on consumer hardware, including smartphones.

Recently, Ardoino gave a live demonstration of QVAC running as a fully local AI assistant. The demo showed the system performing complex reasoning and task automation on a laptop with a below-average GPU, using the Model Context Protocol (MCP) to interface with third-party tools like Asana.

According to Ardoino, QVAC Workbench had received further improvements, calling it “still very alpha mode” but noting that users could “already feel the true potential” ahead of a full QVAC SDK launch.

The official QVAC account confirmed version 0.4.1 with a redesigned interface, expanded document format support for retrieval-augmented generation, and performance fixes for Android devices, including Samsung and Pixel 10.

What the Breakthrough Could Mean

The timing of Ardoino’s teaser, following a rapid sequence of product updates, suggests the upcoming release may involve a significant expansion of QVAC’s capabilities. Possibilities include:

• The full open-source release of the QVAC assistant framework
• An advanced on-device model optimized for edge inference, or
• Deeper integration between AI agents and Tether’s payment rails, which already support autonomous transactions in Bitcoin (BTC) and USDT.

Tether has also invested in adjacent technologies that could feed into such a release. The company acquired a majority stake in Blackrock Neurotech, a brain-computer interface firm, for $200 million, and invested in Generative Bionics, an Italian robotics startup.

Its EVO division placed fourth in a global brain-computer interface AI benchmark in February 2026.

However, skepticism remains warranted. Tether’s history of transparency concerns, combined with the promotional nature of social media teasers, means the gap between announcement and deliverable will be closely watched.

Whether the forthcoming release matches the scale Ardoino has implied should become clear within days.