News

Public · Published

Tether Data Pushes AI Off the Cloud With New 460M Parameter Vision Model

On Wednesday, Tether Data's AI research arm QVAC open‑sourced VisionPsy‑Nano, a 460‑million‑parameter vision‑language model that runs on smartphones without sending data to cloud servers.

Published:

Updated:

What happened

On Wednesday, Tether Data’s AI research arm QVAC open‑sourced VisionPsy‑Nano, a 460‑million‑parameter vision‑language model that runs on smartphones without sending data to cloud servers.

Confirmed

Global impact / market context

Running AI on phones reduces reliance on expensive cloud infrastructure and lowers data‑transfer costs, which can improve user privacy and make AI features more accessible to a wider audience.

Analyst inference

Edge‑computing is becoming popular as companies try to run AI directly on devices to cut cloud costs and improve privacy, and Tether’s new model follows this broader shift.

Analyst inference

What to watch

  1. Adoption by app developers: if popular apps integrate VisionPsy‑Nano, demand for on‑device AI could rise, boosting related hardware sales. Analyst inference
  2. Privacy regulations: stricter data‑privacy rules may favor on‑device processing, encouraging more firms to use models like VisionPsy‑Nano. Analyst inference
  3. Competitive response: other AI providers might release similar edge models, influencing the pace of on‑device AI development. Analyst inference

Affected assets

  • USDT — Tether

Evidence