{"id":18397,"date":"2026-03-17T22:52:40","date_gmt":"2026-03-17T22:52:40","guid":{"rendered":"https:\/\/cryptoted.net\/index.php\/2026\/03\/17\/tethers-qvac-pushes-multi-billion-parameter-ai-models-onto-phones-and-consumer-gpus\/"},"modified":"2026-03-17T22:52:40","modified_gmt":"2026-03-17T22:52:40","slug":"tethers-qvac-pushes-multi-billion-parameter-ai-models-onto-phones-and-consumer-gpus","status":"publish","type":"post","link":"https:\/\/cryptoted.net\/index.php\/2026\/03\/17\/tethers-qvac-pushes-multi-billion-parameter-ai-models-onto-phones-and-consumer-gpus\/","title":{"rendered":"Tether\u2019s QVAC pushes multi\u2011billion\u2011parameter AI models onto phones and consumer GPUs"},"content":{"rendered":"<p> <br \/>\n<br \/><img decoding=\"async\" src=\"https:\/\/crypto.news\/app\/uploads\/2025\/04\/crypto-news-Corruption-and-money-laundering-in-Tether-option07-1380x820-1-1.webp\" \/><\/p>\n<div>\n<p class=\"is-style-lead\">Tether\u2019s QVAC Fabric integrates BitNet LoRA to fine\u2011tune and run multi\u2011billion\u2011parameter AI models on consumer GPUs and flagship phones, pushing serious AI work to the edge.<\/p>\n<div id=\"cn-block-summary-block_5a8c0cc054d306c3dd378478eedd5fd2\" class=\"cn-block-summary\">\n<p>\n        <span class=\"tabs__item is-selected\">Summary<\/span>\n    <\/p>\n<div class=\"cn-block-summary__content\">\n<ul class=\"wp-block-list\">\n<li>QVAC Fabric brings BitNet LoRA fine\u2011tuning and inference to AMD and Intel GPUs, Apple\u2019s Metal stack, and high\u2011end mobile GPUs, claiming 2\u201311x speedups over CPU baselines and up to 90% lower memory use.<a href=\"https:\/\/crypto.news\/wp\/wp-admin\/post-new.php\" target=\"_blank\"\/>\u200b<\/li>\n<li>Tether says it has fine\u2011tuned models up to 3.8 billion parameters on Pixel 9, Galaxy S25, and iPhone 16, and up to 13 billion parameters on iPhone 16, pushing on\u2011device AI far beyond today\u2019s typical sub\u20113B demos.<a href=\"https:\/\/crypto.news\/wp\/wp-admin\/post-new.php\" target=\"_blank\"\/>\u200b<\/li>\n<li>The release fits Tether\u2019s pivot from pure stablecoin issuer to infrastructure player, complementing earlier QVAC initiatives like the 41\u2011billion\u2011token Genesis I dataset and local AI Workbench to challenge Big Tech\u2019s AI moat.<\/li>\n<\/ul><\/div>\n<\/div>\n<p><!-- .cn-block-summary --><\/p>\n<p class=\"is-style-default\">Tether\u2019s AI division has quietly shipped one of its most aggressive non\u2011stablecoin bets to date: a cross\u2011platform BitNet LoRA framework, integrated into its QVAC Fabric stack, that can train and run multi\u2011billion\u2011parameter language models directly on consumer\u2011grade GPUs and flagship smartphones. If the numbers hold up outside Tether\u2019s own benchmarks, this pushes on\u2011device AI from \u201ccute demo\u201d territory into something systemically relevant for both hardware vendors and crypto\u2011aligned infra investors.<\/p>\n<p>    <!-- .cn-block-related-link --><\/p>\n<p>The new QVAC Fabric release brings BitNet LoRA fine\u2011tuning and inference to AMD and Intel GPUs, Apple\u2019s Metal ecosystem, and a range of mobile GPUs in a single framework. Tether claims that, on flagship devices, GPU\u2011based inference is between 2 and 11 times faster than CPU baselines, while memory usage drops by as much as 90% versus full\u2011precision models. In practice, this means you can squeeze significantly larger models, or more concurrent sessions, onto the same hardware envelope\u2014critical for phones and laptops where thermal and RAM ceilings are non\u2011negotiable.<\/p>\n<figure class=\"wp-block-embed is-type-rich is-provider-twitter wp-block-embed-twitter\">\n<div class=\"wp-block-embed__wrapper\">\n<blockquote class=\"twitter-tweet\" data-width=\"550\" data-dnt=\"true\">\n<p lang=\"en\" dir=\"ltr\">Tether AI breakthrough<\/p>\n<p>Tether AI team just released new version of QVAC Fabric  to include the World\u2019s First Cross-Platform BitNet LoRA Framework to Enable Billion-Parameter AI Training and Inference on Consumer GPUs and Smartphones.<\/p>\n<p>Background<br \/>Microsoft&#8217;s BitNet uses one bit\u2026 <a rel=\"nofollow\" href=\"https:\/\/t.co\/ooy3L6775E\">https:\/\/t.co\/ooy3L6775E<\/a><\/p>\n<p>\u2014 Paolo Ardoino \ud83e\udd16 (@paoloardoino) <a href=\"https:\/\/twitter.com\/paoloardoino\/status\/2033894861783376196?ref_src=twsrc%5Etfw\" target=\"_blank\" rel=\"nofollow\">March 17, 2026<\/a><\/p><\/blockquote>\n<\/div>\n<\/figure>\n<p>The headline numbers are provocative: Tether\u2019s team <a href=\"https:\/\/crypto.news\/tethers-ardoino-hints-at-challenging-microsoft-and-amazon-in-ai-sector\/\">says<\/a> it has completed fine\u2011tuning of models up to 3.8 billion parameters on devices like the Pixel 9, Galaxy S25, and iPhone 16, and has pushed fine\u2011tuning to as large as 13 billion parameters on the iPhone 16 specifically. That is a sharp <a href=\"https:\/\/crypto.news\/tether-signs-mou-with-georgia-for-bitcoin-web3-and-p2p-development\/\">escalation<\/a> from the current norm, where most \u201con\u2011device AI\u201d marketing still revolves around sub\u20113B parameter models or offloads heavier workloads to the cloud. If reproducible, this suggests a future where serious personalization and domain\u2011specific adaptation can happen locally, without shipping user data off\u2011device.<a href=\"https:\/\/www.chaincatcher.com\/en\/article\/2252528\" target=\"_blank\" rel=\"nofollow\"\/>\u200b<\/p>\n<p>Strategically, this fits <a href=\"https:\/\/crypto.news\/tether-releases-41-billion-token-dataset-to-democratize-ai-training\/\">Tether\u2019s <\/a>ongoing pivot from pure stablecoin issuer to broader infrastructure operator. The company has already plowed billions into energy, mining, and media; now it is adding edge\u2011AI tooling to the portfolio, with the related QVAC and BitNet LoRA code open\u2011sourced on GitHub for developers to inspect and build on. Open sourcing is not altruism\u2014it is distribution. If QVAC becomes a default path for indie devs and small labs to push models onto consumer hardware, Tether buys cultural and technical relevance in a stack that sits well outside banking regulation\u2019s direct line of fire.<a href=\"https:\/\/www.chaincatcher.com\/en\/article\/2252528\" target=\"_blank\" rel=\"nofollow\"\/>\u200b<\/p>\n<p>For markets, the immediate impact is narrative, not P&amp;L. There is no token here, no obvious \u201cfarm this yield\u201d angle. But there is a clear macro story: as more AI work migrates to the edge, infrastructure power shifts from centralized hyperscalers toward whoever controls key toolchains and hardware abstraction layers. Tether is signaling that it intends to be one of those players, leveraging its balance sheet to seed primitives that reduce dependence on any single cloud or jurisdiction. For crypto, an ecosystem increasingly obsessed with AI\u2011adjacent plays, this is a reminder that not every serious bet needs a ticker symbol attached.<a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/www.chaincatcher.com\/en\/article\/2252528\"\/>\u200b<\/p>\n<p>For now, the obvious questions are technical: how BitNet LoRA\u2019s claimed speedups and memory reductions compare against incumbents like llama.cpp, MLC, or Qualcomm\u2019s own SDKs on the same devices; what the energy and thermal trade\u2011offs look like in real\u2011world use; and how permissive the licenses are for commercial deployment. But if even a conservative slice of Tether\u2019s claims prove out under independent benchmarking, QVAC Fabric\u2019s BitNet LoRA integration will mark a tangible step toward turning high\u2011end smartphones into viable training and inference rigs for mid\u2011sized language models\u2014shifting AI one notch closer to the edge, and giving Tether yet another foothold in critical digital infrastructure.<\/p>\n<p>    <!-- .cn-block-related-link --><\/p><\/div>\n<p><script async src=\"\/\/platform.twitter.com\/widgets.js\" charset=\"utf-8\"><\/script><br \/>\n<br \/><br \/>\n<br \/><a href=\"https:\/\/crypto.news\/tethers-qvac-pushes-multi-billion-parameter-ai-models-onto-phones-and-consumer-gpus\/\">Source link <\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Tether\u2019s QVAC Fabric integrates BitNet LoRA to fine\u2011tune and run multi\u2011billion\u2011parameter AI models on consumer GPUs and flagship phones, pushing serious AI work to the edge. Summary QVAC Fabric brings BitNet LoRA fine\u2011tuning and inference to AMD and Intel GPUs, Apple\u2019s Metal stack, and high\u2011end mobile GPUs, claiming 2\u201311x speedups over CPU baselines and up [&hellip;]<\/p>\n","protected":false},"author":5,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"tdm_status":"","tdm_grid_status":"","footnotes":""},"categories":[23],"tags":[],"kronos_expire_date":[],"class_list":["post-18397","post","type-post","status-publish","format-standard","hentry","category-crypto"],"_links":{"self":[{"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/posts\/18397","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/users\/5"}],"replies":[{"embeddable":true,"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/comments?post=18397"}],"version-history":[{"count":0,"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/posts\/18397\/revisions"}],"wp:attachment":[{"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/media?parent=18397"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/categories?post=18397"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/tags?post=18397"},{"taxonomy":"kronos_expire_date","embeddable":true,"href":"https:\/\/cryptoted.net\/index.php\/wp-json\/wp\/v2\/kronos_expire_date?post=18397"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}