The release of powerful open-weight LLMs like Kimi K3, Qwen3.8, Bonsai, and GLM is reshaping the AI race, challenging proprietary giants and enabling broad local deployment. This surge is accompanied by debates on model control, licensing, and the so-called 'AI communism' critique.
An open source meetup in Shanghai showcased Qwen's new 3.8 release. More than a technical update, it signals a structural shift in open AI's center of gravity toward China, with tangible consequences for those opting for on-premise deployment and dat...
OpenAI’s head of strategic futures compared the rise of open-weight models to communism, sparking debate on digital sovereignty, cost, and control in the LLM ecosystem.
Chinese startup Moonshot AI aims to list in Hong Kong within six months at a valuation above $30 billion. The figure hides a geopolitical bet: capital to sustain the hardware race amid US export curbs, forcing a shift toward local stacks and data sov...
Dean W. Ball of OpenAI analyzes China's Kimi model, revealing a paradox: open-weight can slow CapEx and push toward state-controlled public infrastructure, potentially countered by US strategic regulatory friction.
Moonshot AI updates Kimi, reigniting the debate around so-called 'full AI communism'. The phrase, loaded with politics rather than engineering substance, forces a reflection on open source, data sovereignty, and on-premise hardware.
The pace of open-source releases, with models like Minimax 3 Pro at 2.7 trillion parameters and GLM 5.3, marks a turning point. As enterprise trust in closed vendors wanes — forced to “distill” client knowledge to justify trillion-dollar valuations —...
Axios reports that the Chinese open-source model Kimi K3 matches Anthropic’s Claude Opus, erasing the US lead. A turning point for on-premise deployment and data sovereignty.
Chinese startup Moonshot AI has unveiled Kimi K3, a 2.8-trillion-parameter LLM with open weights. The model aims to compete with American labs and comes amid reports of a $30 billion valuation. But who can really run it self-hosted? AI-RADAR analysis...
The Chinese Moonshot AI model climbs the independent ranking and overtakes Anthropic. Less mainstream noise, more substance for those eyeing high-performance, cost-efficient self-hosted LLMs.
A German research consortium has released Soofi S, an open 30-billion-parameter LLM that tops benchmarks in both English and German. The model marks a step forward for European digital sovereignty, offering a viable path to on-premise self-hosting wi...
Hugging Face CEO Clem Delangue sees enterprises increasingly turning to open models for cost, accessibility, and data ownership. If production AI runs on open source, frontier models risk becoming less relevant, with deep consequences for hardware, l...
