Tiangong AI Hits $800M Annual Revenue; NVIDIA Introduces AI Factory Revenue-Sharing Model; Meta Outsourced Security Testing, Report Says · 0703-441
Editorial standards and source policy: Editorial standards, Team. Content links to primary sources; see Methodology.
The AI industry is rapidly shifting from technological acceleration to value redefinition:
- Kunlun Tech’s AI division (Tiangong AI) has surpassed $800 million in ARR, nearing China’s first non-BAT company to hit the $1 billion ARR milestone [9];
- NVIDIA unveiled a new AI infrastructure partnership model based on revenue sharing, reshaping the commercial logic of foundational AI infrastructure [14];
- Meta was reportedly outsourcing safety testing—feeding dangerous prompts to rival models via third-party contractors—exposing serious ethical risks in AI safety evaluation [0].
🚀 Top Developments
- Meta outsourced staff to impersonate minors while testing competitors’ AI [0]: Sent self-harm and violent prompts to ChatGPT, Gemini, and others—turning safety evaluation into a weaponized competitive tactic.
- Tiangong AI’s annual recurring revenue exceeds $800 million [9]: Over $700 million came from its AI short-video platform; it’s now racing toward becoming China’s first non-BAT AI company with $1B+ ARR.
- NVIDIA launches a new AI infrastructure collaboration model [14]: Partnering with Sharon AI and Firmus to co-build “AI Factories” using revenue-sharing + credit support.
- OpenAI proposes transferring 5% equity to establish a public wealth fund [12]: A formal mechanism to let everyday citizens share in AI’s economic upside.
- “Supervisory Engineering” emerges as a new software layer in the AI programming era [2]: Engineers’ core role now centers on guiding, evaluating, and correcting AI agent outputs.
- Aether AI unveils its causal world modeling roadmap [15]: By explicitly learning causal variables and dynamics, it aims to shift embodied AI from correlation-based fitting toward true physical understanding.
- Goldman Sachs counters three major AI bubble concerns [8]: Highlights infrastructure, power, and cloud services as the three most certain investment themes today.
- Flowith Matrix achieves fully autonomous operation of YouTube kids’ channels [4]: A multi-agent architecture proving AI agents’ viability in content OPC (Operations, Production, and Commercialization) use cases.
🔗 Sources
[0] These ChatGPT “failures”? Meta hired contractors to generate them — https://www.bestblogs.dev/article/d64aca4b?utm_source=rss&utm_medium=feed&utm_campaign=resources&entry=rss_article_item
[2] Supervisory Engineering: The “Middle Ring” of Orchestration Software — https://www.bestblogs.dev/article/63cf901d?utm_source=rss&utm_medium=feed&utm_campaign=resources&entry=rss_article_item
[4] Flowith Launches Yet Another New Product—Matrix: Zero-Human OPC? — https://www.bestblogs.dev/article/a0e06df8?utm_source=rss&utm_medium=feed&utm_campaign=resources&entry=rss_article_item
[8] Goldman Sachs Refutes Three Major AI Bubble Concerns — https://www.bestblogs.dev/article/573efda0?utm_source=rss&utm_medium=feed&utm_campaign=resources&entry=rss_article_item
[9] Tiangong AI Surpasses $800M ARR, Advancing Toward China’s First Non-BAT $1B+ AI Company — https://www.bestblogs.dev/article/7a77e7b9?utm_source=rss&utm_medium=feed&utm_campaign=resources&entry=rss_article_item
[12] OpenAI Proposes Transferring 5% Equity to the U.S. Government to Establish a Public Wealth Fund — https://www.bestblogs.dev/status/2072550767320285654?utm_source