Full Summary
This Tuesday morning, the AI landscape is a volatile mix of groundbreaking launches and unexpected shutdowns, with a new focus on regulation and real-world application. First, a significant shake-up in the AI application market is underway. Both 36Kr and The AI Journal confirm that projects are being discontinued, with some lasting only 180 days. OpenAI is reportedly discontinuing its Sora video generator just six months after launch due to declining downloads and high computing costs. Additionally, the AI model evaluation platform Yupp.ai, despite raising $33 million, announced its closure in March 2026, with its founder citing rapid model improvements making crowdsourced evaluation less relevant. This signals a shift from testing to intense commercial screening for AI applications. But then, major players continue to innovate. Alibaba Group, as reported by NewsBytes and Blockonomi, has launched its first AI models specifically designed for robots, including RynnBrain for spatial perception and Qwen3.7-Max for autonomous AI agent operations. These aim to enable robots to plan and perform multi-step actions, integrating AI into physical machines beyond just chatbots. Meanwhile, Invenci has launched its AI Gateway, confirmed by both The Globe and Mail and The Manila Times. This enterprise platform provides secure and governed access to foundation AI models, allowing organizations to manage permissions, enforce usage standards, and monitor AI use across teams without risking sensitive data. In a move towards regional AI autonomy, TradingView reports that Giotto.ai is opening immediate access to its AI model and operating system for selected institutional partners and public authorities in Europe and Switzerland, with individual users gaining access in July. This aims to provide reliable, portable, and transparent AI under local control. However, the US government is stepping in to regulate. Memeburn reveals that the US Commerce Department has ordered Anthropic to suspend access to its Claude Mythos 5 and Claude Fable 5 AI models for all foreign nationals. This marks the first time the US has applied export controls directly to an AI model, with Anthropic disabling both models globally due to inability to verify user nationality. This action stems from a method found to bypass Fable 5's safety guardrails. This new wave of AI regulation and rapid market shifts means businesses and individuals need to stay vigilant about which AI tools are available, their security, and their compliance with evolving international policies.