The morning briefing.

AI Model Regulation. The US government has lifted export controls on Anthropic's Fable 5 and Mythos 5 AI models, allowing their global availability after a two-week suspension. This decision follows security concerns that the models could be exploited for cyber-attacks. Anthropic stated that a new safety classifier now blocks the identified jailbreak technique in over 99 percent of cases. The incident highlights the ongoing tension between rapid AI development and national security considerations.

AI Investment Boom. Global investment in artificial intelligence continues its rapid ascent, with significant new capital flowing into the sector. Abu Dhabi's MGX has closed one of the largest AI funds ever at $49 billion, backing major players like OpenAI and Anthropic. Concurrently, China's DeepSeek secured a record-breaking $7 billion funding round to invest in domestic AI chips and accelerate AGI development. Prague-based EquiLibre also raised a Series A at a €438 million valuation to scale AI trading agents for Nasdaq.

AI's Societal Impact. The rapid proliferation of AI tools is raising alarms about potential societal consequences, including exacerbating global inequality. A new UN report warns that unequal access to AI infrastructure and models could widen the gap between nations. Meanwhile, Australian creatives are sounding the alarm on copyright issues, as a proposed deal could allow AI companies to mine content in exchange for investment and a compensation fund. The surging demand for AI computing is also pushing global power grids to their breaking point.

Robotics and Enterprise AI. The robotics sector is seeing significant advancements and strategic shifts, particularly in humanoid and enterprise applications. Chinese company UBTECH unveiled its consumer humanoid robot, the U1 Series, with 11,000 pre-orders, signaling a push into the home market. Japan has ambitious plans to deploy 10 million more robots by 2040, including for medical care, leveraging AI. ByteDance's Doubao is also pivoting its AI focus from consumer chat to enterprise tools, mirroring a successful B2B strategy.

Frontier Model Development. The landscape of advanced AI models is evolving, with new strategies for deployment and ethical considerations for development. OpenAI's latest benchmark paper suggests a departure from its single top-tier strategy, revealing plans for three GPT-5.6 Pro model variants. In a notable stance on open-source development, the Godot game engine project announced it will no longer accept AI-authored code contributions, citing concerns about code quality and maintainability.