The insatiable demand for AI infrastructure has triggered a severe global memory crisis, ending the era of cheap storage and forcing consumer tech giants like Apple and AMD to hike hardware prices. Meanwhile, the AI sector is experiencing a massive talent reshuffle and escalating geopolitical tensions, as top researchers flee Google for rivals and companies accuse each other of aggressive model extraction.
The erratic push-and-pull of US AI policy reached a boiling point this week as the Trump administration rapidly backtracked on export bans for Anthropic’s flagship models, officially clearing both Mythos 5 and Fable 5 for broader distribution. This frantic deregulation comes precisely as US sanctions on Chinese tech show glaring vulnerabilities; Chinese researchers revealed their open-weight GLM-5.2 matches Anthropic’s models in cybersecurity, and China successfully built the world’s fastest supercomputer entirely without GPUs. Sweeping hardware embargoes seem to be merely incentivizing China to build resilient, decentralized infrastructure, forcing the US to realize that stifling domestic AI deployment in the name of national security might just hand global dominance to its biggest rival.
OpenAI partnered with AMD, Broadcom, Intel, Microsoft, and NVIDIA to release Multipath Reliable Connection (MRC), a new open networking protocol that reduces wasted GPU time and accelerates large AI training clusters. The protocol is already deployed across OpenAI’s frontier supercomputing sites and is being made available to the entire industry via the Open Compute Project.
According to a cnbeta report, China’s entirely domestically built “LingSheng” supercomputer has reclaimed the global #1 spot on the Top500 list with a sustained performance of 2.19 EFlops. The system introduces an “Online Acceleration” CPU architecture that integrates AI matrix acceleration units, marking a significant milestone in Sino-US supercomputing competition and enabling massive AI-driven scientific research without relying on traditional heterogenous GPU architectures.
Google’s stock suffered a massive drop after two key AI figures—AlphaFold co-creator John Jumper and Transformer co-author Noam Shazeer—left for Anthropic and OpenAI respectively, as reported in this cnbeta coverage. Expanding its own AI footprint, Qualcomm is in advanced talks to acquire AI chip startup Modular for approximately $4 billion to challenge Nvidia’s infrastructure dominance.
US tech sanctions on China are showing serious cracks across multiple fronts. Today, Chinese researchers revealed that their open-weight GLM-5.2 AI model matches Anthropic’s Mythos in cybersecurity, while the country simultaneously reclaimed the title of the world’s fastest supercomputer with a system built entirely without GPUs.