Qianwen AI Glasses S1 Unveils 3D Display Capability
On May 8, Qianwen AI Glasses S1 received a major upgrade, featuring dual-light machinery and dual-camera stereoscopic imaging technology, which enables 3D display capabilities. This upgrade also brings active service capabilities, including ride-hailing, flash shopping, and trip planning, all of which will be available this month. This update covers two products: Qianwen AI Glasses S1 and Quark AI Glasses S1. Users can experience the new features after upgrading to the latest version.
Anthropic Plans to Raise Billions of Dollars in Funding This Summer
On May 8, reports emerged that Anthropic is considering raising hundreds of billions of dollars in funding this summer to significantly expand its computing power, potentially valuing the company at nearly $1 trillion.
SoftBank in Talks with NVIDIA and Foxconn to Develop “Made in Japan” AI Servers
On May 8, reports indicated that SoftBank Group has begun discussing the development of “Made in Japan” artificial intelligence servers with NVIDIA and Foxconn. The plan is expected to be announced as early as next Monday. SoftBank aims to establish a production system by the end of the decade, initially assembling components from external suppliers before gaining full control over the manufacturing process.
NVIDIA to Invest Up to $2.1 Billion in IREN, Plans to Deploy Large-Scale Data Centers
According to a statement on May 7, data center developer IREN has agreed to grant NVIDIA a five-year option to purchase up to 30 million shares at an exercise price of $70 per share. The two companies will also collaborate on deploying computing power worth tens of billions of dollars. As part of the agreement, NVIDIA will invest up to $2.1 billion in IREN. Additionally, the companies announced a $3.4 billion AI cloud services contract to acquire and deploy NVIDIA’s Blackwell processors, with plans to increase the infrastructure to up to 5 gigawatts in the future.
OpenAI Introduces “Trusted Contacts” Safety Feature
On May 7, OpenAI announced the launch of its “Trusted Contacts” feature. Specifically, if OpenAI’s automated systems and trained reviewers detect that an adult user may have discussed self-harm, which could pose a serious safety risk, the user’s trusted contacts will be notified. The feature is not intended to replace professional care or crisis intervention services but rather to provide additional safeguards for users in distress. ChatGPT will continue to encourage users to contact crisis hotlines or emergency services when necessary.
Yingtai Intelligent Collaborates with Google Cloud
Recently, Yingtai Intelligent, a clinical-stage biopharmaceutical company driven by generative AI, announced a collaboration with Google Cloud to accelerate its drug discovery process. According to Yingtai Intelligent, this partnership will combine its Pharma.AI platform with Google Cloud’s scalable computing infrastructure and advanced AI capabilities to enhance the speed and efficiency of its drug discovery process.
China Mobile to Open Trillion-Level Token Service Experience Package, Upgrades MoMA Model Service Platform
On May 8, China Mobile held its 2026 Mobile Cloud Conference in Suzhou. The company’s chairman, Chen Zhongyue, stated that communication services, computing services, and intelligent services have been identified as the company’s main businesses. The announced plans include upgrading the mobile model service platform MoMA, launching the mobile intelligent body framework MobileClaw, building a gigawatt-level AI data park, and opening a trillion-level Token service experience package. China Mobile’s cloud general manager, Sun Shaoling, introduced MoMA, which integrates multiple models, including MiniMax, DouBao, and GLM, and supports unified API calls, small and large model collaboration, and multi-dimensional strategy scheduling. The platform provides three levels of Token services: universal, premium, and confidential. The reasoning engine has been optimized through technologies such as deep operator tuning, domestic multi-chip adaptation, PD heterogeneous separation, and KV cache pooling, resulting in a 20% increase in throughput, a 40% reduction in latency, and a 30% reduction in cost.
