1. China’s Weekly Large Model Token Calls Hit 14.19 Trillion, 9 Straight Weeks Leading the World
Latest data from OpenRouter shows that Chinese AI large models recorded 14.19 trillion token calls in the week ending June 30, 2026, marking 9 consecutive weeks leading the global market. The U.S. model share has dropped sharply from 72% one year ago to just 33% now, with DeepSeek-V4-Flash, Tencent Hy3 Preview, MiniMax M3 and Xiaomi MiMo-V2.5 taking the top four spots in global usage rankings. The surge in demand is driven by the models’ unmatched price competitiveness and stable performance for long-running engineering and coding tasks.
2. Zhipu AI’s GLM-5.2 Jumps 11% in Hong Kong Stock Trading, Market Value Tops HK$970 Billion
Zhipu AI’s Hong Kong-listed stock rallied over 11% in intraday trading on June 30, pushing its market capitalization past HK$970 billion. The surge came after the company released GLM-5.2, an open-source model with 1 million token context window that can autonomously run full large-scale engineering projects for hours without human intervention. The model ranked first among all open-source models on the Artificial Intelligence Global Large Model Intelligence Index, scoring 51 points, and saw its weekly token call volume rise 66% month-on-month to 2.11 trillion.
3. DeepSeek Announces Peak-Valley Pricing for V4 API, Launch Timing Aligns with WAIC 2026
DeepSeek revealed that its official V4 model will launch in mid-July, coinciding with the 2026 World Artificial Intelligence Conference (WAIC) held in Shanghai from July 17 to 20. The company will roll out a new peak-valley pricing mechanism after the launch, setting API prices twice as high during peak hours (9:00–12:00 and 14:00–18:00 daily) to optimize resource allocation and improve service stability for global developers.
4. Meta’s Cloud Business Entry Drags Global Tech Stocks, AI Computing Power Trading Heats Up
Meta’s confirmed plan to enter the public cloud business and sell excess AI computing power triggered a broad sell-off across U.S. and Asian tech stocks in the past three days. The move signals a major shift in the global AI infrastructure landscape, as social media giants with massive idle GPU capacity move to directly compete with traditional cloud providers, pushing down inference costs and accelerating the commoditization of AI computing resources.
5. SAP Cuts Hiring and Travel to Fund Aggressive Global AI R&D Push
Europe’s largest software company SAP announced internal cost-saving measures in a staff memo reviewed by Bloomberg, restricting new hiring exclusively to core AI roles and slashing non-essential internal travel. The company is redirecting hundreds of millions of euros in saved funds to accelerate AI product development, fend off new competitors from generative AI startups, and embed native AI capabilities across its entire enterprise software product line.
6. 2026 AI Index Report Warns Evaluation and Governance Infrastructure Is Falling Behind
The 2026 AI Index Report released this week points out a growing structural gap in the global AI industry: while model capabilities are accelerating faster than ever and adoption curves are steeper than any prior technology, the infrastructure for evaluating, governing and trusting AI systems is lagging far behind. Benchmarks are becoming harder at a faster pace than evaluation methods are improving, and transparency is declining as frontier development concentrates in a small number of large players. PwC estimates AI could expand global GDP by nearly 15% by 2035, a scale comparable to the 19th-century industrial revolution.
7. Global H1 2026 Startup Funding Hits $510 Billion, AI Captures Majority of Q2 Capital
Crunchbase data shows global startup investment reached 510billioninthefirsthalfof2026,anewall−timerecord.AI−relatedcompaniesabsorbedthemajorityofQ2capital,with3DfoundationmodeldeveloperTripoAIraisinganother150 million just one month after its 200millionround,andGermanautonomyanddefensefirmQuantumSystemsclosinga1.2 billion Series D. NVIDIA’s revenue-sharing model remains the most influential infrastructure business model driving the current AI boom.
8. MRDIMM Industry Enters Massive Scaling Phase, 2–3 Years of Rapid Growth Ahead
Semiconductor firm Montage Technology confirmed in a recent investor调研 that MRDIMM, the high-bandwidth server memory module designed specifically for AI inference scenarios, is currently in the second-generation large-scale trial phase. The technology effectively solves the memory bandwidth bottleneck that has long limited AI inference throughput, and the company expects the entire MRDIMM industry to enter a period of massive volume growth over the next 2 to 3 years.
9. Zoom Acquires Common Room to Add AI Agent Capabilities to Its Revenue Suite
Zoom announced the acquisition of customer intelligence platform Common Room for an undisclosed sum, integrating buyer signal analysis and RoomieAI autonomous agents directly into its Zoom Revenue Accelerator product. The move transforms Zoom from a pure video conferencing tool into an end-to-end AI-powered enterprise sales and customer engagement platform, targeting millions of small and medium business users worldwide.
10. Symbotic Acquires ARMS Innovations to Build Full AI Warehouse Orchestration Systems
Warehouse robotics leader Symbotic completed the acquisition of ARMS Innovations this week, shifting its product focus from single-task robot execution to fully AI-powered orchestration systems that coordinate people, machines, maintenance workflows and unexpected disruptions. The integrated solution will enable fully autonomous 24/7 warehouse operations, cutting logistics labor costs by an estimated 40% for large retail and manufacturing clients.