a16z Co-creation: Zhipu GLM-5.2 is the first Chinese AI system to fully benchmark against leading US AI models.
Related
The world's first fully automated robotic pharmacy has been launched, securing $12.6 million in funding to accelerate its commercialization.
Mars Finance reports that Queue, a California-based startup, recently launched the world's first fully automated robotic pharmacy. This technology comes at a time of severe labor shortages and a growing "pharmacy desert" in the United States. Simultaneously, the company announced it has successfully raised $12.6 million, which will be used to accelerate the research and deployment of the automated system. (Cailian Press)
Some large-scale AI models in China are 90% cheaper than those in the US; Chinese AI's high cost-effectiveness is capturing the US market.
According to a report by CNBC on July 7th, influenced by the continued price increases of models from leading US AI vendors, Chinese AI large-scale models are rapidly expanding their application scale in US enterprises due to their cost-effectiveness advantage. Industry insiders point out that the performance of some leading open-source and open weighted models in China is currently about 6 to 9 months behind the technology of top-tier US models such as OpenAI and Anthropic, while the price is 60% to 90% lower, and they can cover the vast majority of routine AI tasks, thus gaining popularity among US enterprises. According to statistics from the AI model aggregation platform OpenRouter, since February 8th of this year, the proportion of Chinese AI models used by US enterprises has exceeded 30% weekly, reaching a peak of 46%; while the average proportion in the previous 12 months was 11%. Another industry statistic shows that in the first week of the launch of Zhipu's latest large-scale model GLM 5.2, the daily average number of word calls increased by 27 times and the number of customers increased by 80 times, making it the fastest-deployed model on the platform in 2026; US AI startup Lindy has significantly reduced costs after switching all its AI business to DeepSeek models, and expects to save millions of dollars within a few months. (CCTV Finance)
Huawei releases "Tao's Law" V2 paper: Kirin 2026 data is publicly disclosed for the first time, supplemented with engineering details and measured data.
According to Beating's monitoring, He Tingbo, head of Huawei's semiconductor division, published the second version of his paper, "Time-Shrinking Theory for Multi-Level Electronic Systems" (Tao's Law), on ChinaXiv, the pre-publishing platform of the Chinese Academy of Sciences, on July 3. Compared to the first version released on May 25, the new version supplements the theoretical framework with a large number of engineering implementation details and quantifiable data, forming a complete eight-chapter discourse system. It also adds principles and physical diagrams covering core technologies such as the τ-layered spatiotemporal model, LogicFolding architecture, bonding interface cross-section, Unified Bus interconnect architecture, and Hi-ONE optical engine. At the engineering implementation level, the second version deeply explains the "gear ratio" concept of LogicFolding: when the hybrid bonding pitch is close to the top-level metal wiring size, the 3D design space shifts from the traditional "macroblock-level discrete optimization" to "cell-level continuous optimization," enabling globally optimal vertical logic partitioning and breaking through the limitation of traditional 3D stacking, which can only be layered by functional blocks. Version V2 also provides, for the first time, a mass production test data table, clearly listing the voltage, frequency, normalized power consumption, area, and power density parameters of the Kirin 2026 and the benchmark Kirin 9030 Pro, further verifying the post-Moore's Law scaling theory centered on the time constant τ through actual measurements. --------------------------------- Click the original link below to join the Beating · Lark AI news channel and monitor global AI hotspots and news 24/7.
The first neurodynamic chip has been developed, which will help the development of technologies such as brain-computer interfaces.
According to Mars Finance, a team led by Professor Yang Yuchao of the School of Integrated Circuits at Peking University, in collaboration with a team led by Researcher Song Zhitang of the Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, has successfully developed the world's first neurodynamics system-on-a-chip based on phase-change memristors. This breakthrough reduces the single-step latency of such complex calculations to 2.12 milliseconds, achieving speedups of 50 to 478 times compared to current advanced graphics processing units (GPUs) in tasks such as cortical reconstruction. This represents a significant breakthrough in real-time computing, a bottleneck that has constrained neurodynamics for half a century. The relevant findings were published in *Science* on the morning of the 3rd. (Cailian Press)
Standard Chartered Bank claims to be the first global systemically important bank to offer USDC minting and redemption services.
Odaily Odaily reports that Standard Chartered Bank, in partnership with Circle, has launched a USDC minting and redemption service for institutional clients. Standard Chartered claims to be the first systemically important bank globally to offer this service, and clients can conduct related transactions without needing a separate Circle account.
The CFTC has launched a full investigation into Polymarket, including allegations of wash trading that have impacted the Robinhood event contract ecosystem; Nasdaq has for the first time distributed TotalView market data on-chain via Pyth Network.
According to Mars Finance and BBX data, the prediction market faced a double whammy yesterday, with traditional exchange infrastructure accelerating its on-chain transformation. Key developments include: Robinhood Markets, Inc. (NASDAQ: $HOOD)'s prediction market/event contract ecosystem suffered a double regulatory blow yesterday: First, the U.S. Commodity Futures Trading Commission (CFTC) launched a full investigation into Polymarket (privately held), covering its social media activities and suspected manipulation; second, a Michigan court ruled to prohibit Kalshi (privately held) from offering sports betting services to Michigan residents. While these two events directly target Polymarket and Kalshi, their strategic importance to Robinhood cannot be ignored—Robinhood, through its subsidiary Robinhood Derivatives LLC, offers event contract products linked to KalshiEx LLC or ForecastEx LLC, making it the largest prediction market distribution channel among regulated brokers in the U.S. The CFTC's escalating investigations and enforcement actions against similar platforms will directly impact Robinhood's compliance architecture and product expansion speed; the sector's average daily trading volume in June reached a record high. Nasdaq, Inc. (NASDAQ: $NDAQ) announced yesterday that it has selected Pyth Network (an on-chain price oracle protocol) as its on-chain distribution partner for TotalView (Nasdaq's full market depth data product). This marks the first time Nasdaq has integrated its core institutional-grade market data onto a blockchain network—TotalView provides full-level buy and sell quotes and transaction data for the entire US stock market, historically only available to traditional financial institutions (paid subscriptions). On-chain distribution means that DeFi protocols, decentralized exchanges, and smart contracts can, for the first time, access Nasdaq-level real-time equity market data as an on-chain pricing basis. The Pyth Network token (PYTH) subsequently rose by over 6%, which the market interpreted as a historic convergence of traditional securities market infrastructure and decentralized finance.