China’s Zhipu Moves to Attract Claude Users Amid Anthropic Restrictions
Chinese AI startup Zhipu has announced support for users of Anthropic’s Claude API who may lose access after reports that Anthropic will restrict Chinese-run entities. In a WeChat post, Zhipu said Claude users could easily switch to its GLM-4.5 model by updating the API URL. To lure developers, the company is offering 20 million free tokens and a coding package it claims costs just one-seventh of Claude’s while delivering three times the usage.
Temu Partners with Horoz Lojistik for Heavy Deliveries in Turkey
Turkey’s Horoz Lojistik has struck a strategic partnership with Chinese e-commerce platform Temu to handle local deliveries, focusing on high-volume orders and packages over 30 kilograms. The deal is designed to ensure faster and more secure shipping for Temu customers across Turkey, while creating synergies in logistics operations.
China’s AI Agents Could Start Generating Revenue by 2026
UBS predicts Chinese companies may begin monetizing AI agents—autonomous software that plans and executes complex tasks—by next year, with 2026 marking a “year of agent monetisation.” Equity strategist Sundeep Gantori said breakthroughs in advanced models like DeepSeek’s R2 will drive the shift. While the U.S. AI agent market already generates $15–20B annually through applications in coding, wealth management, and automation, China lags due to weaker enterprise adoption and consumer-focused use cases.
Scientists Create Syn57, a “Lean-Coded” Synthetic Lifeform
Researchers at Cambridge’s Medical Research Council Lab have engineered Syn57, a synthetic strain of E. coli with just 57 codons—seven fewer than all natural life forms, which use 64. Codons are three-letter DNA sequences that instruct amino acid production, but many are redundant. By rewriting more than 101,000 lines of genetic code and removing these redundancies, scientists demonstrated life can function more efficiently. Earlier milestones included the 2010 creation of a synthetic cell and a 2019 61-codon E. coli. Advances in DNA synthesis now allow genomes to be built from scratch, opening new doors to test alternative genetic codes.
OpenAI to Launch AI-Powered Jobs Platform, Challenging LinkedIn
OpenAI announced plans for an “OpenAI Jobs Platform” to connect candidates with employers using AI, potentially competing with Microsoft’s LinkedIn despite their uneasy partnership. Slated for mid-2026, the platform will also support local businesses and governments seeking AI talent. Alongside, OpenAI will expand its Academy with certification programs covering AI fluency through advanced skills like prompt engineering, aiming to certify 10 million Americans by 2030.
Apple Faces Lawsuit Over Alleged Use of Authors’ Books in AI Training
Apple has been hit with a proposed class-action lawsuit in California, accused of using copyrighted books without permission to train its “OpenELM” AI models. Authors Grady Hendrix and Jennifer Roberson allege their works were part of a pirated dataset Apple exploited without credit or compensation. The case adds to a growing wave of copyright battles in the AI sector, with Microsoft, Meta, and OpenAI also facing claims. Recently, Anthropic agreed to a $1.5B settlement with authors over similar allegations—the largest reported copyright recovery to date.
EU Court Rejects Zalando’s Challenge to Digital Services Act
Zalando lost its bid to overturn its designation as a “very large online platform” (VLOP) under the EU’s Digital Services Act (DSA), marking the first company challenge to the landmark law. The General Court ruled Zalando has more than 83 million monthly active users, far above the 30 million threshold it claimed, making it subject to stricter content moderation rules alongside Google and Meta. Zalando argued its curated marketplace posed no systemic risk of harmful content but plans to appeal.
