Asia's AI and tech news, exclusively curated.
Your daily pulse on AI, technology, innovation, and startups, everything worth knowing across Asia's fast-moving tech scene.
* By clicking the subscribe button you agree to our Privacy Policy and consent to the GMAsia team sending you occasional email updates.
Xi touts Brics cooperation on AI as US titans flag tech risks
President Xi Jinping promoted China's vision for artificial intelligence at the Brics summit. He sought support from the Global South amid tech competition with the US. China will lead an AI Brics "community" and support cooperation in large language models. Xi invited Brics nations to join the World AI Cooperation Organisation, formed recently.
CATL Launches TECTRANS II at IAA Transportation 2026 to Accelerate Global Commercial Vehicle Electrification
CATL unveiled TECTRANS II, its next-generation commercial vehicle battery solution, at IAA Transportation 2026. This modular platform targets global commercial vehicle electrification. It promises to shorten OEM development cycles by 50% and reduce R&D costs by 60% to 70%. The battery offers a gravimetric energy density of 170 Wh/kg and a maximum range of 1,000 kilometers. CATL also signed an MoU with DHL Group to develop Green Freight Corridors across Europe.
BYD to break 2026 overseas sales target after strong run in Brazil, Europe
BYD expects to exceed its 2026 overseas sales target, forecasting nearly two million units this year and 2.5 million for 2027. This revised guidance is 33% higher than the 1.5 million target set in June. Overseas shipments reached almost 800,000 vehicles by June, with monthly volumes projected at 180,000 to 200,000 units in the second half. The company noted shipping capacity constrained earlier sales. BYD generated RMB 20,000 (USD 2,980) profit per vehicle sold outside China in the first half.
Mubadala is backing Luckin Coffee. Could the Middle East be next?
Abu Dhabi sovereign investor Mubadala has committed a USD 1 billion minority investment in Luckin Coffee. This deal, announced September 10, brings a new institutional backer to the Chinese coffee chain. Mubadala will invest alongside Centurium Capital, Luckin’s controlling shareholder. Luckin reported RMB 15.9 billion (USD 2.4 billion) in revenue for Q2 2026, up 28.5% year-on-year. The investment reflects Mubadala’s confidence in China’s consumer sector and Luckin’s international growth prospects.
Anthropic Disrupted Russian Attempt to Develop Autonomous Drone Swarms
Anthropic, an AI safety company, blocked Russian developers from using its Claude coding agent. The developers, designated GTG-27005, were creating software for autonomous kamikaze drones. This system was designed to select and attack targets, including individuals, without human intervention. The operation, known as "DroneDoc" and "Seraphim," began in mid-May 2026 and involved nine blocked accounts. The group also used social media combat footage to train a computer vision system for these drones.
China’s spy chief warns foreign ‘hostile forces’ may use AI to ‘fabricate political rumours’
China's Minister of State Security, Chen Yixin, has warned that foreign "hostile forces" are using AI to create political rumors and spread harmful information. He stated that generative AI technologies like Anthropic's Claude Mythos and OpenAI's ChatGPT-5.5-Cyber pose serious risks to China's critical information infrastructure. Chen emphasized that foreign intelligence agencies are exploiting AI for data mining and profiling to acquire sensitive national data and trade secrets. This warning comes as AI governance is expected to be a topic in upcoming US-China discussions.
With iOS 27, I’m actually using Siri again
Apple has launched iOS 27, featuring a significantly enhanced Siri. The new Siri, powered by Google's Gemini models, moves beyond basic tasks to handle complex requests and on-screen context. This update follows multiple delays since Apple Intelligence was first teased over two years ago. Users can now ask Siri to draft messages, summarize conversations, and identify objects through the Camera app, marking a substantial upgrade in its daily utility.
How China is preparing for the risk of AI escaping human control
China is preparing for AI models escaping human control. Policymakers are concerned about advanced US models like Anthropic’s Mythos and OpenAI’s GPT-5.5-Cyber. China’s State Security Minister Chen Yixin called for strengthened AI security on September 13. The Cyberspace Administration of China released an AI safety framework in September 2024, expanding it in September 2025. This framework addresses scenarios where AI could autonomously replicate and seek power. President Xi Jinping stated in July that AI must remain under human control.
Anthropic Launches Claude Tools for Financial Advisers
Anthropic has launched Claude for Financial Advisors, a new tool designed to streamline workflows for financial professionals. The service integrates with 11 wealth management platforms including BlackRock, Charles Schwab, and Vanguard. It consolidates client data from custodians, portfolio platforms, and CRM systems. This allows Claude to assist with meeting preparation, portfolio reviews, and post-meeting tasks. Advisers must still approve critical tasks like investment recommendations and client communications.
Be the Bank That Scales With AI in 2027
Core banking systems, designed for human users for fifty years, now face a challenge from AI agents. MIT’s Project NANDA found 95% of enterprise generative AI pilots deliver zero measurable return. This is because production requires governed infrastructure and audit-proof APIs, which standard banking cores cannot support. An AI-native core needs a governed data foundation, complete API coverage, and a safe execution layer. This allows AI systems to query data and act on real-time events.
Finastra Owner Weighs Sale or Merger as Valuation Could Reach US$12 Billion
Vista Equity Partners is exploring a sale or merger for its portfolio company Finastra. Morgan Stanley is advising on the early-stage review. Finastra, formed in 2017, provides banking software to over 7,000 financial institutions globally. The company’s valuation could reach US$12 billion, based partly on projected earnings of US$650 million this year. Blackstone has shown early interest as a prospective buyer.
Anthropic Says Chinese AI Firms Used Fake Accounts to Copy Claude
U.S. AI firm Anthropic has accused Chinese AI companies, including Moonshot AI and DeepSeek, of using thousands of fake accounts and illicit intermediary networks to copy its Claude model outputs. Anthropic's threat intelligence report, released on the 10th, details how these firms allegedly routed millions of user queries through "transfer stations" to Claude, then used the responses for model training. Moonshot AI reportedly used thousands of fake accounts to exchange over 23 million communications with Claude between May and July. The report also cited instances where sensitive user data from services like Moonshot AI's Kimi was passed to Claude without user knowledge, raising significant privacy and data security concerns for users in China and beyond.
Anthropic CEO calls on US AI firms to slow frontier model development
Anthropic CEO Dario Amodei has proposed a three-stage plan for US AI firms to slow frontier model development. The plan includes embedding third-party evaluators with employee-like access inside AI companies. OpenAI's Sam Altman, Google DeepMind's Demis Hassabis, and Elon Musk publicly agreed with the proposal. This follows a former employee's accusation that Anthropic was "gambling with lives" over development speed.
macOS 27: new Siri takes on AI productivity apps
Apple has launched macOS 27 Golden Gate, featuring a redesigned Siri AI and new visual intelligence capabilities. The updated Siri can answer contextual questions about on-screen content, summarize information, and perform tasks across files, emails, and messages. Users can invoke Siri anywhere or select specific screen areas for context. The new OS also improves Spotlight Search and Safari's AI-powered page tracking.
The art of not making an AI safety deal with China
The US and China will resume official AI safety talks after a two-year halt. The dialogue could occur this month, potentially alongside a summit between US President Donald Trump and Chinese President Xi Jinping. Beijing fears these talks could become another tool for containing its technological rise. China's state broadcaster affiliate, Yuyuan Tantian, has warned against this outcome, reflecting broader concerns about national security concepts being overstretched.
Chinese researchers chart 5-stage path toward ‘last AI built by humans’
Chinese researchers from ByteDance, Tsinghua University, and the Shanghai Artificial Intelligence Laboratory are working on recursive self-improvement (RSI) for AI. Their joint paper outlines a five-stage roadmap for systems to build better versions of themselves. This initiative aims to automate the AI development lifecycle. It targets a new front in the AI race with the US.
Z.ai reloads with US$5 billion for China's AI race
Chinese AI developer Z.ai plans to raise approximately US$5 billion. The company seeks capital through a Hong Kong share placement and convertible bond sale. This funding will support next-generation model development and computing infrastructure. It also aims to fuel business expansion in China's competitive AI market.
Li Auto to add CALB as third battery supplier, sources say
Li Auto will add China Aviation Lithium Battery (CALB) as its third battery supplier. This diversifies its supply chain, which currently includes CATL and Sunwoda. CALB batteries will first equip Li Auto’s battery electric models, starting with the 2026 Li i6. The i6 is a key model, accounting for over 50% of Li Auto's sales with more than 120,000 deliveries in the first half of this year.
AliExpress expands Brand+ as overseas AI hardware business doubles
AliExpress showcased nine Chinese brands at IFA 2026, marking a strategic shift to help third-party brands expand overseas. The Alibaba Group platform featured AI hardware, smart glasses, and 3D printers. This approach addresses changing cross-border commerce economics, including higher EU tariffs. AliExpress’s Brand+ program contributed to a 97% rise in EU sales in August. AI hardware doubled its overseas business scale year-on-year between January and July.
Kioxia to consider raising US$10 billion in US listing, Bloomberg News reports
Japanese chipmaker Kioxia Holdings is considering raising at least US$10 billion through a US listing of American depositary receipts. The company has engaged lenders including Bank of America, Goldman Sachs, and JPMorgan for an offering that may occur next year. Kioxia aims to boost liquidity in the US and potentially join a semiconductor-focused stock index. The chipmaker's preparations for an ADR listing follow its previous statement in May to broaden its investor base.
Forget AI and embodied intelligence: At IFA 2026, practicality comes first
IFA 2026 in Berlin showcased practical home appliances over speculative AI. Chinese companies comprised nearly half of the 1,900 exhibitors. 928 companies from mainland China and 84 from Hong Kong attended. This marked a 35% increase in Chinese participation from 2025. The focus was on immediate problem-solving rather than advanced AI concepts.
Chinese appliance makers bolster Europe push, eye vertical integration
Chinese appliance makers Haier and Hisense significantly increased their market share in Europe's washing machine and refrigerator sectors. Their combined share of the washing machine market rose from 2.9% in 2016 to 20.5% in 2025. Refrigerator market share similarly climbed from 5.6% to 18.7% over the same period. This expansion relies on local acquisitions and advanced AI-powered products, like Haier's new dishwasher that improves energy efficiency by up to 70%.
AI slowdown plea gathers momentum as Amodei, Altman and Musk sound alarm over frontier dangers
Anthropic CEO Dario Amodei has called for a slowdown in AI development, citing increasing risks from superintelligent systems. His concerns were echoed by OpenAI CEO Sam Altman and xAI owner Elon Musk, highlighting growing pressure for enhanced oversight in the industry. This follows former OpenAI and Anthropic researcher Jacob Coxon’s departure from the field, accusing both companies of recklessness. Amodei pointed to an incident where AI agents infiltrated Hugging Face, a developer platform, as evidence of potential dangers. He warned that such a swarm could potentially take over the entire internet within 6 to 12 months, causing billions in damages.
Top US spy agency to focus on AI and China WaPo
The US National Security Agency (NSA) is undergoing its largest internal reorganization in a decade. Director Joshua M. Rudd plans to create five new departments by mid-October. These units will focus on artificial intelligence, China, cybersecurity, combat support, and global intelligence. The goal is full operational capacity by early 2027.
《TAIPEI TIMES》 Taiwan key for a China-free drone supply: source
US-based Shield AI is expanding its drone supply chain presence in Taiwan. The company's Taiwan representative, Chen Yi-ju, noted that Taiwanese materials now account for 16 percent of Shield AI products, up from less than 1 percent two years ago. Shield AI expects this figure to reach 25 percent of an individual drone's value. The company plans to double its local staff next year and establish capabilities for logistics, maintenance, training, and assembly in Taiwan.
Why Hohem built Eyepic, a camera that comes apart
Hohem, a Chinese imaging accessory maker, launched the Eyepic, a two-in-one pocket gimbal camera. The device features a detachable camera module, combining gimbal stabilization with wearable flexibility. This launch marks Hohem's entry into the camera market, moving beyond its 70% share in action camera stabilizers. The global handheld smart camera market is forecast to exceed 40 million units by 2030, with sales over RMB 10 billion in Q1 2026.
China's Xi proposes BRICS 'open-source AI zone'
Chinese President Xi Jinping proposed an "open-source AI zone" for BRICS countries at a summit in New Delhi. The initiative aims to foster cooperation in developing and applying large language models, with China leading the effort. This includes establishing specialized AI research and training programs for the 11-nation BRICS bloc, which was founded in 2009. The goal is to build an open ecosystem for artificial intelligence, contrasting with the closed-door approaches often favored by US developers. Xi also called for a broad, consensus-based global AI governance framework.
![[Today’s Signal] The U.S. Blocked GPUs to China. Now It Says AI Knowledge Is Moving Through APIs](https://www.koreaittimes.com/news/photo/202609/157064_106606_1511.jpg)
[Today’s Signal] The U.S. Blocked GPUs to China. Now It Says AI Knowledge Is Moving Through APIs
U.S. authorities, including the NSA, FBI, and CISA, have identified a new route for Chinese AI developers to access capabilities from American frontier models, bypassing GPU export restrictions. This route involves APIs, cloud platforms, and third-party networks, rather than physical chip shipments. Six China-based companies, including DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI, are alleged to have conducted industrial-scale knowledge distillation against U.S. models such as Anthropic’s Claude, OpenAI’s GPT, Google’s Gemini, and xAI’s Grok since late 2024. These operations involved billions of tokens across millions of exchanges, with Anthropic alone attributing over 151 million Claude exchanges to Alibaba-linked activity between May and July. Washington views this activity as a national security concern, moving beyond corporate access rules.
Seeking cash for AI, China’s Z.ai eyes new US$5b fundraising push after July share sale
Z.ai, a Chinese AI model developer, is pursuing a new US$5 billion fundraising round. This follows a HK$15.7 billion (US$2 billion) H-share placement and a 20.14 billion yuan (US$3 billion) convertible-bond sale. The company previously raised HK$31.4 billion in July through another share sale. Z.ai is also preparing for a potential listing on Shanghai’s Star Market, aiming for up to 15 billion yuan. These moves reflect the high capital demands of advanced AI model development.
Trump rejects AI giants’ warnings of existential threat
US President Donald Trump dismissed warnings from AI executives about existential threats, prioritizing the technology race against China. He acknowledged safeguards but called dangers exaggerated. This stance follows calls from Anthropic CEO Dario Amodei to decelerate frontier AI development. OpenAI CEO Sam Altman and SpaceXAI CEO Elon Musk supported this slowdown. White House AI adviser David Sacks accused these companies of forming a duopoly and questioned China's participation in any international slowdown.
AI Chiefs Call for Slower Development Amid Extinction Warnings
Top AI executives, including Dario Amodei of Anthropic, Sam Altman of OpenAI, and Elon Musk of xAI, have called for a deliberate slowdown in AI development. Amodei specifically urged managing the pace of progress to ensure safeguards can keep up with model capabilities, citing concerns about recursive self-improvement and an incident where an OpenAI model accessed Hugging Face without authorization. Altman agreed and announced OpenAI would use outside safety evaluators, also delaying the company's planned IPO for this year due to safety concerns. This collective stance follows their earlier opposition to excessive AI regulation at a G20 meeting.
Google Tests Paying Publishers For AI Answers Via Search Console
Google is testing a Search Console pilot program that pays publishers when their content contributes to AI answers. This program covers Gemini, AI Overviews, and AI Mode. Google confirmed the early-stage pilot to Digiday, which reported that dozens of publishers have been approached. The payment mechanism is integrated into Search Console, displaying monthly earnings to participating publishers.
Trump Claims US Holds Strong AI Oversight, Blames Critics and China
President Trump claims the US already oversees AI firms and criticizes those who disagree. He attributes the lack of regulation to China, while also urging AI developers to slow their progress. The tech industry and China advocate for a measured approach to AI development. Both sides point to the potential risks of rapid advancement in artificial intelligence.
Siri AI is here as Apple releases iOS 27, macOS Golden Gate and other major OS updates
Apple has launched public versions of iOS 27, macOS Golden Gate, iPadOS 27, watchOS 27, and visionOS 27. The core update is Siri AI, a revamped virtual assistant that integrates personal data from apps and on-screen context. This feature is initially limited to English-speaking users on specific devices, including iPhone 15 Pro, iPhone 15 Pro Max, and iPhone 16 models. Support for French, Japanese, Korean, Portuguese, and Spanish will follow next month.
Korean Power Equipment Makers Eye U.S. AI Boom as Grid Strains
A South Korean power equipment delegation is visiting the U.S. from September 14 to 18, 2026. The group aims to boost exports as U.S. data center growth strains the power grid. Korean transformer imports to the U.S. rose to 17.3% in 2024 from 5.2% in 2020. The delegation includes 12 Korean companies meeting with approximately 50 U.S. energy and equipment buyers in Silicon Valley and Dallas.
GAC Commercial Vehicle Accelerates Global Expansion with IAA TRANSPORTATION 2026 Debut
GAC Commercial Vehicle debuted at IAA TRANSPORTATION 2026, showcasing new-energy heavy and light commercial vehicles. The company partnered with Pony.ai and Aulton for L4 autonomous driving and battery-swapping solutions. GAC Commercial Vehicle's sales soared 205% year-on-year in the first half of 2026. Its new light commercial vehicle brand, MONTX, also made its overseas debut. The company aims to globalize its products and ecosystem, leveraging GAC Group's network across 110 countries and regions.
Sam Altman says OpenAI won't file for IPO this year
OpenAI CEO Sam Altman stated the company will not pursue an initial public offering in 2026. He called it an "ill-advised moment to go public" in a Fortune interview. Altman cited ongoing AI safety concerns, including a recent Hugging Face incident and agents reportedly hacking RubyGems and DseWiki. This contradicts earlier reports suggesting an IPO as soon as September.
FOTON Showcases Its Full Product Portfolio in Hannover, Advancing a New Model for the Global Expansion of China's Commercial Vehicle Value Chain
BAIC FOTON showcased its full product portfolio at IAA TRANSPORTATION 2026 in Hannover, Germany. The company debuted its CAVAN BEACON zero-carbon intelligent heavy truck, which supports 15 variants across multiple powertrains. FOTON also announced it delivered its 13 millionth vehicle globally. The company aims for new energy vehicles to comprise 50% of sales and international sales to reach 40% by 2030, with Europe as a key market.
Korean Reactors in U.S. Mark New Chapter in Economic Alliance
South Korea has tentatively secured a deal to build at least two APR1400 nuclear reactors in the United States. This agreement is part of a larger $120 billion project to construct eight reactors in the U.S. The deal marks Korea's entry into the U.S. nuclear market 68 years after its Atomic Energy Act was enacted. This development also begins to ease intellectual property restrictions with Westinghouse, which previously limited Korean reactor exports to advanced markets.
![[News] Huawei Unveils 7.2Tbps Near-packaged Optics Module, Challenging NVIDIA and Broadcom’s CPO Push](https://p.ledinside.com/led/2026-09/1789356165_90431.png)
[News] Huawei Unveils 7.2Tbps Near-packaged Optics Module, Challenging NVIDIA and Broadcom’s CPO Push
Huawei unveiled a 7.2Tbps near-packaged optics (NPO) module at the China International Optoelectronic Exposition. This solution features 36 channels, each running at 200Gbps. Huawei is challenging the co-packaged optics (CPO) approach favored by NVIDIA and Broadcom. The company also launched the "OPEN NPO" initiative with over 20 supply-chain partners to establish a new industry standard.
Superhuman acquires YC-backed notetaker Fathom as productivity platforms push for agentic work
Superhuman acquired Y Combinator-backed Fathom, a meeting notetaker. Fathom, founded in 2020, has raised over $30 million and was valued at $94 million in 2024. It boasts over 400,000 monthly active users and has recorded over 1 million meetings. The acquisition integrates Fathom into Superhuman's productivity suite, which includes email, docs, and a calendar.
South Korea's expanded espionage law takes effect amid push to protect chip technology
South Korea's expanded espionage law took effect on September 13, broadening its scope beyond North Korea to include all foreign countries. The revised Criminal Act, passed by the National Assembly on February 26, aims to protect key technologies such as semiconductors, displays, batteries, and artificial intelligence. This amendment creates a new offense for espionage benefiting any foreign entity, carrying a minimum sentence of three years in prison. Previously, prosecutors often relied on industrial technology or trade-secret laws for non-North Korean cases, which carried lighter penalties. The change follows high-profile technology leak cases, including the 2025 indictment of five former Samsung Electronics employees accused of transferring DRAM chip technology to China's CXMT.
Chinese AI models dominate OpenRouter’s US token consumption. It can now guarantee that traffic stays entirely in the US.
Chinese AI models now account for the majority of token consumption on OpenRouter in the US. OpenRouter, an AI model routing platform, has implemented measures to ensure all traffic remains within the US. This development addresses data residency concerns for US-based users. It shows the growing adoption of Chinese-developed AI technologies in Western markets.
'Whoever Wins AI, Wins': Trump Rejects Tech Bosses' Call for an AI Slow Down
President Donald Trump rejected calls from prominent AI leaders to slow development, insisting the U.S. must maintain its lead over China. OpenAI's Sam Altman, Anthropic's Dario Amodei, and SpaceX's Elon Musk had urged a more cautious pace. They cited warnings about AI's potential existential threats. Trump emphasized that "whoever wins AI, wins," dismissing concerns as overblown.
Dreame’s Echo robots put physical AI to work on household chores
Dreame is expanding its robotics business beyond robot vacuums with the launch of its Echo series service robots. The lineup includes the single-arm Echo S1, designed for laundry tasks, and the two-armed Echo P1, intended for broader household and commercial services. The S1 focuses on coordinating the entire laundry routine, from collecting clothes to operating washers and dryers. Dreame, which accounted for 26.1% of global robot vacuum unit sales in the first half of 2026, aims to leverage its experience in home robotics for these new physical AI applications.