China

Amazon closes AI research lab in Shanghai as global focus shifts

Amazon is shutting down its AI research lab in Shanghai, marking another step in its gradual withdrawal from China. The move comes amid continuing US–China trade tensions and a broader trend of American tech companies reassessing their presence in the country.

The company said the decision was part of a global streamlining effort rather than a response to AI concerns.

A spokesperson for AWS said the company had reviewed its organisational priorities and decided to cut some roles across certain teams. The exact number of job losses has not been confirmed.

Before Amazon’s confirmation, one of the lab’s senior researchers noted on WeChat that the Shanghai site was the final overseas AWS AI research lab and attributed its closure to shifts in US–China strategy.

The team had built a successful open-source graph neural network framework known as DGL, which reportedly brought in nearly $1 billion in revenue for Amazon’s e-commerce arm.

Amazon has been reducing its footprint in China for several years. It closed its domestic online marketplace in 2019, halted Kindle sales in 2022, and recently laid off AWS staff in the US.

Other tech giants including IBM and Microsoft have also shut down China-based research units this year, while some Chinese AI firms are now relocating operations abroad instead of remaining in a volatile domestic environment.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Alibaba’s Qwen3 upgrade beats OpenAI and DeepSeek on benchmarks

Alibaba has unveiled a significant update to its flagship open‑source Qwen3 family, spotlighting the Qwen3‑235B‑A22B‑Instruct‑2507‑FP8 model.

However, this revision delivers enhanced capabilities across multiple domains, such as instruction understanding, logical reasoning, text analysis, mathematics, science, coding, and tool integration, and pushes Qwen3 to the top of several key benchmarks.

The upgraded model scored 70.3 on the American Invitational Mathematics Exam in competitive metrics, well ahead of DeepSeek‑V3 (46.6) and OpenAI’s GPT‑4o (26.7).

In MultiPL‑E, which evaluates coding, it achieved 87.9, beating DeepSeek (82.2) and OpenAI (82.7), though Anthropic’s Claude Opus 4 edged ahead with 88.5.

A notable technical advancement is the eightfold increase in context capacity to 256k tokens, allowing it to process longer documents in non‑thinking mode.

The open‑source release on reputable platforms like HuggingFace and ModelScope reinforces Alibaba’s commitment to building a transparent, high‑performance AI ecosystem.

This update intensifies competition in China’s AI landscape, with Alibaba closing the benchmark gap versus Western leaders and rival Chinese startups such as DeepSeek, whose upgraded R1‑0528 has reportedly matched Qwen3 in some reasoning tasks.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Android malware infects millions of devices globally

Millions of Android-based devices have been infected by a new strain of malware called BadBox 2.0, prompting urgent warnings from Google and the FBI. The malicious software can trigger ransomware attacks and collect sensitive user data.

The infected devices are primarily cheap, off-brand products manufactured in China, many of which come preloaded with the malware. Models such as the X88 Pro 10, T95, and QPLOVE Q9 are among those identified as compromised.

Google has launched legal action to shut down the illegal operation, calling BadBox 2.0 the largest botnet linked to internet-connected TVs. The FBI has advised the public to disconnect any suspicious devices and check for unusual network activity.

The malware generates illicit revenue through adware and poses broader cybersecurity threats, including denial-of-service attacks. Consumers are urged to avoid unofficial products and verify devices are Play Protect-certified before use.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Salt Typhoon targets routers in sweeping campaign

Since early 2025, the Chinese-linked hacking group Salt Typhoon has aggressively targeted telecom infrastructure worldwide, compromising routers, switches and edge devices used by clients of major operators such as Comcast, MTN and LG Uplus.

Exploiting known but unpatched vulnerabilities, attackers gained persistent access to these network devices, potentially enabling further intrusions into core telecom systems.

The pattern suggests a strategic shift: the group broadly sweeps telecom infrastructure to establish ready-made access across critical communication channels.

Affected providers emphasised that only client-owned hardware was breached and confirmed no internal networks were compromised, but the campaign raises deeper concerns.

Experts warn that such indiscriminate telecommunications targeting could threaten data security and disrupt essential services, revealing a long-term cyber‑espionage strategy.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Chinese scientists develop new method to turn Moon dust into fuel

Chinese researchers have developed an innovative method to extract water and generate fuel from lunar soil, offering a potentially game-changing solution for future Moon missions.

By harnessing a ‘photothermal strategy’—a technique that converts light into heat—they demonstrated how carbon dioxide exhaled by astronauts could be transformed into oxygen, hydrogen, and carbon monoxide using water extracted directly from Moon dust.

The breakthrough, detailed in the journal Joule, centres on integrating water extraction with gas conversion in a single, energy-efficient step.

Professor Lu Wang from the Chinese University of Hong Kong, Shenzhen, described the discovery as a surprise, especially the soil’s unexpected catalytic potential. The integrated system could dramatically reduce reliance on costly Earth-supplied water, currently at about $83,000 per gallon.

The team used actual lunar samples collected during China’s Chang’E-5 mission to test their method. These samples, retrieved from the Moon’s near side in 2020, helped validate the concept in controlled lab conditions.

However, researchers caution that challenges such as radiation, microgravity, and extreme temperature fluctuations may complicate implementation on the Moon’s surface.

Even so, the results underscore the rapid progress of space programme in China. From relative obscurity just two decades ago, the country now positions itself as a frontrunner in building a sustainable Moon base, with ambitions to establish a permanent outpost by 2035.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

China now the top DDoS target, says Cloudflare

Cloudflare’s latest DDoS threat report reveals that business competitors initiate most known attacks. Of the customers who identified attackers, 63% blamed rivals, 21% pointed to state-linked actors, and 5% admitted self-inflicted disruptions caused by misconfigurations.

The Q2 report shows China as the most targeted country, followed by Brazil and Germany, while Ukraine, Singapore and Indonesia are listed among the top sources of DDoS traffic. Telecommunications, internet services and gaming are the industries most frequently targeted by attackers.

Cloudflare highlighted that the locations identified as sources often reflect the presence of botnets, proxies or VPNs, not the actual location of threat actors. Countries like the Netherlands appear high on the list due to favourable privacy laws and strong network infrastructure.

The company urged broader participation in its threat intelligence feed to help mitigate risks. Over 600 providers currently use Cloudflare’s data to remove abusive accounts and stop the spread of DDoS attacks across the internet.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Huang says Chinese AI science is world class

NVIDIA’s CEO has praised China’s open-source AI work during a high-profile visit to Beijing, spotlighting DeepSeek as an example of world-class science and engineering.

At the International Supply Chain Expo, Huang highlighted the global value of Chinese contributions to AI research.

He told former Alibaba executive Wang Jian that Chinese researchers publish more AI papers than any other nation, calling their output A-plus in both science and engineering. Huang named DeepSeek, Alibaba, Tencent, MiniMax, and Baidu’s Ernie Bot globally significant AI leaders.

The visit coincides with a policy shift by the US government allowing Nvidia to resume H20 chip exports to China. Huang confirmed that export licenses are expected soon, potentially unlocking billions in lost revenue.

Industry watchers see this development as pivotal for Nvidia, which had paused shipments under earlier restrictions. The chipmaker’s renewed access to China’s semiconductor market could reshape global AI infrastructure over the next year.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

US moves to block Chinese tech in undersea Internet cables

The United States is moving to strengthen the security of its digital infrastructure by proposing new regulations on undersea internet cables.

The Federal Communications Commission (FCC) plans to ban the connection of submarine cables to the US if they involve Chinese technology or equipment. It also aims to prohibit companies that are flagged as national security threats, including Huawei and ZTE.

Additionally, the FCC will seek public comments on further measures to protect undersea cable infrastructure, as part of an ongoing review of regulations overseeing the global network of submarine cables that carry 99% of international internet traffic.

These policy moves reflect US concerns over China’s role in internet infrastructure and potential espionage. Since 2020, regulators have blocked multiple cable projects linking the US to Hong Kong due to security risks.

Recent incidents, such as suspected sabotage in the Baltic Sea, Taiwan’s accusations of cable cutting by Chinese vessels, and Houthi-linked damage to cables in the Red Sea, highlight the growing vulnerability of global communications networks.

These steps guard against foreign adversary ownership, cyber threats, and physical sabotage.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Salt Typhoon compromises critical US infrastructure

A US state’s Army National Guard network was thoroughly compromised by the Chinese cyberespionage group Salt Typhoon from March to December 2024. According to a confidential federal memo, hackers extracted highly sensitive information, including administrator credentials, network maps, and interstate communication data, raising alarm over data leaked across all 50 states and four US territories.

Security analysts caution that the breach goes beyond intelligence gathering. With access to National Guard systems, integral to state-level threat response and civilian support, the group is poised to exploit vulnerabilities in critical infrastructure, particularly during crises or conflict.

Salt Typhoon, linked to China’s Ministry of State Security, has a track record of penetrating telecommunications, energy grids, transport systems, and water utilities. Often leveraging known vulnerabilities in Cisco and Palo Alto equipment, the group has exfiltrated over 1,400 network configuration files from more than 70 US critical infrastructure providers.

Federal agencies, including DHS and CISA, are sounding the alarm: this deep infiltration presents a serious national security threat and indicates a strategic shift in cyber warfare. Navigating Sun Typhoon’s persistent access through local and federal networks is now a top priority in defending the critical systems on which communities rely.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Tech giants work to avert an AI‑driven energy crisis

The AI boom is triggering alarms in the energy sector, with data centres expected to consume 3% of the world’s electricity by 2030, double today’s share. This projection has spurred tech firms to pursue aggressive efficiency strategies.

Control systems powered by AI regulate cooling more precisely, while liquid cooling systems replace inefficient air-based approaches. Cutting-edge chips and more innovative software have already yielded energy savings of 20–30% per workflow.

Infrastructure improvements mean data-centre support systems now consume just 10% of the energy that powering the servers does. Despite these advances, the total energy demand continues to rise, prompting investments in technology upgrades and low-carbon power sources.

A global race is underway as US and Chinese companies push for increasingly efficient AI chips. Still, experts warn that overall consumption will grow, so efficiency alone won’t be enough. Broad energy planning and sustainable strategies are critical to avoid a looming power crunch as AI proliferates.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!