Chinese President Xi Jinping said 2025 marked a year of major breakthroughs for the country’s AI and semiconductor industries. In his New Year’s address, he said that Chinese technology firms had made significant progress in AI models and domestic chip development.
China’s AI sector gained global attention with the rise of DeepSeek. The company launched advanced models focused on reasoning and efficiency, drawing comparisons with leading US systems and triggering volatility in global technology markets.
Other Chinese firms also expanded their AI capabilities. Alibaba released new frontier models and pledged large-scale investment in cloud and AI infrastructure, while Huawei announced new computing technologies and AI chips to challenge dominant suppliers.
China’s progress prompted mixed international responses. Some European governments restricted the use of Chinese AI models over data security concerns, while US companies continued engaging with Chinese-linked AI firms through acquisitions and partnerships.
Looking ahead to 2026, China is expected to prioritise AI and semiconductors in its next five-year development plan. Analysts anticipate increased research funding, expanded infrastructure, and stronger support for emerging technology industries.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
Nvidia is in discussions with Taiwan Semiconductor Manufacturing Co. (TSMC) about expanding production of its H200 AI chips, following large requests from Chinese technology companies for 2026 deliveries, according to people familiar with the talks.
Those sources stated that Chinese firms have placed orders for more than 2 million H200 chips for 2026, far exceeding Nvidia’s current stock of roughly 700,000 units. Work on the additional output is expected to start in the second quarter of 2026, though the extra volume Nvidia wants has not been publicly detailed.
The H200 is viewed by Chinese buyers as a significant step up from the previously available H20 chips, which were restricted, helping to explain the rush to secure supply. Sources said Nvidia has indicated pricing around $27,000 per chip, while an eight-chip module could cost about 1.5 million yuan, and some prospective buyers see that premium as worthwhile given the performance boost.
The order talks also sit under a cloud of policy uncertainty. While the Trump administration recently allowed H200 exports to China under a framework that includes a 25% fee, Chinese authorities have not yet approved imports, and officials are weighing how such sales could affect the country’s push to build up domestic AI chip suppliers, potentially including rules that tie purchases to local alternatives.
Nvidia stated that it continually manages its supply chain and argued that licensed sales to approved Chinese customers would not impact its ability to serve US clients, while TSMC declined to comment. Separately, a report cited by other coverage stated that ByteDance is considering spending approximately 100 billion yuan on Nvidia chips in 2026, contingent upon the success of H200 product sales in China.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
President Donald Trump has ordered the unwinding of a small US semiconductor-related deal, citing national security concerns tied to China. In an executive order dated 2 January 2026, Trump said there was ‘credible evidence’ that HieFo Corporation is controlled by a citizen of the People’s Republic of China, and that the acquisition ‘might take action that threatens to impair the national security of the United States.’
The order bars HieFo from owning any interest in the purchased Emcore assets and gives the company 180 days to divest them (unless the US foreign-investment review committee grants an extension). It also imposes restrictions meant to prevent access to sensitive, non-public technical information while the divestment is pending.
The transaction at the centre of the order was completed on 30 April 2024, and involved Emcore’s digital chips business and indium-phosphide wafer fabrication operations. Emcore previously stated that the sale price was approximately $2.92 million, while other reports described it as a roughly $2.9 million deal that also included the assumption of $1 million in liabilities.
Officials did not publicly specify the risk, but the executive order follows a review by the Committee on Foreign Investment in the United States (CFIUS), the interagency body that screens certain foreign deals for national security implications. HieFo and Emcore had not publicly responded at the time of publication, and the move is likely to reinforce the message that Washington is prepared to intervene, even after a deal has closed, when advanced manufacturing and China-linked ownership intersect.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
South Korea recorded its highest-ever export figures in 2025, driven largely by surging global demand for semiconductors used in AI technologies. Official data shows total exports exceeded $700 billion, marking a year-on-year increase despite ongoing trade pressures and economic uncertainty.
The semiconductor sector led the growth, with exports reaching a record $173.4 billion, up more than 20 per cent from the previous year. Strong demand for high-value memory chips used in AI data centres pushed shipments higher throughout the year, including a sharp rise in December that capped ten consecutive months of growth.
South Korea’s dominance in the chip market is underpinned by global leaders such as Samsung Electronics and SK Hynix, both key suppliers to the AI industry. The government is also doubling down on the sector, with President Lee Jae Myung pledging to triple national spending on AI in a bid to position the country among the world’s top AI powers.
Other export sectors also posted strong results. Car exports climbed to a record $72 billion, while agriculture and cosmetics benefited from sustained global interest in South Korean food, beauty products, and pop culture. These gains helped offset weaker shipments to the United States and China.
Exports to those two major partners declined amid tariffs on steel, automobiles, and machinery, although Seoul secured a reduced US tariff rate late in the year. While officials hailed the export record as a sign of economic resilience, they cautioned that global trade uncertainty and the durability of semiconductor demand could pose challenges ahead.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
More than 108,000 users of ManageMyHealth may have had their information exposed following a data breach affecting one of the country’s largest patient portals. The incident occurred on Wednesday and is believed to have affected between 6% and 7% of the platform’s 1.8 million registered users.
ManageMyHealth said affected users will be contacted within 48 hours with details about whether and how their data was accessed. Chief executive Vino Ramayah said the company takes the protection of health information extremely seriously and acknowledged the stress such incidents can cause.
He confirmed that the Office of the Privacy Commissioner has been notified and is working with the company to meet legal obligations.
Health Minister Simeon Brown described the breach as concerning but stated that there was no evidence to suggest that Health New Zealand systems, including My Health Account, had been compromised. He added that health services were continuing to operate as normal and that there had been no clinical impact on patient care.
Health New Zealand said it is coordinating with the National Cyber Security Centre and other agencies to understand the scope of the breach and ensure appropriate safeguards are in place.
Officials stressed expectations around security standards, transparency and clear communication, while ongoing engagement with primary care providers and GPs continues.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
AI is rapidly becoming the starting point for many everyday activities, from planning and learning to shopping and decision-making. A new report by PYMNTS Intelligence suggests that AI is no longer just an added digital tool, but is increasingly replacing traditional entry points such as search engines and mobile apps.
The study shows that AI use in the United States has moved firmly into the mainstream, with more than 60 per cent of consumers using dedicated AI platforms over the past year. Younger users and frequent AI users are leading the shift, increasingly turning to AI first rather than using it to support existing online habits.
Researchers found that how people use AI matters as much as how often they use it. Heavy users rely on AI across many aspects of daily life, treating it as a general-purpose system, while lighter users remain cautious and limit AI to lower-risk tasks. Trust plays a decisive role, especially when it comes to sensitive areas such as finances and banking.
The report also points to changing patterns in online discovery. Consumers who use standalone AI platforms are more likely to abandon older methods entirely, while those encountering AI through search engines tend to blend it with familiar tools. That difference suggests that the design and context of AI services strongly influence user behaviour.
Looking ahead, the findings hint at how AI could reshape digital commerce. Many consumers say they would prefer to connect digital wallets directly to AI platforms for payments, signalling a potential shift in how intent turns into transactions. As AI becomes a common entry point to the digital world, businesses and financial institutions face growing pressure to adapt their systems to this new starting line.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Google has filed a lawsuit against a Chinese-speaking cybercriminal network it says is behind a large share of scam text messages targeting people in the United States. The company says the legal action is aimed at disrupting the group’s online infrastructure rather than seeking damages.
According to the complaint, the group, known as Darcula, develops and sells phishing software that allows scammers to send mass text messages posing as trusted organisations such as postal services, government agencies, or online platforms. The tools are designed to be easy to use, enabling people with little technical expertise to run large-scale scams.
Google says the software has been used by hundreds of scam operators to direct victims to fake websites where credit card details are stolen. The company estimates that hundreds of thousands of payment cards have been compromised globally, with tens of thousands linked to victims in the United States.
The lawsuit asks a US court to grant Google the authority to seize and shut down websites connected to the operation, a tactic technology companies increasingly use when criminal networks operate in countries beyond the reach of US law enforcement. Investigations by journalists and cybersecurity researchers suggest the group operates largely in Chinese and has links to individuals based in China and other countries.
The case highlights the growing scale of text-based fraud in the US, where cybercrime losses continue to rise sharply. Google says it will continue combining legal action with technical measures to limit the reach of large scam networks and protect users from increasingly sophisticated phishing campaigns.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
A significant shift away from global views on central bank digital currencies has been made with the decision to allow China’s digital yuan to earn interest starting in January 2026. Wallet balances will now accrue interest at demand deposit rates, marking a shift from the widely held view that retail CBDCs should function purely as digital cash.
Central banks in Europe and the United States have long argued against interest-bearing CBDCs, warning they could destabilise financial systems by drawing deposits away from commercial banks.
Institutions such as the European Central Bank, the Federal Reserve and the Bank for International Settlements have stressed that digital currencies should not become savings instruments.
China’s move, however, effectively repositions the digital yuan closer to a deposit-like form of money rather than a simple cash substitute.
The policy applies to verified individual and corporate wallets, while anonymous wallets remain excluded. Digital yuan balances are also now covered by China’s deposit insurance scheme, offering the same protection as bank deposits.
Analysts say these design choices, combined with China’s two-tier distribution model that keeps commercial banks as intermediaries, aim to limit risks of bank disintermediation while encouraging wider adoption.
China’s decision could influence global debates as dozens of countries continue to explore the use of digital currencies. While Europe remains committed to a non-interest-bearing digital € and the United States has formally banned a retail CBDC, China is testing whether an interest-paying digital currency can coexist with traditional banking.
The experiment is likely to be closely watched as policymakers reconsider what role digital money should play in future financial systems.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Face-to-face interviews and oral verification could become a routine part of third-level assessments under new recommendations aimed at addressing the improper use of AI. Institutions are being encouraged to redesign assessment methods to ensure student work is authentic.
The proposals are set out in new guidelines published by the Higher Education Authority (HEA) of Ireland, which regulates universities and other third-level institutions. The report argues that assessment systems must evolve to reflect the growing use of generative AI in education.
While encouraging institutions to embrace AI’s potential, the report stresses the need to ensure students are demonstrating genuine learning. Academics have raised concerns that AI-generated assignments are increasingly difficult to distinguish from original student work.
To address this, the report recommends redesigning assessments to prioritise student authorship and human judgement. Suggested measures include oral verification, process-based learning, and, where appropriate, a renewed reliance on written exams conducted without technology.
The authors also caution against relying on AI detection tools, arguing that integrity processes should be based on dialogue and evidence. They call for clearer policies, staff and student training, and safeguards around data use and equitable access to AI tools.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
xAI is expanding its AI infrastructure in the southern United States after acquiring another data centre site near Memphis. The move significantly increases planned computing capacity and supports ambitions for large-scale AI training.
The expansion centres on the purchase of a third facility near Memphis, disclosed by Elon Musk in a post on X. The acquisition brings xAI’s total planned power capacity close to 2 gigawatts, placing the project among the most energy-intensive AI data centre developments currently underway.
xAI has bought a third building called MACROHARDRR. Will take @xAI training compute to almost 2GW.
xAI has already completed one major US facility in the area, known as Colossus, while a second site, Colossus 2, remains under construction. The newly acquired building, called MACROHARDRR, is located in Southaven and directly adjoins the Colossus 2 site, as previously reported.
By clustering facilities across neighbouring locations, xAI is creating a contiguous computing campus. The approach enables shared power, cooling, and high-speed data infrastructure for large-scale AI workloads.
The Memphis expansion underscores the rising computational demands of frontier AI models. By owning and controlling its infrastructure, xAI aims to secure long-term access to high-end compute as competition intensifies among firms investing heavily in AI data centres.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
