Huawei chairman Xu outlined the company’s roadmap for AI computing platforms, revealing plans to launch the Atlas 950 SuperPoD in Q4 2026. The system will use over 8,000 Ascend GPUs across 128 racks, covering 1,000 sq metres, and offer 6.7 times more computing power and 15 times more memory.
A year later, the Atlas 960 SuperPod will debut with up to 15,488 Ascend 960 chips, achieving 30 exaflops of computing power and 4,460TB of memory. Xu said the two systems will stay the world’s most potent super nodes, with uses beyond AI in general-purpose computing in China.
Huawei faces Western sanctions limiting access to advanced semiconductor nodes. Xu said assembling less advanced chips into super pods lets Huawei compete with rivals like Nvidia at a system level despite lower individual chip performance.
Over the next three years, Huawei will launch three new Ascend chip series: the 950 line, 950PR and 950DT, the 960, and the 970. The 950PR, optimised for early-stage inference and recommendations, will ship in Q1 2026, while the 950DT with 2Tb/s bandwidth will launch in Q4 2026.
The 960 will double its predecessor’s computing power and memory capacity and arrive in Q4 2027.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Scientists have unveiled an AI tool capable of predicting the risk of developing over 1,000 medical conditions. Published in Nature, the model can forecast certain cancers, heart attacks, and other diseases more than a decade in advance.
Developed by the German Cancer Research Centre (DKFZ), the European Molecular Biology Laboratory (EMBL), and the University of Copenhagen, the model utilises anonymised health data from the UK and Denmark. It tracks the order and timing of medical events to spot patterns that lead to serious illness.
Researchers said the tool is exceptionally accurate for diseases with consistent progression, including some cancers, diabetes, heart attacks, and septicaemia. Its predictions work like a weather forecast, indicating higher risk rather than certainty.
The model is less reliable for unpredictable conditions such as mental health disorders, infectious diseases, or pregnancy complications. It is more accurate for near-term forecasts than for those decades ahead.
Though not yet ready for clinical use, the system could help doctors identify high-risk patients earlier and enable more personalised, preventive healthcare strategies. Researchers say more work is needed to ensure the tool works for diverse populations.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
Google CEO Sundar Pichai announced that its Gemini 2.5 Deep Think AI model achieved a ‘gold-medal performance’ at the 2025 ICPC World Finals. The ICPC World Finals is among the most prestigious university-level programming contests.
The model solved ten out of twelve problems, including one no human team could complete during the competition. The AI solved eight issues in the first 45 minutes and completed two more in the following three hours, demonstrating exceptional algorithmic reasoning and coding skills.
Gemini’s participation was conducted live and remotely under official ICPC rules, starting ten minutes after human teams, with the same five-hour limit.
Incredible milestone: an advanced version of Gemini 2.5 Deep Think achieved gold-medal performance at the ICPC World Finals, a top global programming competition, solving an impressive 10/12 problems. Such a profound leap in abstract problem-solving – congrats to @googledeepmind!
Google explained that Gemini’s performance results from reinforcement learning, allowing the AI to reason, generate code, test solutions, and refine its approach based on feedback. Internal tests indicate Gemini 2.5 would have matched the top 20 human coders at past ICPC finals.
For general users, a lighter version of Gemini 2.5 Deep Think is available through the Gemini app for Google AI Ultra subscribers, offering access to advanced problem-solving capabilities in a more accessible format.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Ireland has designated 15 authorities to monitor compliance with the EU’s AI Act, making it one of the first EU countries fully ready to enforce the new rules. The AI Act regulates AI systems according to their risk to society and began phasing in last year.
Governments had until 2 August to notify the European Commission of their appointed market surveillance authorities. In Ireland, these include the Central Bank, Coimisiún na Meán, the Data Protection Commission, the Competition and Consumer Protection Commission, and the Health and Safety Authority.
The country will also establish a National AI Office as the central coordinator for AI Act enforcement and liaise with EU institutions. A single point of contact must be designated where multiple authorities are involved to ensure clear communication.
Ireland joins Cyprus, Latvia, Lithuania, Luxembourg, Slovenia, and Spain as countries that have appointed their contact points. The Commission has not yet published the complete list of authorities notified by member states.
Former Italian Prime Minister Mario Draghi has called for a pause in the rollout of the AI Act, citing risks and a lack of technical standards. The Commission has launched a consultation as part of its digital simplification package, which will be implemented in December.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
NVIDIA and the UK are accelerating plans to build the nation’s AI infrastructure, positioning the country as a hub for AI innovation, jobs and research.
The partnership, announced by Prime Minister Keir Starmer and NVIDIA CEO Jensen Huang earlier in the year, has already resulted in commitments worth up to £11 billion.
A rollout that includes AI factories equipped with 120,000 NVIDIA Blackwell GPUs across UK data centres, supporting projects such as OpenAI’s Stargate UK.
NVIDIA partner Nscale will host 60,000 of these GPUs domestically while expanding its global capacity to 300,000. Microsoft, CoreWeave and other partners are also investing in advanced supercomputing facilities, with new projects announced in England and Scotland.
NVIDIA is working with Oxford Quantum Circuits and other research institutions to integrate AI and quantum technologies in a collaboration that extends to quantum computing.
Universities in Edinburgh and Oxford are advancing GPU-driven quantum error correction and AI-controlled quantum hardware, highlighting the UK’s growing role in cutting-edge science.
To prepare the workforce, NVIDIA has joined forces with techUK and QA to provide training programmes and AI skills development.
The government has framed the initiative as a foundation for economic resilience, job creation and sovereign AI capability, aiming to place Britain at the forefront of the AI industrial revolution.
Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Zuckerberg’s Meta has unveiled a new generation of smart glasses powered by AI at its annual Meta Connect conference in California. Working with Ray-Ban and Oakley, the company introduced devices including the Meta Ray-Ban Display and the Oakley Meta Vanguard.
These glasses are designed to bring the Meta AI assistant into daily use instead of being confined to phones or computers.
The Ray-Ban Display comes with a colour lens screen for video calls and messaging and a 12-megapixel camera, and will sell for $799. It can be paired with a neural wristband that enables tasks through hand gestures.
Meta also presented $499 Oakley Vanguard glasses aimed at sports fans and launched a second generation of its Ray-Ban Meta glasses at $379. Around two million smart glasses have been sold since Meta entered the market in 2023.
Analysts see the glasses as a more practical way of introducing AI to everyday life than the firm’s costly Metaverse project. Yet many caution that Meta must prove the benefits outweigh the price.
Chief executive Mark Zuckerberg described the technology as a scientific breakthrough. He said it forms part of Meta’s vast AI investment programme, which includes massive data centres and research into artificial superintelligence.
The launch came as activists protested outside Meta’s New York headquarters, accusing the company of neglecting children’s safety. Former safety researchers also told the US Senate that Meta ignored evidence of harm caused by its VR products, claims the company has strongly denied.
Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Hammersmith and Fulham Council has approved a £3m upgrade to its CCTV system to see facial recognition and AI integrated across the west London borough.
With over 2,000 cameras, the council intends to install live facial recognition technology at crime hotspots and link it with police databases for real-time identification.
Alongside the new cameras, 500 units will be equipped with AI tools to speed up video analysis, track vehicles, and provide retrospective searches. The plans also include the possible use of drones, pending approval from the Civil Aviation Authority.
Council leader Stephen Cowan said the technology will provide more substantial evidence in a criminal justice system he described as broken, arguing it will help secure convictions instead of leaving cases unresolved.
Civil liberties group Big Brother Watch condemned the project as mass surveillance without safeguards, warning of constant identity checks and retrospective monitoring of residents’ movements.
Some locals also voiced concern, saying the cameras address crime after it happens instead of preventing it. Others welcomed the move, believing it would deter offenders and reassure those who feel unsafe on the streets.
The Metropolitan Police currently operates one pilot site in Croydon, with findings expected later in the year, and the council says its rollout depends on continued police cooperation.
Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!
AI is reshaping healthcare, according to experts featured in Microsoft’s new podcast seriesThe AI Revolution in Medicine, Revisited.
Peter Lee, President of Microsoft Research, spoke with clinicians and researchers about how AI is helping doctors work more effectively and patients access better care. From note-taking during visits to accelerating drug discovery, the technology is already proving its value.
A doctor of UC San Diego Health in the US said AI helps physicians draft longer, more empathetic responses to patient messages, reducing mental strain. Meanwhile, Stanford’s Dr Roxana Daneshjou described how AI detected a dosage error in a medical summary, acting as a crucial safeguard.
Bill Gates highlighted how AI could boost healthcare in low-income regions by providing medical intelligence where doctors are scarce. Other guests suggested the technology may even blur traditional boundaries between medical specialties while accelerating drug development.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
The US tech giant, Google, has introduced a new experimental app for Windows that combines web search, file discovery and Google Lens in a single interface.
The tool, known as the Google app for Windows, is part of Search Labs and is designed to allow users to find information faster instead of interrupting their workflow.
An app that can be launched instantly using the Alt+Space shortcut, opening a Spotlight-like bar similar to Apple’s macOS. Users can search local files, installed applications, Google Drive content and web results. It supports multiple modes, including AI-generated answers, images, videos, shopping and news.
A dark mode is available for those who prefer night-time use, and the search bar can be resized or repositioned on the desktop instead of staying fixed.
Google has also built its Lens technology, allowing users to select and search images directly on screen, translate text or solve mathematical problems. An AI Mode offers detailed replies, though it can be disabled or customised through the settings menu.
The experimental app is currently limited to English-speaking users in the US and requires Windows 10 or Windows 11. Google has not yet confirmed when it will expand availability to more regions.
Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Scaling law estimation helps organisations make better decisions about architecture, optimisers and dataset sizes before devoting extensive compute budgets.
The team assembled over 485 pre-trained models across 40 families (including Pythia, OPT, Bloom, LLaMA and others) and tracked almost 1.9 million performance metrics. Using that dataset, they fit more than 1,000 scaling laws and assessed variables such as the number of parameters, the token count, intermediate training checkpoints, and seed effects.
Practical recommendations include discarding training data from very early stages (before about 10 billion tokens), using several small models across sizes rather than only large ones, and using intermediate checkpoints rather than waiting for final model loss.
The guide also notes that a 4 percent absolute relative error (ARE) is near best-case for prediction quality, though up to 20 percent ARE remains useful depending on budget.
Because training large models can cost millions, these scaling laws also help those without huge resources to approximate outcomes more safely. AI model inference scaling laws are still under development and are flagged as important future work.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
