Weekly #226 – Google fined $35M in Australia, EU–US seal $750B trade deal, Nvidia’s new China AI chip

 Logo, Text

15 – 22 August 2025


 License Plate, Transportation, Vehicle, Adult, Male, Man, Person, Car, Book, Comics, Publication, Face, Head, Animal, Antelope, Mammal, Wildlife, Canine, Dog, Pet, Police Dog

Dear readers,

Last week, on Monday, Google agreed to pay a A$55 million (US$35.8 million) fine in Australia after regulators found it restricted competition by striking revenue-sharing deals with Telstra and Optus to pre-install its search app on Android phones, sidelining rival platforms. The Australian Competition and Consumer Commission (ACCC) said the arrangements, which were in place from 2019 to 2021, limited consumer choice and blocked competitors’ visibility. Google admitted that the deals harmed competition and pledged to drop similar practices, while Telstra and Optus confirmed that they no longer pursue such agreements. The settlement, which still requires court approval, comes amid wider legal and regulatory challenges for Google in Australia, including a recent loss in a case brought by Epic Games and growing scrutiny over its role in app distribution and social media access.

The United States and the European Union have agreed on a new Framework Agreement on Reciprocal, Fair, and Balanced Trade, aiming to reset one of the world’s largest trade relationships. The deal includes EU commitments to eliminate tariffs on US industrial goods, expand access for American agricultural and seafood products, and procure $750 billion in US energy exports and $40 billion in AI chips by 2028. In return, the US will cap tariffs on key EU goods, ease automobile tariffs, and pursue cooperation on steel, aluminium, and supply chain security. Both sides pledged deeper collaboration on defence procurement, digital trade, cybersecurity, sustainability rules, and standards harmonisation, while also working to resolve disputes over deforestation, carbon border taxes, and non-tariff barriers.

OpenAI CEO Sam Altman warned that the US risks underestimating China’s rapid AI progress, arguing that export controls on advanced semiconductors are an unreliable long-term solution. Speaking in San Francisco, he said chip restrictions and policy-driven approaches often fail due to workarounds, while China is quickly expanding its AI capacity and accelerating domestic alternatives through firms like Huawei.

On the same front, Nvidia is quietly developing a new AI chip for China, the B30A, based on its advanced Blackwell architecture, just as Washington debates how much US technology Beijing should be allowed to access. Positioned between the weaker H20 and the flagship B300, the B30A retains key features like high-bandwidth memory and NVLink, making it more powerful than China’s current scaled-down H20 approved model while staying within export limits. The move follows President Trump’s recent openness to allowing scaled-down chip sales to China, though bipartisan lawmakers remain wary of boosting Beijing’s AI capabilities. Nvidia, which relies on China for 13% of its revenue, also plans to release the lower-end RTX6000D for AI inference in September, reflecting efforts to comply with US-China export policy while fending off rising domestic rivals like Huawei, whose chips are improving but still lag in software and memory. Meanwhile, Chinese regulators have warned firms about potential security risks in Nvidia’s products, underscoring the political tensions shaping the company’s commercial strategy.

Private conversations with xAI’s chatbot Grok were unintentionally exposed online after its ‘share’ button generated public URLs that became indexed by Google and other search engines, raising serious concerns about user privacy and AI safety. The leaked chats included sensitive and dangerous content, from hacking crypto wallets to drug-making instructions, despite xAI’s ban on harmful use. The flaw, reminiscent of earlier issues with other AI platforms like ChatGPT, has damaged trust in xAI and highlighted the urgent need for stronger privacy safeguards, such as blocking the indexing of shared content and adopting privacy-by-design measures, as users may otherwise hesitate to engage with chatbots.

Meta is launching a new research lab dedicated to superintelligence, led by Scale AI founder Alexandr Wang, as part of its push to regain momentum in the global AI race after mixed results with its Llama models and ongoing talent losses. Mark Zuckerberg is reportedly considering a multibillion-dollar investment in Scale, signalling strong confidence in Wang’s approach, while the lab’s creation, separate from  Meta’s FAIR division, underscores Meta’s shift toward partnerships with top AI players, mirroring strategies used by Microsoft, Amazon, and Google.

Japanese technology giant SoftBank has announced plans to buy a $2 billion stake in Intel, signalling a stronger push into the American semiconductor industry. The investment comes as Washington debates greater government involvement in the sector, with reports suggesting President Donald Trump is weighing a US government stake in the chipmaker. SoftBank will purchase Intel’s common stock at $23 per share. Its chairman, Masayoshi Son, said semiconductors remain the backbone of every industry and expressed confidence that advanced chip manufacturing will expand in the US, with Intel playing a central role.

For the main updates, reflections and events, consult the RADAR, the READING CORNER and the UPCOMING EVENTS section below.

Join us as we connect the dots, from daily updates to main weekly developments, to bring you a clear, engaging monthly snapshot of worldwide digital trends.

DW Team


RADAR

Highlights from the week of 15 – 22 August 2025

closeup shot pile cryptocurrency cash golden coins

The Frontier Stable Token marks the first government-backed stablecoin in the US, with Wyoming positioning itself as a leader in digital finance innovation.

download

With the V3.1 upgrade now live and the R1 label missing, observers are debating whether DeepSeek has postponed or abandoned its R2 reasoning model entirely.

collaborating with ai agents improves productivity

The former Twitter chief executive argues that AI agents will soon dominate the internet instead of humans, with individuals likely to deploy dozens to manage daily online activity.

3098222

Strangeworks acquires German firm Quantagonia to expand European operations and bring AI-powered, quantum-ready planning technology to more organisations.

rm373batch13 083

Government plan tackles talent shortages and chip supply disruption.

people wearing futuristic high tech virtual reality glasses

Nearly nine in ten developers are using AI agents to speed up coding, testing, and localisation, while also adapting games to players in real time.

children playing grass

Regulators urge safeguards for AI toys as children gain interactive companions that teach and engage instead of relying solely on human interaction or screens.

rubaitul azad u4F54GIZWGI unsplash

Malware steals passwords, crypto data and system info without clicks.

Alibaba logo screenshot youtube publiultra footage

The platform allows users to conduct market research, plan products, design prototypes, check regulations, and find distributors in minutes rather than weeks.

representation user experience interface design 3

Competitors Grok, Claude, and Copilot trail far behind as ChatGPT leads in downloads and consumer spending per user.

108177567 1753687050071 Alibaba AI glasses

HTC’s entry into this market is significant as it competes with established players like Meta, Google, and Samsung, each developing or already offering advanced smart glasses technology.

desktop photo laptop with ai hologram interface chat concept ai artificial intelligence

Key concerns include the potential for widespread job displacement as AI systems replace human workers, significant environmental harm due to the substantial energy usage of AI models, and privacy erosion…

handcuffs on laptop cyber crime concept

A $45 million Bitcoin donation accepted without checks has turned into a major Czech political scandal, now focused on money laundering and drug trafficking.

INTEL

Trump criticises Intel leadership, urging board action.


READING CORNER
BLOG featured image 2025 95

As the Trump-Putin summit brought Alaska into focus, its overlooked telegraph cables reveal a fascinating history: in the late 19th century, Alaska was on the brink of becoming a telecommunication hub connecting the US to Europe via Siberia.

BLOG featured image 2025 97

Can AI replace the transmission of wisdom? The world of education is changing radically and rapidly. Generative AI tools are now capable of writing essays, solving math problems, summarising textbooks, and even personalising learning experiences at scale.

 
BLOG featured image 2025 96

English dominates the AI landscape, but this hegemony can hinder our understanding of AI’s deeper, non-technical aspects. The recent explosion of AI jargon often obscures meaning and can lead to cognitive confusion. Embracing our native languages allows us to deflate this jargon, fostering clearer, common-sense comprehension of AI concepts.

AI Hub AI and Law Ethics

AI offers tools to expand access to justice globally, but without transparency, oversight, and human-rights safeguards, it risks deepening bias, exclusion, and eroding public trust.

Investment diplomacy v2

How is money shaping foreign policy? Learn how countries use sovereign wealth funds and strategic investments as powerful tools for foreign policy and soft power.

GPT-5 criticised for lacking flair as users seek older ChatGPT options

OpenAI’s rollout of GPT-5 has faced criticism from users attached to older models, who say the new version lacks the character of its predecessors.

GPT-5 was designed as an all-in-one model, featuring a lightweight version for rapid responses and a reasoning version for complex tasks. A routing system determines which option to use, although users can manually select from several alternatives.

Modes include Auto, Fast, Thinking, Thinking mini, and Pro, with the last available to Pro subscribers for $200 monthly. Standard paid users can still access GPT-4o, GPT-4.1, 4o-mini, and even 3o through additional settings.

Chief executive Sam Altman has said the long-term goal is to give users more control over ChatGPT’s personality, making customisation a solution to concerns about style. He promised ample notice before permanently retiring older models.

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

Musk acknowledges value in ChatGPT-5’s modesty after public spat

Elon Musk has taken an unexpected conciliatory turn in his feud with Sam Altman by praising a ChatGPT-5 response, ‘I don’t know’, as more valuable than overconfident answers. Musk described it as ‘a great answer’ from the AI chatbot.

Initially sparked by Musk accusing Apple of favouring ChatGPT in App Store rankings and Altman firing back with claims of manipulation on X, the feud has taken on new dimensions as AI itself seems to weigh in.

At one point, xAI’s Grok chat assistant sided with Altman, while ChatGPT offered a supportive nod to Musk. These chatbot alignments have introduced confusion and irony into a clash already rich with irony.

Musk’s praise of a modest AI response contrasts sharply with the often intense claims of supremacy. It signals a rare acknowledgement of restraint and clarity, even from an avowed critic of OpenAI.

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

OpenAI softens GPT-5 tone after users complain it felt cold

OpenAI has updated GPT-5 to make its tone noticeably warmer and more engaging, without reverting to the overly flattering style some users criticised in GPT-4o. The change is rolling out, aiming to balance emotional resonance with substance.

CEO Sam Altman stated the adjustment directly responds to users finding GPT-5 too formal or robotic. The update is subtle yet visible, enhancing conversational warmth while avoiding sycophantic tendencies.

OpenAI also expands user control by offering three interaction modes, Auto, Fast, and Thinking, which adapt response style to user preference. These changes empower users to shape the tone and depth of their AI interactions.

Reacting to public frustration, OpenAI has reinstated GPT-4o (along with GPT-4.1, o3, and GPT-5 Thinking mini) for paid subscribers, while promising more customisation options in future updates.

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

ChatGPT dominates mobile AI market

ChatGPT’s mobile app has surpassed $2 billion in worldwide consumer spending since its launch in May 2023, according to Appfigures. Revenue from January to July 2025 alone reached $1.35 billion, a 673% increase from the same period in 2024.

The app has also dominated downloads, with an estimated 690 million lifetime installs, including 318 million added in 2025. India leads in total downloads at 13.7%, followed by the US, which accounts for 38% of revenue.

Competitors such as Grok, Claude, and Copilot remain far behind, with Grok generating just $25.6 million in 2025.

Consumer spending per download reinforces ChatGPT’s lead, averaging $2.91 globally and $10 in the US. The figures highlight OpenAI’s dominance in the mobile AI assistant market, despite ongoing criticism from X owner Elon Musk, who has alleged that the App Store suppresses competition.

Apple has rejected these claims.

The AI market continues to heat up as Microsoft integrates OpenAI’s GPT-5 into its Copilot offerings. Elon Musk has predicted intense competition, while OpenAI CEO Sam Altman has downplayed Musk’s criticism, emphasising innovation and collaboration as the sector expands.

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

Sam Altman admits OpenAI holds back stronger AI models

OpenAI recently unveiled GPT-5, a significant upgrade praised for its advances in accuracy, reasoning, writing, coding and multimodal capabilities. The model has also been designed to reduce hallucinations and excessive agreeableness.

Chief executive Sam Altman has admitted that OpenAI has even more powerful systems that cannot be released due to limited capacity.

Altman explained that the company must make difficult choices, as existing infrastructure cannot yet support the more advanced models. To address the issue, OpenAI plans to invest in new data centres, with spending potentially reaching trillions of dollars.

The shortage of computing power has already affected operations, including a cutback in image generation earlier in the year, following the viral Studio Ghibli-style trend.

Despite criticism of GPT-5 for offering shorter responses and lacking emotional depth, ChatGPT has grown significantly.

Altman said the platform is now the fifth most visited website worldwide and is on track to overtake Instagram and Facebook. However, he acknowledged that competing with Google will be far harder.

OpenAI intends to expand beyond ChatGPT with new standalone applications, potentially including an AI-driven social media service.

The company also backs Merge Labs, a brain-computer interface rival to Elon Musk’s Neuralink. It has partnered with former Apple designer Jony Ive to create a new AI device.

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

GPT-5 impresses in reasoning but stumbles in flawless coding

OpenAI’s newly released GPT-5 draws praise and criticism in equal measure, as developers explore its potential for transforming software engineering.

Launched on 7 August 2025, the model has impressed with its ability to reason through complex problems and assist in long-term project planning. Yet, engineers testing it in practice note that while it can propose elegant solutions, its generated code often contains subtle errors, demanding close human oversight.

Benchmark results showcase GPT-5’s strength. The model scored 74.9% on the SWE-bench Verified test, outperforming predecessors in bug detection and analysis. Integrated into tools such as GitHub Copilot, it has already boosted productivity for large-scale refactoring projects, with some testers praising its conversational guidance.

Despite these gains, developers report mixed outcomes: successful brainstorming and planning, but inconsistent when producing flawless, runnable code.

The rollout also includes GPT-5 Mini, a faster version for everyday use in platforms like Visual Studio Code. Early users highlight its speed but point out that effective prompting remains essential, as the model’s re-architected interaction style differs from GPT-4.

Critics argue it still trails rivals such as Anthropic’s Claude 4 Sonnet in error-free generation, even as it shows marked improvements in scientific and analytical coding tasks.

Experts suggest GPT-5 will redefine developer roles rather than replace them, shifting focus toward oversight and validation. By acting as a partner in ideation and review, the model may reduce repetitive coding tasks while elevating strategic engineering work.

For now, OpenAI’s most advanced system sets a high bar for intelligent assistance but remains a tool that depends on skilled humans to achieve reliable outcomes.

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

Employees of OpenAI eye multi-billion dollar stock sale

According to a source familiar with the talks, OpenAI employees are preparing to sell around $6 billion worth of shares to major investors, including SoftBank Group and Thrive Capital.

The deal, still at an early stage, would push the company’s valuation to $500 billion, up from its current $300 billion.

SoftBank, Thrive and Dragoneer Investment Group are already among OpenAI’s backers, and their participation in the secondary share sale would further strengthen ties with the Microsoft-supported AI company.

Reports suggest the size of the sale could still change as discussions continue.

The planned deal follows SoftBank’s leadership role in OpenAI’s $40 billion primary funding round earlier this year. Employee share sales often reflect strong investor demand and highlight the rapid growth of companies in competitive markets.

OpenAI has seen user numbers and revenues soar in 2025, with weekly active ChatGPT users climbing to about 700 million, up from 400 million in February.

The company doubled its revenue in the first seven months of the year, hitting an annualised run rate of $12 billion, and is expected to reach $20 billion by the end of the year.

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

New OpenAI hire shares savvy interview strategies

Bas van Opheusden, who joined OpenAI as a technical staff member in July, has published a comprehensive eight-page guide for aspiring applicants, offering strategic advice spanning recruiter calls, coding interviews, compensation discussions and more.

He suggests treating recruiter conversations as strategic briefings, which are key for understanding the hiring manager’s priorities, team dynamics, role expectations, and organisational goals.

Van Opheusden recommends taking notes during calls, ideally using a dual-screen setup, and arranging windows so it appears you’re maintaining eye contact.

He also shared a standard error: arriving at coding interviews without remembering the exact role he’d applied for, underscoring the importance of clear preparation and role alignment.

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

OpenAI’s GPT-5 faces backlash for dull tone

OpenAI’s GPT-5 launched last week to immense anticipation, with CEO Sam Altman likening it to the iPhone’s Retina display moment. Marketing promised state-of-the-art performance across multiple domains, but early user reactions suggested a more incremental step than a revolution.

Many expected transformative leaps, yet improvements mainly were in cost, speed, and reliability. GPT-5’s switch system, which automatically routes queries to the most suitable model, was new, but its writing style drew criticism for being robotic and less nuanced.

Social media buzzed with memes mocking its mistakes, from miscounting letters in ‘blueberry’ to inventing US states. OpenAI quickly reinstated GPT-4 for users who missed its warmer tone, underlining a disconnect between expectations and delivery.

Expert reviews mirrored public sentiment. Gary Marcus called GPT-5 ‘overhyped and underwhelming’, while others saw modest benchmark gains. Coding was the standout, with the model topping leaderboards and producing functional, if simple, applications.

OpenAI emphasised GPT-5’s practical utility and reduced hallucinations, aiming for steadiness over spectacle. At the same time, it may not wow casual users, its coding abilities, enterprise appeal, and affordability position it to generate revenue in the fiercely competitive AI market.

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