Scientists prevail in Johnson & Johnson talc case

A US federal judge has rejected Johnson & Johnson’s trade libel case against three scientists whose research linked cosmetic talc exposure to mesothelioma, ruling that the company failed to show the researchers acted with actual malice.

US District Judge Jamar K. Walker granted summary judgment to doctors Theresa Emory, John Maddox and Richard Kradin in the Eastern District of Virginia, ending the remaining claim in litigation brought by Johnson & Johnson subsidiary Pecos River Talc LLC. The company has said it plans to appeal and seek a trial over what it describes as false statements concerning talc and asbestos exposure.

The dispute centred on a scientific paper published in March 2020 in the American Journal of Industrial Medicine. The researchers presented a case series involving 75 people with malignant mesothelioma whose only known exposure to asbestos, the paper said, was through cosmetic talc. The group was described as additional to 33 patients covered by an earlier study.

Johnson & Johnson challenged the accuracy of those representations. Evidence developed during the litigation showed that eight of the 75 patients had also appeared in the earlier study, contradicting the description of the group as entirely additional. The company also argued that some patients had possible asbestos exposure from sources other than cosmetic talc.

Walker found that mistakes had occurred in the underlying data but concluded that the record did not establish the legal standard required for trade libel. Pecos River had to produce clear and convincing evidence that the scientists published statements they knew were false or acted with reckless disregard for their truth.

The judge determined that the company had not created a genuine dispute of material fact on actual malice. The researchers had compared anonymised patient data and information supplied through lawyers, circumstances that could make duplicate cases difficult to identify. The court found no evidence showing that errors apparent during later litigation had been obvious to the researchers when the study was prepared.

The decision is significant beyond the immediate talc dispute because it addresses the boundary between corporate litigation and scientific publication. The researchers’ lawyers argued that allowing defamation-style claims to proceed without evidence of deliberate or reckless falsehood could have a chilling effect on researchers studying products that are already the subject of major litigation.

Johnson & Johnson takes the opposite view. It has maintained that the scientific literature used against the company in talc lawsuits contains serious flaws and that expert witnesses working with plaintiffs’ lawyers have helped create misleading claims about its products. The company has repeatedly said its cosmetic talc was safe, asbestos-free and did not cause cancer.

The Virginia case formed part of a broader legal strategy challenging experts involved in talc litigation. Johnson & Johnson also pursued separate litigation against Jacqueline Moline, whose earlier research examined mesothelioma cases associated with cosmetic talc. A federal judge in New Jersey dismissed that case in 2024, finding that the challenged research did not support the company’s allegations of fraud or libel.

The latest ruling does not establish that cosmetic talc causes mesothelioma, nor does it resolve the wider scientific debate over asbestos contamination and cancer risk. Its immediate legal finding is narrower: the company did not present sufficient evidence that Emory, Maddox and Kradin published their disputed statements with the level of knowledge or recklessness required to sustain the trade libel claim.

Mesothelioma is an aggressive cancer affecting the lining surrounding organs, most commonly the lungs, and asbestos exposure is its principal established cause. Litigation involving cosmetic talc has focused heavily on allegations that some talc deposits or finished products contained asbestos fibres. Johnson & Johnson has consistently disputed claims that its products contained asbestos.

The ruling arrives as the company seeks to reduce its exposure to more than a decade of talc litigation. Johnson & Johnson announced in July a proposed settlement committing at least $5.5 billion to resolve about 76,000 ovarian cancer claims. The agreement requires participation by firms representing at least 95 per cent of the remaining claims before it can take effect.

The settlement follows three unsuccessful attempts to resolve large numbers of talc claims through bankruptcy proceedings involving subsidiaries. A bankruptcy judge rejected the latest restructuring plan in 2025, after which Johnson & Johnson returned to defending cases individually while pursuing challenges to plaintiffs’ expert testimony.

AI skills reshape cybersecurity hiring across G7

Artificial intelligence skills are rapidly becoming a standard requirement for cybersecurity professionals, with their presence in job advertisements across the G7 economies doubling within a year as employers redesign security teams around AI agents and human oversight.

AI capabilities were specified in 28.5% of cybersecurity job postings between October 2025 and March 2026, compared with 14.2% in the corresponding period a year earlier. By March, the proportion had climbed to 29.7%, highlighting how quickly knowledge of AI-powered systems is moving from a specialist qualification towards a mainstream cybersecurity requirement.

The shift comes as overall demand for cybersecurity workers continues to expand. Job postings across Canada, France, Germany, Italy, Japan, the United Kingdom and the United States grew 9.5% during the six months ending March 2026. The growth accelerated from 6.8% in the preceding six-month period.

The findings form part of new research by the AI Workforce Consortium, established by Cisco in 2024 with technology and employment groups including Accenture, Cornerstone, Eightfold AI, Google, IBM, Indeed, Intel, Microsoft and SAP. Pearson has since joined the consortium.

The changing requirements go beyond learning how to operate generative AI applications. Security engineering, cloud security, detection and response engineering and security operations roles increasingly demand an “agentic” combination of skills that allows professionals to direct, supervise and verify autonomous AI systems.

Security operations centres illustrate the transformation. AI agents can now perform high-volume tasks including alert triage, threat-intelligence correlation and execution of established response procedures. Human analysts are consequently moving away from processing every alert themselves towards overseeing automated decisions, investigating ambiguous threats and determining when machine-generated conclusions require intervention.

This evolution is making some cybersecurity positions more strategic rather than simply eliminating human work. Demand for ethical reasoning skills in cybersecurity postings increased 533% year on year, while systems-thinking requirements rose 251% and stakeholder-engagement skills increased 125%.

Employers are increasingly looking for professionals who can understand the technical output of AI systems while judging its reliability, implications and potential risks. Such abilities are becoming particularly important as autonomous agents gain access to security tools capable of taking actions rather than merely producing recommendations.

The transition is also exposing an imbalance between senior and entry-level recruitment. Cybersecurity postings carrying senior titles increased 65% during the six months ending March 2026, while junior-titled positions grew only 5.9%.

Employers appear to be raising expectations for people entering the profession even as the global shortage of cybersecurity workers remains severe. The worldwide cybersecurity skills gap exceeds 4.8 million positions, leaving organisations competing for experienced specialists while entry-level candidates struggle to demonstrate the broader capabilities employers increasingly demand.

A May survey of cybersecurity leaders showed the difficulty particularly clearly. Forty-nine per cent identified practical experience with AI agents as one of the hardest competencies to find among entry-level applicants. Technical cybersecurity depth was cited by 48%, while 45% identified deficiencies in human and professional skills.

The mismatch is placing pressure on universities, certification providers and corporate training programmes to move beyond conventional classroom instruction. Training increasingly needs to expose students to AI-enabled security environments where they can evaluate agent-generated findings, challenge automated recommendations and understand when autonomous actions should be stopped or escalated.

Internships, apprenticeships and laboratory-based learning are consequently becoming more important as employers seek evidence that applicants can operate alongside AI systems before joining production security teams.

The transformation builds on a broader acceleration in AI-related technology employment. Earlier labour-market analysis across G7 economies found AI capabilities appearing across 78% of the technology roles examined, while demand for AI security skills had risen sharply. Skills involving responsible AI, foundation-model adaptation and multi-agent systems were also among the fastest-growing specialist requirements.

Investment patterns, however, indicate a possible workforce weakness. More than a third of cybersecurity leaders plan to invest in AI-powered security capabilities over the next one to two years, while only about a quarter place comparable priority on investing in workforce skills.

TDCX expands Hyderabad base as India workforce surges

TDCX is expanding its operations in Hyderabad with a new campus at Meenakshi Eco Park, increasing capacity as its India workforce grows by more than 600% over three years.

The expansion strengthens the Singapore-headquartered customer experience specialist’s presence in one of the country’s largest technology hubs and underlines India’s growing importance in supporting global technology, digital economy and high-growth clients.

The Meenakshi Eco Park facility gives TDCX additional space to scale customer experience operations, technology functions and specialised services as demand from international clients becomes increasingly complex. The company entered the market in 2022 with a 45,000-square-foot office at Skyview 20 in Hyderabad, establishing the city as both a service delivery location and a digital innovation hub.

Since then, TDCX has expanded rapidly as companies seek a combination of skilled labour, multilingual capabilities and technology expertise. Its headcount in the country has risen by more than sixfold over three years, making the Hyderabad operation an increasingly significant part of its global delivery network.

The expansion comes as customer experience outsourcing is being reshaped by artificial intelligence, automation and growing demand for higher-value services. Large technology companies are increasingly looking beyond traditional contact-centre functions, seeking partners capable of handling digital advertising, e-commerce, trust and safety, fintech, mobility and technically demanding customer interactions.

TDCX has positioned its operations around that shift. Its services combine customer support with revenue operations, content and community management, digital transformation, AI-enabled analytics and consultancy. The company says its model is designed to combine automation with human judgement, particularly where customer queries or platform decisions require more complex handling.

Hyderabad provides access to a large pool of technology and engineering professionals as well as employees experienced in business process management. The city has become an important global capability centre destination, attracting major investments from technology, financial services and digital businesses seeking skilled workers and comparatively competitive operating costs.

Meenakshi Eco Park, in Hyderabad’s western technology corridor, is emerging as a major commercial development. The project has a planned built-up area of about six million square feet and has attracted large corporate occupiers. Uber has leased more than 900,000 square feet at the development under a long-term agreement, while investment management company Invesco has also taken substantial office space there.

TDCX’s move reflects a broader transformation in the outsourcing sector, where companies are placing greater emphasis on specialised skills rather than simply increasing the number of customer service agents. Generative AI can now resolve routine inquiries, summarise conversations and provide real-time assistance to employees, encouraging service providers to direct more workers towards complex customer interactions, technical support and revenue-generating functions.

The company is also using AI-supported tools in its customer operations. Such systems can assist agents by retrieving information, suggesting responses and analysing customer sentiment while automation handles repetitive processes. Human employees remain central to interactions requiring judgement, empathy or an understanding of regulatory and cultural factors.

TDCX operates across Asia-Pacific, Europe and the Americas and employs more than 20,000 people across its global network. Its operations cover more than 50 languages and dialects, allowing multinational clients to consolidate customer experience programmes across different markets.

India is becoming an increasingly important part of that network because of its combination of English-language proficiency, engineering talent and established outsourcing infrastructure. While traditional business-process work remains substantial, employers are increasingly recruiting workers with skills in data analytics, software development, cybersecurity, artificial intelligence and digital operations.

That development is also altering competition for talent in Hyderabad. Global capability centres, cloud computing companies, financial institutions and technology groups are expanding in the city, increasing demand for employees who combine technical proficiency with communication and customer-management skills.

TDCX initially identified Hyderabad as its entry point because of the city’s technology ecosystem and availability of specialists in areas including frontend development, user interface and user experience design, Angular, React and full-stack engineering. Its technology teams have worked on software intended to generate deeper insights into customer behaviour and improve service outcomes.

The new campus provides additional capacity for the next phase of that strategy. TDCX is expected to use its expanded Hyderabad footprint to support international businesses seeking round-the-clock digital customer operations while developing capabilities around AI-assisted services and other technology-intensive customer experience functions.

ByteDance and Tencent secure first H200 deliveries

ByteDance and Tencent have each received about 10,000 Nvidia H200 artificial intelligence processors as Beijing permits limited shipments into mainland China while continuing to steer the bulk of approved hardware towards offshore deployment.

The deliveries mark the first substantial flow of H200 processors to the two technology groups after months of uncertainty over whether approvals from Washington would translate into physical shipments. Other major technology companies are expected to seek similar clearance as demand for high-performance computing capacity intensifies across China’s artificial intelligence industry.

US authorities have cleared ByteDance and Tencent to purchase as many as 100,000 H200 processors each. Beijing, however, is restricting the number that can be deployed on the mainland as part of a policy designed to balance the immediate computing requirements of leading AI developers against its longer-term objective of building a self-sufficient semiconductor industry.

Chinese regulators have indicated that companies may place larger quantities of the processors in Hong Kong, which lies outside the mainland customs boundary. Individual shipments into the mainland remain subject to tighter regulatory scrutiny, effectively creating a two-tier system for access to Nvidia hardware.

The arrangement gives companies including ByteDance and Tencent access to powerful processors for training large artificial intelligence models without opening the mainland market fully to imported chips. It also protects demand for processors developed by domestic suppliers led by Huawei Technologies and a growing group of semiconductor designers backed by major technology companies.

Nvidia's H200 is based on its Hopper architecture and carries 141GB of HBM3e memory with memory bandwidth of 4.8 terabytes per second. The processor was designed for generative AI, large language models and high-performance computing workloads, although it now sits behind Nvidia's newer Blackwell generation and the company's more advanced computing systems.

Access to even the older H200 remains important for Chinese AI developers because large-scale model training requires enormous clusters of processors with high memory capacity and mature software support. Domestic alternatives have improved rapidly, but Nvidia's CUDA software ecosystem and established data-centre architecture continue to give its hardware advantages for some sophisticated training workloads.

Washington opened a pathway for H200 sales to China in December 2025 after years of progressively tighter restrictions on advanced semiconductor exports. The Commerce Department followed in January by introducing case-by-case licensing for H200 processors and comparable products, subject to security, compliance and supply requirements.

The policy stopped short of allowing China access to Nvidia's most capable systems. More advanced Blackwell and subsequent architectures remain restricted, preserving a performance gap between the processors available to customers in China and those being deployed by leading US technology companies.

The opening nevertheless prompted debate in Washington over whether even H200 exports could accelerate China's artificial intelligence development. Supporters of controlled sales have argued that allowing older US processors to remain commercially relevant in China could prevent domestic competitors from displacing Nvidia entirely. Critics contend that powerful accelerators can still strengthen capabilities considered strategically sensitive.

Beijing faces a parallel policy dilemma. Its technology companies need greater computing power to remain competitive with US AI developers, but unrestricted purchases of Nvidia processors could weaken the commercial prospects of domestic semiconductor manufacturers that have received extensive policy support.

Huawei has emerged as the principal domestic alternative, while Alibaba and other technology groups are also expanding proprietary processor programmes. The competitive landscape is shifting towards systems that combine locally designed chips, networking equipment and software rather than relying entirely on imported accelerators.

ByteDance requires substantial computing resources to develop and operate its artificial intelligence models and services, while Tencent is increasing investment in AI infrastructure alongside its cloud, gaming and advertising businesses. Tencent's capital spending has risen sharply as it expands data-centre capacity and model development.

Infrastructure constraints could complicate the strategy of routing large volumes of H200 equipment through Hong Kong. High-density AI servers demand significant electricity, cooling and networking capacity, and available data-centre infrastructure in the territory is considerably smaller than the mainland facilities operated by China's technology groups.

That limitation means Hong Kong can function as an approved destination for processors but may not immediately accommodate every accelerator authorised for purchase. Companies could consequently continue pressing Beijing for permission to move selected systems into mainland data centres where large AI clusters are already operating.

India pulls ahead in enterprise AI adoption

India has emerged as the Asia-Pacific leader in enterprise artificial intelligence investment and adoption, with businesses accelerating spending while pushing AI deeper into everyday operations.

Autodesk’s 2026 State of Design & Make: AI Pulse shows 91% of organisations surveyed in India increased their AI investment during the past year, compared with 85% across Asia-Pacific and 81% globally. The findings point to a widening shift from experimental projects towards operational deployment, particularly in industries where design, engineering, construction, manufacturing and digital content depend on complex data and workflows.

AI assistants and chatbots are already used by 63% of respondents in India, the highest proportion in Asia-Pacific, while every organisation covered by the India survey reported using at least one form of AI. The figures place the country ahead of several mature technology markets in the region in the breadth of enterprise adoption.

Investment is increasingly moving towards technologies that can perform defined tasks rather than simply generate text or images. Generative AI remains the largest investment priority, cited by 74% of respondents, followed by process automation at 58% and AI-powered decision-support systems at 54%. About 69% expect to introduce agentic AI within the next 12 months, signalling growing corporate interest in systems capable of carrying out multi-step activities with limited human intervention.

Businesses are also reporting measurable gains. About 88% of respondents in India said AI had improved productivity, 84% reported stronger innovation and 75% cited improvements in decision-making. Globally, Autodesk found that 98% of leaders across the design and make industries use at least one AI tool and 84% associate AI with productivity gains.

The findings reinforce evidence that businesses across the country are moving beyond isolated proof-of-concept programmes. Separate enterprise research published this year found 40% of organisations reporting significant or full AI usage, compared with about 28% globally, while large-scale deployment was particularly strong in product development, strategy and operations, marketing and sales, and supply chains.

Workforce preparedness has become another differentiating factor. Autodesk’s survey found 70% of respondents in India considered the AI training available to employees adequate, while 88% reported leadership support for AI adoption. That combination of management backing and workforce development is helping companies shift AI projects from technology teams into business functions.

The picture is not uniformly positive. Regulatory uncertainty was identified as a challenge by 50% of respondents, while 49% pointed to shortages of specialised skills and 47% cited difficulties integrating AI with existing systems. Other research on enterprise AI deployment has similarly identified data quality, systems integration, talent availability and uncertain returns as obstacles to scaling applications beyond individual departments.

Confidence in data security nevertheless remains high among the Autodesk respondents, with 93% saying their AI systems could securely handle business data. That level of confidence is important as companies introduce models into engineering, construction, manufacturing and media workflows containing commercially sensitive intellectual property and proprietary design information.

Market expansion is also becoming linked with AI strategy. About 87% of respondents in India said their organisations were prioritising expansion into new markets, suggesting companies increasingly view AI not only as an efficiency tool but as infrastructure for growth, product development and faster decision-making.

Autodesk compiled the AI Pulse study with Statista Plus Research from a survey conducted between January and February 2026. It covered 2,500 industry leaders and experts across 16 countries and regions spanning architecture, engineering, construction and operations, design and manufacturing, and media and entertainment. The company classified 19% of global respondents as early adopters that had already incorporated advanced technologies such as large language models or agentic AI into their workflows.

Open-source adoption meets enterprise hurdles in India

Open-source technologies are gaining ground across India’s enterprise sector as companies accelerate artificial intelligence deployments, but ageing systems, cybersecurity concerns, regulatory obligations and shortages of specialised skills are complicating the shift.

Businesses are increasingly evaluating open-source software, AI models and cloud-native platforms as alternatives to proprietary technologies. The attraction is being reinforced by the need to control costs, avoid dependence on individual vendors and customise AI systems around industry-specific requirements. Yet the transition is exposing weaknesses in enterprise technology estates that were built over decades and were not designed for data-intensive AI workloads.

Legacy infrastructure remains one of the biggest obstacles. Banks, telecommunications companies, manufacturers and large service providers often operate critical applications on older platforms that cannot easily communicate with modern open-source frameworks. Moving such workloads can involve extensive application redesign, data migration and testing while companies must continue running essential operations without disruption.

The challenge has become more prominent as enterprises seek to take AI projects beyond experimentation. Nearly three-quarters of organisations in India have encountered difficulties moving AI initiatives beyond the proof-of-concept stage, with fragmented data, outdated infrastructure and processes designed for earlier generations of technology among the principal constraints.

Open source is nevertheless playing a growing role in AI development. About 76% of startups in India use open-source AI, reflecting the lower entry costs and greater flexibility offered by openly available models and development tools. The country’s AI market is projected to expand from about $6 billion in 2024 to almost $32 billion by 2031 as companies, startups and public institutions increase deployment.

Enterprise interest has followed the same direction. A study of technology decision-makers found that 71% of companies surveyed planned to increase their use of open-source solutions, while almost half said more than half of the AI solutions they were already using were based on open-source technologies. Governance, AI expertise and technology integration ranked among the main barriers to expansion.

Security is a particularly sensitive issue for enterprises adopting open-source components. Companies must track vulnerabilities across libraries, frameworks and software dependencies while ensuring that updates do not disrupt production systems. The growing use of generative and agentic AI adds further concerns around access controls, sensitive corporate information, model behaviour and the provenance of training or application data.

Information-security spending in India is projected to reach $3.4 billion during 2026, an increase of 11.7% from the previous year, as companies respond to AI-enabled threats and tougher regulatory requirements. Identity attacks, credential compromise and deepfake-enabled fraud are pushing businesses towards stronger detection, governance and resilience measures.

Compliance is also influencing technology choices. Enterprises handling personal or sensitive information must consider requirements created by the Digital Personal Data Protection framework, sector-specific regulations and data-residency obligations. Multinational companies face an additional layer of complexity because AI applications may process information across jurisdictions with different privacy and governance rules.

These requirements are driving demand for architectures that combine open technologies with enterprise-grade management, security and support. Technology suppliers are consequently positioning hybrid platforms as a way to give companies the flexibility of open ecosystems while retaining controls over models, data and infrastructure.

IBM and Yotta Data Services, for example, announced plans this year for an agentic AI platform hosted on Yotta’s Shakti Cloud, aimed at helping enterprises and government organisations deploy AI while addressing data residency, security and regulatory compliance requirements.

Cloud investment is simultaneously accelerating. Public-cloud spending in India is forecast to reach $17.5 billion in 2026, rising 28.1% from $13.7 billion in 2025. Platform-as-a-service expenditure alone is expected to reach about $6.4 billion as organisations modernise technology foundations for AI and integrate data across applications.

Skills could prove just as important as infrastructure. Enterprises need engineers capable of managing containers, Kubernetes environments, open-source databases, machine-learning frameworks, cybersecurity tools and increasingly complex AI stacks. Demand is also shifting from general software skills towards specialists who understand data engineering, model governance, application modernisation and AI operations.

India possesses a large technology workforce and has recorded strong demand for AI professionals, but enterprise requirements are evolving faster than many organisations can retrain employees. The resulting gap is encouraging companies to expand internal training, recruit specialised engineers and work more closely with technology vendors and open-source communities.

Copper gains ground in India’s electric vehicle boom

Copper is becoming an increasingly strategic input for India’s automotive manufacturing industry as electric vehicles, advanced electronics and charging infrastructure lift demand for the highly conductive metal.

Prime Minister Narendra Modi highlighted the scale of the transformation in his Independence Day address on August 15, saying about 25 lakh electric vehicles were sold in 2025-26, compared with roughly 1.5 lakh in 2009-10. The figures underline how electrification has moved from a niche segment towards a sizeable part of the wider vehicle market.

The shift has consequences well beyond battery production. Electric vehicles use copper in traction motors, high-voltage wiring, inverters, battery-management systems, charging connections and power electronics. Conventional vehicles are also becoming more copper-intensive as manufacturers add advanced driver-assistance systems, sensors, infotainment, connectivity and electronically controlled functions.

Battery-electric cars can require several times as much copper as comparable internal-combustion vehicles, although the precise quantity varies by vehicle size, motor architecture and battery configuration. Industry estimates commonly place copper content in a typical battery-electric passenger vehicle at around 60-80 kg, against roughly 20-25 kg for a conventional model.

India’s expanding electric two-wheeler, three-wheeler, passenger-car and commercial-vehicle markets therefore represent an increasingly important source of metal demand. Charging networks add another layer, requiring copper for cables, transformers, switchgear and grid connections as public and private charging installations expand.

The trend coincides with wider growth in electronics manufacturing. Vehicles increasingly resemble mobile electronic platforms, incorporating semiconductor-based control units, radar and camera systems, digital displays and sophisticated electrical architectures. Even before full electrification, rising electronic content was increasing the amount of copper wiring and connectors required per vehicle.

This development is creating opportunities for manufacturers of copper rods, wires, foils, busbars, connectors and specialised alloys. Suppliers capable of producing high-conductivity components to automotive specifications are becoming more important as carmakers localise components and seek shorter supply chains.

The challenge lies in securing enough material. Domestic copper demand is estimated at about 1.8 million tonnes annually, while refined production remains substantially lower. Imports have expanded sharply over the past decade, reflecting both rising consumption and limitations in domestic mining and processing capacity.

The government’s long-term Copper Vision anticipates demand increasing sixfold by 2047 and envisages about five million tonnes a year of additional smelting and refining capacity by 2030. The strategy also places greater emphasis on recycling, overseas mineral assets and long-term concentrate supply arrangements.

Hindalco Industries remains a major domestic refined-copper producer, while Adani Group’s Kutch Copper complex has added substantial new smelting capacity. Hindustan Copper, the country’s principal domestic copper miner, is seeking to expand production and strengthen access to overseas resources.

Supply security has consequently become intertwined with foreign economic policy. Companies and state enterprises have been exploring copper opportunities in resource-rich countries including Chile and Peru. Hindustan Copper has pursued arrangements involving concentrate sourced from Chile, while major private groups have examined opportunities in overseas mines and long-term raw-material partnerships.

The international backdrop adds urgency. Global copper demand is being pushed higher by electric vehicles, renewable power, electricity grids, data centres and other digital infrastructure. Current projections show copper recording the largest absolute increase in demand among several key energy-transition minerals through 2040.

Mine development, however, typically takes many years. Ore grades are declining at some established deposits, while new projects face high capital costs, environmental approvals and geopolitical risks. Even with announced projects proceeding, global copper supply could remain materially below projected requirements during the next decade.

For vehicle manufacturers, copper availability is only one part of the equation. Prices can affect the cost of motors, wiring harnesses, chargers and power-electronic assemblies. Automakers are responding through lightweight wiring designs, higher-voltage systems and engineering changes that reduce material use without compromising conductivity or safety.

Recycling could also assume a larger role. Copper can be repeatedly recovered without losing its fundamental electrical properties, giving scrap from vehicles, electrical equipment and industrial machinery significant economic value. Building stronger collection and secondary-refining systems could reduce import exposure as vehicle scrappage volumes rise.

SEBI blocks two firms after Sensex auction trades

The Securities and Exchange Board of India has barred Copthall Mauritius Investment Limited and Mansi Share and Stock Broking Private Limited from the securities market after detecting alleged manipulation during the BSE Sensex closing auction, and ordered the impounding of about ₹3.68 crore in alleged wrongful gains.

The action followed trading on August 13, when weekly derivatives contracts linked to the Sensex expired. SEBI issued an ex-parte interim order on August 19, only six days after the disputed transactions, marking a swift regulatory intervention in the Closing Auction Session, or CAS, introduced at the beginning of the month.

SEBI calculated prima facie wrongful gains of ₹2.96 crore for Copthall and ₹71.64 lakh for Mansi, taking the combined amount to ₹3,67,80,773. Copthall is a Mauritius-based investment entity owned by JPMorgan Chase, while the restriction on Mansi applies to its proprietary trading account.

The case centres on unusually aggressive orders placed in Sensex constituent shares during the closing auction. Copthall allegedly entered large buy orders at prices materially above reference levels, pushing the indicative equilibrium prices of several stocks higher. Its derivatives portfolio included long call and short put positions that stood to gain from an upward movement in the benchmark.

Surveillance data identified three abrupt Sensex movements during the auction, involving swings of 362.02 points, 132.67 points and 405.08 points over intervals ranging from two seconds to 28 seconds. During one two-second surge, Copthall accounted for 99.91% of the relevant buy-order value, while its share was 96.09% during another sharp movement.

Many of those orders were placed close to the permitted upper price boundary and across multiple Sensex constituents within seconds. A substantial proportion was later cancelled after influencing the indicative auction price, strengthening SEBI's preliminary view that the trading pattern had been designed to affect price discovery rather than simply execute genuine investment demand.

Mansi allegedly employed an opposite strategy. The brokerage placed aggressive sell orders in eight Sensex constituents at prices significantly below their reference levels, temporarily suppressing the benchmark's indicative equilibrium price. Its outstanding put-option positions benefited from the downward move. Most of the sell orders were subsequently withdrawn after the relevant derivatives positions had been squared off.

SEBI has not alleged at this stage that Copthall and Mansi acted together. Its preliminary examination indicates that the two entities independently used opposing strategies during the same closing auction, each producing price movements favourable to its own derivatives positions. A detailed investigation remains under way.

Pending that investigation, both entities have been restrained from accessing the securities market and from participating directly or indirectly in equity-market closing auctions. They cannot place, modify or cancel CAS orders until further directions. Restrictions have also been imposed on debits from relevant bank and demat accounts, while disposal of assets requires regulatory permission. The entities have been directed to provide inventories of their assets, investments and accounts within 15 days.

The enforcement action is significant because CAS has been operating only since August 3. The system replaced the earlier method of calculating closing prices of futures-and-options stocks from the volume-weighted average of trades during the final 30 minutes of continuous trading. Under the auction method, orders are accumulated near the close and matched at a single equilibrium price intended to reflect maximum executable volume.

The 20-minute mechanism begins at 3.15 pm and was introduced to improve end-of-day price discovery and bring the market closer to practices used by major global exchanges. The official closing price is particularly important because it affects benchmark calculations, derivatives settlements, portfolio valuations and the net asset values of funds.

Its early implementation has nevertheless produced sharp movements in the Nifty 50 and Sensex around the close, exposing liquidity differences between exchanges and creating challenges for derivatives traders. SEBI Chairman Tuhin Kanta Pandey has ruled out abandoning the system while signalling that operational weaknesses will be addressed.

Regulators are also working with brokers to deepen liquidity during the auction process. Brokers are expected from September to begin accepting orders during the five-minute transition period between 3.15 pm and 3.20 pm, when order entry has so far been unavailable. Exchanges have separately improved visibility of indicative benchmark prices and introduced shorter-duration securities-lending contracts to make hedging and short-selling easier.

Banks accelerate offshore bond funding

Banks have raised about $4 billion through international bond markets this year, as cheaper hedging, strong investor demand and expanding foreign-currency business revive offshore borrowing after years of relatively subdued issuance.

HDFC Bank, ICICI Bank, Axis Bank and State Bank of India have been among the principal issuers, while another group of lenders is preparing deals that could keep the fundraising cycle active through the remainder of 2026. The rush has broadened from large banks to lenders making their first appearances, or returning after lengthy absences, in the international debt market.

The revival has been driven largely by a Reserve Bank of India initiative introduced in June that sharply reduced the cost of hedging qualifying overseas borrowings. The facility allows eligible foreign-currency borrowings with minimum maturities of three years to be swapped at a fixed annual rate of 1.5 per cent, reducing one of the biggest disadvantages banks face when raising dollars abroad.

HDFC Bank was among the first major beneficiaries, raising $750 million through five-year senior unsecured bonds issued through its International Financial Services Centre banking unit at GIFT City. The securities carry a coupon of 5.067 per cent and mature in June 2031.

Axis Bank followed with an $800 million dual-tranche transaction. It sold $500 million of perpetual Additional Tier 1 capital notes carrying a 6.875 per cent coupon and $300 million of five-year senior unsecured notes paying 5.348 per cent. The senior securities mature in June 2031.

ICICI Bank has emerged as one of the most active borrowers. After returning to the public dollar market in July with a $1 billion five-year issue — its first such benchmark transaction since 2017 — the lender raised another $750 million in August. The second transaction was priced at 105 basis points above the five-year US Treasury yield after investor demand exceeded $2.2 billion.

State Bank of India has also stepped up overseas fundraising. The lender raised $600 million through a private placement of three-year dollar bonds after postponing an earlier planned public transaction when heavy supply pushed borrowing costs higher. It subsequently returned to the public market with plans for at least $500 million of five-year bonds through its London branch, with bankers expecting the eventual size to potentially reach $1 billion.

The next phase is likely to involve a wider pool of borrowers. Kotak Mahindra Bank, YES Bank, IDFC First Bank and Federal Bank have accelerated plans to raise a combined $1.85 billion through dollar bonds with maturities of up to five years. Kotak is entering the international bond market for the first time, while YES Bank is returning after the restructuring that followed its 2020 crisis.

Their plans have gained urgency after the central bank brought forward the closing date for a separate zero-cost foreign-exchange swap facility covering Foreign Currency Non-Resident deposits to August 31 from September 30. The measure had attracted $52.3 billion in such deposits between June and mid-August, prompting the earlier closure. Other concessional windows covering external commercial borrowings and overseas foreign-currency borrowings remain available until the end of December.

Banks have sought foreign-currency liabilities partly to support lending and deposit businesses linked to overseas customers while diversifying funding beyond domestic deposits. International issuance can also broaden the institutional investor base, particularly among global asset managers, pension funds and banks seeking investment-grade Asian credit.

Pricing remains sensitive to US Treasury yields and the volume of simultaneous supply. SBI had delayed a planned $1 billion deal in June when competing issuance increased borrowing costs, illustrating how an overcrowded calendar can erode some of the advantage provided by cheaper hedging.

That constraint has not halted the broader expansion. Overseas fundraising by banks reached about $6.3 billion after the central bank's June measures, compared with roughly $850 million earlier in the year, while lenders were preparing billions of dollars of additional bond and loan transactions before the end-August deposit-swap deadline.

Amazon email redesign raises phishing concerns

Amazon’s simplified order confirmation emails are drawing scrutiny from cybersecurity specialists who warn that vague product descriptions could make phishing messages harder for shoppers to distinguish from genuine communications.

The redesigned emails replace specific product names and images with broad labels such as “Household item”, “Essentials item”, “Decor” or “Garden”. Customers must open Amazon’s app or website to discover exactly what has been ordered or shipped, a change the company says reduces the amount of personal information transmitted outside its own platforms.

Confusion over the system has grown as the classifications have sometimes appeared unrelated to purchases. Dan Landau, a 38-year-old marketing executive, ordered a portable photo printer but received an Amazon confirmation indicating that his “luggage” was coming. A separate purchase of pool-maintenance products was described under three different categories — “Decor”, “Outdoors” and “Garden”.

“The categories are very vague, and they seem to have little to do with what is actually being ordered,” Landau said.

Amazon customers have reported similar experiences since July, with detailed descriptions and product thumbnails disappearing from order messages and being replaced by generic classifications. Emails sent in June still commonly displayed the precise items purchased.

Amazon says the redesigned messages are intended to steer customers towards the “Your Orders” section of its app and website, where consolidated information about purchases and delivery status is available. The company also says reducing order information contained in email helps improve customer privacy by limiting data exposed outside Amazon-controlled services.

Cybersecurity specialists, however, say the trade-off could create another vulnerability. Genuine order emails traditionally gave shoppers an immediate way to recognise a legitimate purchase without clicking anything. Removing identifying details means customers who want to know what an email refers to may become accustomed to following links.

Arun Vishwanath, founder of the Cyber Hygiene Academy and an adviser to the US National Security Agency’s research-focused Science of Security and Privacy programme, said the approach could increase rather than reduce risk.

“It’s definitely not in the best interest of customers,” Vishwanath said. “If anything, it increases risk.”

Phishing campaigns commonly imitate legitimate corporate emails before directing recipients towards counterfeit websites designed to collect passwords, payment credentials or other personal information. Malicious links can also be used to deliver malware.

The problem is particularly significant for major online retailers because fraudulent order notices have long been among the most common impersonation tactics. Amazon-related scams frequently claim that an unfamiliar purchase has been made, that a delivery requires confirmation or that payment details must be updated.

Americans reported about $3.5 billion in losses from imposter scams during 2025. Amazon has also remained one of the businesses most frequently impersonated by fraudsters. Federal data for 2023 recorded roughly 34,000 reports involving scammers pretending to represent Amazon, placing it behind only Best Buy and its Geek Squad brand.

Amazon has argued that authentic communications remain identifiable and advises customers to verify purchases by opening its app or website directly rather than relying on links in messages. Legitimate communications can also be checked through the Message Centre within an Amazon account.

Security specialists nevertheless contend that changing the appearance and usefulness of legitimate messages can give criminals an opportunity. Erich Kron, a security awareness adviser at KnowBe4, said fraudsters commonly exploit periods when customers are adapting to unfamiliar formats.

“Anytime some change like this happens, bad actors are going to jump on it,” Kron said, adding that unfamiliarity can encourage mistakes.

The risk is greater on smartphones, where checking the actual destination behind a hyperlink is less straightforward than hovering over a link with a mouse on a desktop computer. Customers receiving a vague notice about an unidentified shipment could therefore click through simply to determine what is being delivered.

Another concern is credential reuse. Passwords obtained through counterfeit shopping pages can be tested automatically against banks, credit-card providers and other retailers through credential-stuffing attacks, potentially turning one compromised retail account into a wider financial-security problem.

There is no conclusive evidence so far that Amazon’s email redesign has produced an increase in phishing attacks. Abnormal AI, which develops artificial-intelligence-based email security systems, has not identified data establishing such a rise. Its assessment nevertheless indicates that the format could make imitation campaigns easier to construct.

Amazon has intensified efforts against impersonation fraud, including removing tens of thousands of phishing websites and fraudulent telephone numbers. The company says customers should never provide passwords, one-time authentication codes or payment information in response to unsolicited communications.

Pentagon widens university scrutiny over foreign research ties

The Pentagon has ordered 30 leading US universities to audit foreign research partnerships within two weeks, escalating Washington’s effort to prevent federally funded science and advanced technology from benefiting institutions linked to China’s military and other strategic rivals.

The universities must examine collaborations involving organisations identified by the Pentagon as research-security risks and terminate arrangements considered problematic or face possible loss of federal funding. Harvard University, Massachusetts Institute of Technology and Johns Hopkins University are among the institutions affected, while other schools identified include Georgetown University, Cornell University, Duke University, New York University and several University of California campuses. The reviews are due by August 31.

The directive follows an updated Pentagon assessment identifying about 130 foreign universities and research organisations whose activities are considered capable of increasing the likelihood that US government-funded research and development could be diverted or exploited. Most are based in China, although institutions from Russia and Iran are also included.

The move has gained added significance following a congressional investigation into Harvard’s dealings with China. Two House committees alleged this month that Harvard maintained research relationships with institutions connected to China’s defence research establishment and operated an affiliated non-profit that could have been used to circumvent federal foreign-funding disclosure requirements. Harvard has faced growing scrutiny over research security as well as its wider financial links with overseas donors.

The congressional report identified Harvard Global Research and Support Services, commonly called Harvard Global, as an affiliated organisation capable of accepting certain awards when Harvard University itself could not accept a sponsor’s terms for legal or financial reasons. Archived material cited by lawmakers said Harvard Global could become the formal applicant and award recipient while Harvard researchers carried out the work using university resources.

Lawmakers characterised the arrangement as a potential mechanism for bypassing Section 117 of the Higher Education Act, which requires universities receiving federal financial assistance to disclose qualifying foreign gifts and contracts. Harvard Global began operations around 2012, and its annual revenue grew from about $621,000 in its 2012 tax filing to more than $7 million in 2023, the congressional investigation found.

The investigation does not establish that the roughly $630 million Harvard has disclosed as having come from Chinese sources was routed through Harvard Global. The larger figure refers to Harvard’s overall reported funding from Chinese sources over decades, while the congressional allegation concerns whether the affiliated organisation could have provided a structure through which some transactions escaped normal disclosure rules. Federal records place Harvard’s total reported foreign funding at about $4.5 billion.

That distinction is important because congressional investigators have alleged a disclosure loophole rather than demonstrated that hundreds of millions of dollars were secretly channelled through the non-profit. The House report said Harvard Global’s earlier website language was removed after committees sent the university a draft of their findings. It also noted that federal guidance treats intermediaries operating for the benefit of universities as potentially subject to foreign-funding disclosure requirements.

Harvard’s foreign-funding compliance has previously been examined by federal authorities. An Education Department review opened in 2020 was closed in December 2024 after Harvard attested that its disclosures were current and that systems were in place to comply with Section 117. A new review was opened in April 2025 after authorities identified what they described as incomplete or inaccurate disclosures. That review remained open when the House committees issued their report.

Research-security concerns extend well beyond Harvard. A separate congressional examination published in 2025 identified about 1,400 scientific papers produced between June 2023 and June 2025 through collaborations involving Pentagon-funded US researchers and Chinese partners. The work was associated with roughly 700 defence grants worth more than $2.5 billion, with research covering artificial intelligence, semiconductors, hypersonics, advanced materials and propulsion technologies. More than half of the publications involved organisations affiliated with China’s defence research or industrial system.

University representatives have cautioned that Washington risks treating legitimate scientific cooperation as evidence of wrongdoing. The American Council on Education has argued that universities have worked with federal authorities for years to strengthen safeguards and that institutions should not automatically be penalised for partnerships established before particular foreign organisations were placed on security lists.

Google shifts Pixel hardware production beyond China

Google is preparing to move production of its entire Pixel hardware portfolio out of China by 2027, accelerating a supply-chain restructuring that will make Vietnam and India its principal manufacturing bases for smartphones, smartwatches and wireless earbuds.

The Alphabet-owned company has informed suppliers of the planned transition after successfully expanding production of premium Pixel phones in Vietnam and building greater manufacturing capacity in India. The strategy would end China's role as a final manufacturing centre for Pixel-branded devices, although Chinese suppliers are expected to remain embedded in parts of the wider component network.

Vietnam has emerged as the cornerstone of Google's high-end Pixel manufacturing programme. The company moved beyond straightforward assembly there during 2026 by assigning suppliers responsibility for new-product introduction work on flagship Pixel models, including engineering, production-process development, verification and manufacturing preparation. That shift represented an important test of whether Google's supply chain outside China could handle the technically demanding stages required before mass production.

Successful development and manufacturing of high-end Pixel phones in Vietnam has strengthened confidence that the 2027 target is achievable. Vietnam already hosts an extensive electronics manufacturing ecosystem and has become an alternative base for technology groups seeking to spread production risk across Asia.

India is simultaneously taking on a larger role, particularly as Google develops export capacity alongside production for the domestic market. Pixel smartphones are already assembled there, with contract manufacturer Dixon Technologies among the companies participating in Google's supply chain. Production has expanded from serving local customers towards supplying overseas markets, placing the country in a stronger position within Google's global hardware operation.

The redistribution gives Google several manufacturing options rather than shifting dependence from one country to another. Vietnam is positioned heavily around flagship development and assembly, while India offers a large electronics manufacturing base, growing component capacity and access to one of the world's biggest smartphone markets.

Google's manufacturing expansion has coincided with a broader commercial push in India. The company began selling Pixel phones, watches and earbuds directly through its own online store in 2025 after previously relying primarily on authorised retailers and Flipkart. It has also explored a physical retail presence while widening local manufacturing.

The strategy comes as technology companies seek to reduce risks created by trade tensions between Washington and Beijing, tariff uncertainty and concentrated production networks. For Google, the task is less complicated than for Apple because Pixel volumes are considerably smaller and its manufacturing footprint is less deeply entrenched in China.

Google's smartphone business, while still well behind Samsung and Apple globally, has been gaining ground in important premium markets. Pixel shipments rose 14 per cent year on year during the first quarter of 2026 even as worldwide smartphone shipments fell, helped by Google's emphasis on artificial intelligence, computational photography and its integrated Android software experience.

Momentum has been particularly visible in India. Google recorded 68 per cent year-on-year growth in the country's ultra-premium smartphone segment during the second quarter of 2026, supported by wider offline distribution, marketing and stable pricing. That performance followed 39 per cent growth in the premium segment during the first quarter.

Pixel's improving position gives Google a stronger commercial reason to create a more resilient manufacturing system. The brand had already entered the world's five largest premium smartphone vendors after Pixel sales doubled year on year during the first half of 2025, supported by the Pixel 9 series and expansion into additional markets.

The 2027 restructuring is expected to encompass Pixel smartphones, Pixel Watch devices and Pixel Buds rather than smartphones alone. That breadth distinguishes the plan from earlier diversification measures that shifted selected handset models or portions of production outside China.

China nevertheless remains difficult to remove completely from electronics supply chains because many components, materials, manufacturing tools and specialist suppliers remain concentrated there. Moving final production does not necessarily mean eliminating Chinese-origin parts, making supplier diversification a longer process than relocating assembly lines.

Google's experience with Vietnam reflects that challenge. Earlier Pixel generations had already been assembled there, but transferring new-product development is strategically more significant because engineers and suppliers must establish production processes before commercial manufacturing begins. Completing those tasks for flagship devices demonstrates capabilities that previously remained concentrated largely within China's mature electronics ecosystem.

India is following a different trajectory, using large-scale smartphone assembly as a foundation for deeper component manufacturing. Government incentives and expanding domestic electronics investment have encouraged contract manufacturers to increase capacity, while export production allows facilities to operate beyond demand from the local market.

Sebi warns against attempts to manipulate closing auction

Securities and Exchange Board of India Chairman Tuhin Kanta Pandey has warned market participants against attempts to manipulate the newly introduced Closing Auction Session, saying the regulator has stronger surveillance capabilities under the mechanism and will act sternly against misconduct.

Pandey said on Wednesday that any attempt to undermine or “defame” the Closing Auction Session, or CAS, through manipulation would attract regulatory action. He also reinforced Sebi’s position that the mechanism, introduced on August 3, is here to stay despite concerns raised by brokers and other market participants during its opening weeks.

The warning comes as Sebi and stock exchanges work to improve liquidity and price discovery during the 20-minute auction that determines closing prices for eligible shares. The regulator has asked brokers to facilitate order placement during the five-minute transition between 3.15 pm and 3.20 pm from September, a change aimed at increasing participation before orders are matched.

CAS operates between 3.15 pm and 3.35 pm on trading days. During its initial phase, it applies to shares in the equity cash market on which derivative contracts are available. Other securities continue to use the earlier method for calculating closing prices.

The previous framework determined the closing price using the volume-weighted average price of trades conducted during the last 30 minutes of continuous trading. Under CAS, buy and sell orders are brought together in a separate auction, with the closing price established around the level at which the maximum executable quantity can be matched.

Pandey said Sebi’s ability to identify manipulation is greater under the auction framework than under the volume-weighted average system. The regulator views the mechanism as an important market-structure reform designed to make closing-price formation more transparent and bring domestic practices closer to those followed by major international markets.

The rollout initially produced sharp movements in benchmark indices and prompted concerns over thin liquidity during the auction. The Nifty 50 and BSE Sensex experienced unusual closing-stage divergences during some sessions, drawing attention from traders, brokers and institutional investors.

Market participants have linked some of the early volatility to limited depth in the securities lending and borrowing market. Short-selling constraints can make it difficult for participants to respond quickly when auction prices diverge from levels prevailing during continuous trading.

Sebi has maintained that it has not detected manipulation behind the unusual price movements observed since CAS began. Pandey earlier said cautious participation during the initial phase contributed to some of the differences in closing levels and argued that greater familiarity with the mechanism should improve liquidity.

Institutional participation has already strengthened. Mutual funds accounted for only about 5-6 per cent of activity during the first day of the closing auction, but their share subsequently climbed to roughly 20-25 per cent. Greater participation from institutions, arbitrage desks and other large market players is expected to narrow price discrepancies and deepen the auction order book.

Stock exchanges have also introduced adjustments intended to make the system easier for traders to follow. Indicative index prices are now displayed during the closing auction, giving participants greater visibility into the potential effect of individual share prices on benchmark indices.

Shorter-tenure contracts have also been introduced in the securities lending and borrowing segment, addressing concerns that the earlier structure provided insufficient flexibility for participants seeking to hedge or arbitrage positions around the closing auction.

Sebi has indicated that feedback from brokers and investors remains under examination even as it rules out abandoning CAS. Pandey said earlier this week that the regulator was analysing issues raised during implementation and would consider changes where required to make the mechanism function more efficiently.

The auction is particularly significant for passive investment products such as exchange-traded funds and index funds. Closing prices influence index calculations, portfolio valuations and the execution of large institutional orders. Sebi expects a deeper closing auction to reduce tracking errors and provide investors with a more representative end-of-day price.

Closing auctions are already standard features at several major international exchanges, where substantial institutional trading takes place around the market close. Sebi designed the domestic framework around the principle of establishing a single equilibrium price through the matching of aggregate supply and demand rather than relying on an average of earlier trades.

Korean brain-controlled robotic hand makes Beijing debut

Dynamic Solution has unveiled a brain-computer interface robotic hand at the World Robot Conference 2026 in Beijing, demonstrating how electrical activity measured from the scalp can be converted into commands for dexterous robotic movement without surgical implants.

The South Korean technology company is presenting the system during the five-day conference that opened on August 19 at the Beiren Etrong International Exhibition and Convention Center in Beijing E-Town. The opening day has been designated “Release Day”, focusing on new products, technological achievements and collaborative initiatives. More than 300 companies and research organisations are participating, with more than 2,000 exhibits expected.

Dynamic Solution’s robotic hand combines brain-computer interface technology with individually actuated fingers capable of performing multiple gripping movements. The system analyses electroencephalography, or EEG, signals produced when a user imagines moving a hand and converts recognised patterns into commands that drive the robotic device.

The approach is non-invasive because brain activity is detected by electrodes positioned on the scalp rather than by surgically implanted interfaces. That distinction could prove important commercially because non-invasive systems avoid the medical risks and regulatory complexities associated with implanting electrodes in the brain, although they generally face greater challenges in obtaining highly precise signals through the skull and surrounding tissue.

The robotic hand is approximately the size of an adult male hand and uses separate actuators for individual fingers. Contact areas incorporate anti-slip material intended to improve the stability of grasping. Dynamic Solution is seeking to combine such mechanical capabilities with artificial intelligence and neural-signal interpretation, placing the product within the rapidly developing field of physical AI, where intelligent software directly controls machines operating in the physical world.

The company has also acquired a non-exclusive licence for X-HAND and remote wearable robot technologies developed by the Electronics and Telecommunications Research Institute. The underlying platform incorporates wearable and haptic technologies designed to reproduce movement and tactile feedback, providing a potential route towards systems in which information flows both from the user to the robot and back from the robotic device.

Dynamic Solution, formerly known as Neofect, has spent years developing rehabilitation and wearable technologies. Its intellectual-property portfolio includes technologies covering sensing, movement assistance, upper-limb rehabilitation robots, finger-motion measurement and wearable robotic hands. The company says it holds 114 related patents in domestic and overseas markets.

The Beijing demonstration also comes as South Korea expands national investment and institutional support for brain-computer interfaces, humanoid systems and physical AI. The Ministry of Science and ICT formally launched the K-Moonshot programme this year with 12 national technology missions extending to 2035, including separate programmes for BCI, humanoids and physical AI. The initiative aims to use artificial intelligence to accelerate scientific research and double research productivity by 2030.

That government programme is broader than Dynamic Solution’s non-invasive robotic-hand project and includes work on implanted BCI technologies. Its significance for companies developing EEG-based systems lies mainly in the research ecosystem being built around neuroscience, robotics, artificial intelligence and commercialisation rather than direct state sponsorship of a single product.

Dynamic Solution has been positioning itself within that ecosystem through partnerships with hospitals, research institutes and BCI specialists. It signed an agreement with Ybrain this year covering brain-signal precision control, AI robotics, BCI commercialisation and joint participation in national research projects. The companies have identified applications that could ultimately include assistive technology for people with severe motor disabilities, wearable robots and industrial systems.

The company is also involved in collaborative work linking South Korean and German researchers, including Seoul National University Bundang Hospital and Charité in Berlin, as it investigates the clinical potential of brain-controlled robotic technologies.

Commercial viability will depend heavily on performance that cannot be judged from a product demonstration alone. Dynamic Solution has not publicly released detailed benchmark data covering the X-HAND system’s EEG decoding accuracy, command latency, error rates or reliability across large groups of users. Those measurements will be crucial for determining whether a brain-controlled hand can progress from controlled demonstrations to dependable everyday use.

Non-invasive BCI systems also have to cope with electrical noise, variations between users, movement artefacts and changing signal patterns over time. Machine-learning models can improve interpretation, but practical assistive devices must respond quickly while avoiding unintended commands.

The potential applications nevertheless extend well beyond exhibition demonstrations. A reliable interface capable of translating motor intention into robotic movement could eventually assist people with paralysis or severe mobility impairment, while related technologies could be adapted for rehabilitation, teleoperation and human-machine interaction.

US sanctions deepen confrontation with International Criminal Court

The United States has imposed sanctions on International Criminal Court President Tomoko Akane of Japan and senior prosecutor Abdoulaye Seye of Senegal, widening Washington’s campaign against the Hague-based tribunal over cases involving Israel and challenges to US sovereignty.

The designations announced on Tuesday target two of the court’s most senior figures under an executive order signed by President Donald Trump in February 2025. The order authorises financial and travel restrictions against foreign nationals involved in ICC attempts to investigate, arrest, detain or prosecute protected US or allied personnel without the consent of their governments.

Secretary of State Marco Rubio said Akane and Seye had directly participated in such efforts. Washington argues that the ICC has exceeded its mandate by pursuing officials from countries that are not parties to the Rome Statute, including the United States and Israel. The administration has made curbing the court’s ability to act against US personnel and officials of close allies a foreign-policy priority.

The sanctions block property and financial interests that the two officials may hold under US jurisdiction and generally prohibit Americans and US companies from dealing with them. Their broader effect can extend well beyond the United States because international banks and technology companies frequently depend on the US financial system. The Treasury Department has authorised transactions needed to wind down dealings involving Akane until September 17.

Akane, who joined the ICC bench in 2018 and became president in 2024, has been a prominent defender of the court’s judicial independence during its confrontation with Washington. She has argued that political or economic pressure cannot determine how judges interpret the Rome Statute or decide cases.

Seye is a Senegalese lawyer serving in the Office of the Prosecutor. His designation places additional pressure on the prosecutorial structure responsible for investigating alleged crimes connected with the war in Gaza. The sanctions bring the number of ICC judges and prosecutors targeted by Washington since Trump returned to office to at least 13, following successive measures imposed through 2025.

The dispute centres heavily on the ICC’s Palestine investigation. Judges issued arrest warrants on November 21, 2024, for Israeli Prime Minister Benjamin Netanyahu and then defence minister Yoav Gallant over alleged war crimes and crimes against humanity. Israel rejects the allegations and disputes the court’s authority over its nationals.

Washington takes the same jurisdictional position, stressing that neither the United States nor Israel has joined the Rome Statute. The ICC maintains that membership is not the only basis for jurisdiction. Palestine acceded to the statute in 2015 and accepted the court’s jurisdiction over alleged crimes committed on its territory, giving prosecutors a legal basis to investigate conduct there regardless of the nationality of an alleged perpetrator.

That disagreement has developed into a broader struggle over the reach of international criminal justice. The United States has also objected to the ICC’s earlier Afghanistan investigation, which examined alleged abuses involving US military and intelligence personnel. Trump’s executive order describes ICC action against non-consenting states as an extraordinary threat to US national security and foreign policy.

The court and its supporters counter that allowing nationality alone to shield people from prosecution for offences allegedly committed on the territory of a member state would severely restrict the Rome Statute. The ICC was created as a court of last resort to prosecute genocide, crimes against humanity, war crimes and, under defined circumstances, aggression when national judicial systems cannot or will not genuinely pursue cases. It currently has 125 states parties.

The sanctions campaign has also moved into US courts. Three ICC judges filed a lawsuit against the Trump administration in Manhattan in June, arguing that sanctions imposed on them exceeded presidential authority and were intended to coerce judicial decision-making. They described restrictions on banking, commercial services and personal transactions as punishment for carrying out their judicial functions.

Baidu revenue slips as ad weakness eclipses AI growth

Baidu’s second-quarter revenue fell short of market expectations as a steep decline in its search advertising business outweighed strong growth in artificial intelligence cloud infrastructure, highlighting the cost and complexity of the company’s transition towards an AI-led business model.

Revenue for the three months ended June fell 4% from a year earlier to 31.33 billion yuan, or about $4.62 billion, compared with market expectations of roughly 31.96 billion yuan. The results sent Baidu’s US-listed shares sharply lower on Tuesday as investors weighed continued weakness in its traditional cash-generating operations against rapid expansion in AI services.

Online marketing services revenue dropped 19% year on year to 13.1 billion yuan. The decline reflected softer advertising demand as businesses remained cautious about marketing expenditure amid prolonged weakness in the property sector and restrained consumer spending.

The advertising contraction overshadowed continued growth in Baidu’s AI-powered operations. Revenue from its Core AI-powered Business rose 25% to 12.5 billion yuan and accounted for half of Baidu General Business revenue during the quarter.

AI Cloud Infrastructure revenue climbed 50% to 7.3 billion yuan, supported by rising demand from companies adopting generative AI applications and seeking access to computing infrastructure. Revenue from GPU Cloud services surged 283%, accelerating from growth of 184% in the first quarter.

The figures nevertheless showed some sequential moderation. AI Cloud Infrastructure revenue declined from 8.8 billion yuan in the March quarter, while total Core AI-powered Business revenue slipped 8% quarter on quarter from 13.6 billion yuan.

Baidu is seeking to convince investors that the rapid expansion of its AI operations can eventually offset structural pressure on search advertising. Chief executive Robin Li told analysts that sustained investment in technology, applications and talent would remain central to the company’s strategy as it works to strengthen its position in China’s increasingly competitive AI market.

Competition has intensified as Alibaba, ByteDance, Tencent and AI start-ups including Moonshot AI roll out new foundation models and applications. Baidu was among the earliest major technology groups in China to launch a ChatGPT-style product through its Ernie platform, but rivals have narrowed the gap as successive generations of models have entered the market.

Li said Baidu intended to return Ernie to the frontier of artificial intelligence development, signalling renewed emphasis on improving its foundation-model capabilities. The company has increasingly sought to commercialise AI through cloud computing, enterprise applications, marketing services and autonomous driving rather than relying primarily on consumer chatbot adoption.

Revenue from AI Applications increased 3% to 2.5 billion yuan during the quarter. AI-native marketing services generated 2.6 billion yuan, little changed from a year earlier. Baidu said artificial intelligence features were gaining broader usage across products including Baidu Wenku and Baidu Drive.

The shift is requiring substantial spending on computing infrastructure and personnel, creating pressure on profitability even as AI revenue expands. Research and development expenses stood at 4.6 billion yuan during the quarter, 5% higher than in the previous three months, although they were 10% lower year on year.

Net income attributable to Baidu fell to 2.32 billion yuan from 7.32 billion yuan a year earlier. Diluted earnings per American depositary share were 5.74 yuan, while adjusted diluted earnings came to 7.22 yuan.

Operating income was 3.02 billion yuan, producing an operating margin of 10%. Adjusted operating income reached 3.8 billion yuan, with an adjusted margin of 12%. Baidu generated operating cash flow of 3.4 billion yuan and ended June with cash and investments totalling 283.1 billion yuan.

The company’s broader business also continues to contend with weakness at video-streaming platform iQIYI. Revenue from iQIYI fell 5% year on year to 6.3 billion yuan. Baidu General Business revenue declined 4% to 25.2 billion yuan.

Its autonomous-driving operation, Apollo Go, is meanwhile expanding beyond China. The service has begun open-road testing in London through partnerships involving Uber and Lyft and launched fully driverless commercial operations in Dubai. It has also secured permits for fully driverless testing in Hong Kong and conducted road testing in Switzerland.

India prepares first Arctic cargo voyage to Russia

India is preparing to send its first cargo vessel through Russia’s Northern Sea Route in 2027, opening a new maritime option that could deepen commercial links with Moscow while testing the economics and operational risks of Arctic shipping.

The plan has emerged from talks between New Delhi and Russia’s Arkhangelsk region over the commercial potential of the route. Arkhangelsk, on the Northern Dvina River near the White Sea, is an established cargo and industrial centre and is increasingly being positioned as a western gateway for traffic moving through Russia’s Arctic waters.

The Northern Sea Route, or NSR, runs for roughly 5,600 km along Russia’s northern coastline between the Barents Sea region and the Bering Strait. It can provide a significantly shorter connection between parts of Europe and Asia than routes through the Suez Canal, although the commercial advantage varies sharply depending on a ship’s origin, destination, season, cargo and ice conditions.

For India, the proposed voyage would be significant because New Delhi and Moscow have spent several years examining alternatives to established trade corridors. Their July 2024 summit declaration formally backed cooperation in developing shipping between the two countries through the Northern Sea Route, while the issue remained part of bilateral discussions during the Russia-India annual summit in December 2025.

Interest has grown as Russia expands Arctic infrastructure and seeks to increase the NSR’s role in international commerce. Transit cargo on the route reached a record 3.2 million tonnes in 2025, while the number of container transit voyages rose to 24 from 14 a year earlier. The total number of voyages along the route climbed 14% to 1,521. Russia also operates a large nuclear-powered icebreaker fleet to support navigation through difficult conditions.

The commercial attraction lies principally in distance and time. A China-Europe container transit through the NSR in 2025 took about 20 days, compared with around 40 days using the longer southern route. Such savings, however, cannot automatically be applied to India-Russia trade because geography is less favourable for some cargo movements originating from ports in the Arabian Sea or along India’s western coast.

A pilot voyage would therefore allow shippers to test actual freight costs, insurance charges, ice-class requirements, port handling, navigation support and seasonal reliability before considering regular services. Thick Arctic ice still restricts much of the route for long periods, and conventional vessels may require icebreaker assistance or specialised hull specifications.

Studies of Arctic shipping also caution that shorter distance does not always translate into proportionate savings. Lower operating speeds, weather constraints, sea-ice diversion, higher insurance premiums and specialised vessel requirements can erode some of the theoretical cost advantage over Suez. Environmental risks are also greater in remote Arctic waters, where oil spills, mechanical failures and rescue operations can be more difficult to manage.

Russia nevertheless continues to accelerate use of the corridor. During the 2026 navigation season, crude shipments towards Asia began at a faster pace than the previous year, while additional domestically built Arc7 ice-class LNG carriers entered service. These ships are designed to operate in severe Arctic conditions and can move through ice that would stop ordinary commercial vessels.

China’s expanding role is another factor shaping India’s calculations. Russia and China have developed container links across the NSR, and a roadmap envisages increasing bilateral cargo traffic on the route to 20 million tonnes by 2030. Regularised Arctic services between Chinese and European ports are also demonstrating that container traffic can move through the corridor during favourable navigation periods.

For New Delhi, the Arctic route would complement rather than immediately replace established maritime corridors. India is already pursuing the International North-South Transport Corridor through Iran and the Caspian region, while the Chennai-Vladivostok maritime corridor has been promoted as another link with Russia’s Far East.

Arkhangelsk could become particularly relevant if cargo arriving through Arctic waters can connect efficiently with Russia’s rail and inland logistics network. Its location gives Russia a potential distribution point for goods moving between the NSR, European Russia and neighbouring markets.

Delhi Police defends force as court orders scrutiny

Delhi Police has rejected allegations that excessive force was used against student protesters at Jantar Mantar, telling the Supreme Court that officers exercised restraint while confronting a crowd that attempted to march towards Parliament on July 20. The court, however, has decided to constitute a high-powered committee to examine competing allegations of violence by police personnel and protesters.

The police defence was presented in a counter-affidavit as the Supreme Court considered a batch of petitions seeking an independent investigation into the handling of demonstrations linked to protests over examination paper leaks. Police said the proposed march towards Parliament had no permission and described the attempt to proceed beyond the designated protest area as unlawful.

Police estimated that more than 30,000 protesters were present around Jantar Mantar and adjoining areas, spread across roughly three kilometres, while about 5,000 police personnel had been deployed to manage the gathering. The affidavit said sections of the crowd breached barricades and the situation deteriorated as demonstrators attempted to advance towards Parliament.

Authorities maintained that force was applied in stages and only after attempts at crowd control failed. They disputed allegations that their response amounted to indiscriminate violence, arguing that photographs and video clips cited by petitioners presented an incomplete account of the confrontation.

The police also alleged that anti-social elements and people with criminal records had entered the demonstration and contributed to violence. More than 240 police and security personnel were injured, the affidavit said, while around 200 protesters and other members of the public suffered injuries. The figures presented to the court are higher than some initial assessments issued immediately after the July 20 confrontation.

Petitioners have offered a sharply different account. They have alleged indiscriminate baton charges, use of tear gas, electronic batons and projectile weapons against demonstrators, many of whom were students. One of the petitions specifically challenges the alleged use of pellet guns for civilian crowd control.

The Supreme Court had already found that the allegations disclosed a prima facie case warranting an independent and impartial examination. Material placed before the court included claims that students sustained pellet and rubber-bullet injuries, that one person suffered serious damage to his eyesight and that journalists and other civilians were assaulted. The court has also taken note of allegations that protesters attacked police personnel with stones and other objects.

A bench comprising Chief Justice Surya Kant and Justices Joymalya Bagchi and V Mohana has sought to keep scrutiny focused on both sides of the confrontation. Earlier proceedings produced the court's observation that neither a police officer who used excessive force nor a criminal posing as a student protester should receive protection.

The court had previously directed authorities to preserve CCTV recordings, drone footage, body-camera recordings, police wireless communications and other evidence connected with the demonstrations. It also ordered protection of protesters' personal and digital information and directed that children below 18 without criminal antecedents who had been detained or arrested in connection with the protests be released.

The July 20 confrontation developed after protesters participating in the youth-led movement gathered at Jantar Mantar and attempted to march towards Parliament. The campaign had grown around anger over examination irregularities and demands for greater accountability in the education system. Police used batons and tear gas as demonstrators pushed towards restricted areas, while protesters accused the authorities of attempting to suppress a largely student-driven movement.

The dispute has consequently moved beyond the question of whether police were entitled to stop an unauthorised march. The proceedings now centre on whether the methods employed were proportionate to the threat faced and whether individual officers or protesters crossed legal limits.

Delhi University suspends scholar after professor assault

Delhi University has suspended a Faculty of Law PhD scholar and barred him from its premises after he allegedly assaulted a professor inside a campus building, prompting a police case and an independent university inquiry.

Shubham Singh, 31, was suspended with immediate effect following the incident at Umang Bhawan at about 10.30 am on Monday, August 17. The scholar, admitted to the Faculty of Law during the 2024-25 academic session, has also been prohibited from entering university premises while disciplinary proceedings are under way.

The professor has been identified as Anupam Jha, a Faculty of Law teacher who has served at the university for 23 years. CCTV footage of the episode shows a man waiting near the entrance before approaching the professor. The footage circulating online appears to show the confrontation developing rapidly, with others nearby attempting to intervene.

Delhi Police have registered an FIR at Maurice Nagar police station and begun an investigation. The university has separately constituted a three-member committee to establish the circumstances surrounding the confrontation and recommend further action. The committee is headed by chemistry professor Rajeev Gupta and includes professors Vipul Singh and Sudha Singh.

Jha has said he suffered an injury to his elbow and bruising to his hand during the episode. He said Singh was not his student and that interaction between them had been limited. Jha linked the dispute to an earlier encounter following Singh's PhD admission process, saying the scholar had confronted him in July 2025 over questions asked during an admission interview.

Singh has presented a different account of the background to the confrontation. He described his actions as stemming from accumulated frustration connected with the PhD interview process and questions raised during his admission attempts. Singh completed his LLB in 2020 and LLM in 2024 before enrolling for doctoral research at the Faculty of Law.

The scholar said his proposed research concerned the continuation of colonial-era approaches to relations between the state and religion. He said questions during the admission process about terminology used in the proposal had remained a source of dissatisfaction. The university inquiry is expected to examine these claims alongside CCTV material, accounts from those present and other evidence surrounding the alleged assault.

The disciplinary order was issued by the Office of the Proctor under provisions governing student conduct. University Proctor Manoj Kumar Singh confirmed that an inquiry had been ordered after the institution received CCTV footage relating to the incident. The entry ban was imposed under Ordinance XV-B, which provides the university with disciplinary powers in cases involving student misconduct.

The episode has intensified concern among Delhi University teachers over safety on campus. The Democratic Teachers' Front said suspending the scholar did not by itself adequately address the seriousness of the incident and called for swift disciplinary measures as well as enforceable safeguards to protect teachers from violence, intimidation and abuse.

The teachers' organisation said the assault should not be viewed as an isolated episode. It referred to another case in which Faculty of Law teachers were allegedly subjected to verbal abuse and argued that repeated incidents were undermining confidence among faculty members that threats or attacks would bring sufficiently strong institutional action.

Faculty groups have also pointed to an October 2025 confrontation at Dr Bhim Rao Ambedkar College, affiliated with Delhi University, where Delhi University Students' Union joint secretary Deepika Jha was accused of slapping professor Sujit Kumar during a dispute. That incident was recorded on video and triggered protests by teachers as well as a university-level investigation.

Delhi University later imposed disciplinary action against Deepika Jha, while the episode remained a reference point for faculty organisations pressing the administration to adopt a firmer response to physical confrontations involving teachers. The Democratic Teachers' Front has now sought clear preventive measures and faster action whenever faculty members face violence or threats.

Pant becomes first India batter with 100 Test sixes

Rishabh Pant became the first player from India to hit 100 sixes in Test cricket on Tuesday, adding a world pace record to the landmark during the opening Test against Sri Lanka at Galle.

The wicketkeeper-batter reached three figures for career Test sixes when he launched fast bowler Lahiru Kumara over the boundary during India’s second innings on the fourth day. The stroke also completed Pant’s 20th Test half-century and underlined his position as one of the most aggressive batters in the longer format.

Pant reached 100 sixes in his 89th Test innings, comfortably the quickest any batter has attained the milestone. Adam Gilchrist had previously required 130 innings, while Ben Stokes took 151 and Brendon McCullum 170.

The 28-year-old also became only the fourth player to reach 100 Test sixes. Stokes leads the all-time list with 138, followed by former New Zealand captain McCullum with 107. Pant moved alongside former Australia wicketkeeper Gilchrist on 100.

Pant had entered the Galle Test with 97 sixes from 50 matches. He struck one maximum during his first-innings 39, taking him to 98, before clearing the ropes twice in the second innings to become the first player from India to reach the century mark.

Virender Sehwag and Rohit Sharma had previously set the benchmark among players from India with 91 Test sixes each. Pant has passed them despite playing significantly fewer matches, reflecting the unusually high frequency with which he has attacked the boundary during his Test career.

His milestone six came during an important phase of India’s second innings. Pant walked in with the score at 81 for three after captain Shubman Gill was dismissed, with India seeking to build on a first-innings advantage of 178 runs.

Pant was initially measured, waiting until his 13th delivery for his first boundary. He then accelerated sharply. A powerful six off Kumara sent the ball out of the ground and forced a replacement before Pant targeted the Sri Lanka attack more aggressively as India sought to stretch the lead beyond manageable proportions.

The landmark arrived when Pant again attacked Kumara. His fifty came from 53 balls, with the six taking his career tally to 100. He continued to score quickly and was eventually dismissed for 66 from 69 deliveries, an innings that contained a succession of unconventional attacking strokes.

Pant’s contribution strengthened India’s already commanding position after the tourists had controlled much of the match. India made 462 in their first innings, built around Devdutt Padikkal’s maiden Test century. Padikkal scored 167, while KL Rahul contributed 82 and Dhruv Jurel made 51.

Sri Lanka responded with 284 despite a century from Sonal Dinusha and 80 from Niroshan Dickwella. Left-arm spinners Manav Suthar and Ravindra Jadeja shared seven wickets, giving India a sizeable advantage before their second innings began.

Pant’s achievement adds another distinctive record to a Test career defined by high-risk batting in demanding situations. Since making his debut against England at Nottingham in 2018, he has established himself as a rare wicketkeeper capable of changing the course of Test matches through sustained attacking play.

His six-hitting rate is particularly striking because Test cricket traditionally offers fewer opportunities for sustained power hitting than limited-overs formats. Pant required 4,918 deliveries faced to reach 100 sixes, substantially fewer than the other players to have achieved the landmark.

Gilchrist needed 6,578 deliveries to complete his century of sixes, while Stokes required 9,042 and McCullum 9,756. Pant’s numbers therefore reflect not simply longevity but the frequency with which he has used the aerial route throughout his career.