Militaries accelerate AI use as safeguards lag

Military adoption of artificial intelligence is advancing faster than international efforts to set common safeguards, a gap highlighted this week at the Beijing Xiangshan Forum as defence officials and experts warned that automated systems are compressing decision times and increasing escalation risks.

The three-day forum, which ended on Thursday, brought together about 2,000 military officers, diplomats and academics from more than 100 countries and organisations. China’s Ministry of National Defence said one discussion track focused on the risks and regulation of military applications of emerging technologies.

Several delegates argued that human judgement must remain central as armed forces deploy autonomous systems, machine-assisted targeting tools and AI-enabled command functions. Jürg Lauber, vice-president of the International Committee of the Red Cross, said growing autonomy in weapons systems made preservation of human judgement and control over the use of force increasingly important.

Pakistan’s Defence Secretary Muhammad Ali said AI was accelerating military decision-making while misinformation and disinformation could erode public trust before governments had time to respond. Thailand’s Defence Minister Adul Boonthumjaroen said the central challenge was not simply which country would become more powerful, but how an international order could be built around fast-moving technologies.

The debate comes amid weaker momentum around military AI governance. At the third Responsible AI in the Military Domain summit in A Coruña, Spain, in February, neither the United States nor China endorsed the meeting’s outcome document, known as Pathways to Action. The declaration remains open for additional endorsements and is now backed by more than 40 countries.

That marked a change from the first REAIM summit in The Hague in 2023, when both Washington and Beijing supported the Call to Action. The United States also backed the 2024 Blueprint for Action adopted at the second summit in Seoul, while China did not. The 2026 document moved further towards practical measures covering oversight, accountability, testing, risk management and responsible deployment.

The absence of both AI powers from the latest endorsement list has sharpened attention on middle powers such as the Netherlands, South Korea, Spain, Singapore, Canada and Australia, which have played roles in sustaining multilateral discussions. Analysts argue these countries may have greater space to advance operational standards even if the largest military and technology powers remain cautious about binding constraints.

Military applications are meanwhile expanding rapidly. Armed forces are using AI for intelligence analysis, target identification, logistics, cyber operations, drone navigation, battlefield awareness and command support. The spread of low-cost unmanned systems has also increased pressure for faster automated processing, particularly where communication links are disrupted or operators must respond within seconds.

Governments broadly agree on responsible use, but differences persist over what should be legally binding, how much human control is required and whether certain applications should be prohibited. Nuclear command systems remain among the most sensitive areas. Security experts from the United States and China have called for clearer red lines, human oversight of critical systems and dedicated channels to manage AI-related incidents, but the proposals have not been adopted by either government.

The United States and China have pursued separate national approaches to AI policy while continuing strategic competition over advanced semiconductors, computing power and model development. Both governments have acknowledged risks from military AI, yet their rivalry has complicated efforts to translate general principles into shared mechanisms.

REAIM’s 2026 Pathways to Action calls for responsible design, development, testing, deployment and use of military AI, along with stronger oversight and accountability across the technology’s life cycle. It also stresses capacity-building and knowledge-sharing among states, reflecting concern that many governments lack the technical expertise needed to assess increasingly complex systems.

The Beijing forum showed that anxiety over these risks is no longer confined to specialist meetings. Representatives from several Asian countries raised concerns about autonomy, compressed strategic timelines and misinformation, while discussion also covered potential benefits of AI for defence planning and operational efficiency.

Google widens Gemini push across student learning

Google is expanding Gemini deeper into education, giving eligible college students in India a year of Google AI Plus at no cost while introducing personalised study tools, interactive learning features and broader support for exam preparation.

The offer is part of a wider international push covering more than 140 countries and places Gemini more directly into students’ everyday academic routines. Eligible students in India receive Google AI Plus for 12 months, including higher Gemini usage limits, access to advanced artificial intelligence capabilities and 400 GB of cloud storage.

The programme is open to eligible university and college students aged 18 or above who are enrolled at qualifying degree or certificate-granting institutions. Students have to verify their academic status before activating the plan. The offer can be redeemed until December 31, 2026, and requires a valid payment method at registration.

Google AI Plus normally costs ₹399 a month in India. Unless cancelled, subscriptions activated through the student promotion will convert to the prevailing paid rate after the free period. Students whose earlier 2025 AI Pro student trial has expired may also qualify, subject to eligibility verification.

At the centre of the education expansion is a dedicated Student Hub within Gemini. It brings together study notebooks, flashcards, practice quizzes and other learning functions in a single interface. Students can upload syllabi, lecture notes and course materials and use them to build personalised learning programmes.

Study notebooks are designed to identify gaps rather than simply produce answers. A student can begin with a diagnostic quiz, after which Gemini creates shorter lessons based on areas requiring greater attention. Subsequent quizzes track performance and allow the programme to adjust the study plan as the learner progresses.

The notebooks, introduced on desktop earlier this year, have now expanded to mobile devices. They can incorporate course deadlines, notes and supporting material while connecting with other Google learning services, giving students a more structured way to manage subjects across a semester.

Gemini is also adding interactive visualisations aimed at subjects where static explanations may be insufficient. AI-generated tables, diagrams and three-dimensional simulations can demonstrate concepts such as molecular structures, mathematical relationships and scientific processes, allowing users to rotate, explore or modify visual material.

Gemini Live is being extended into research and study workflows as well. Students can initiate detailed research tasks and later discuss the results conversationally using voice. This gives the service a tutoring-style dimension alongside its established role as a generative AI assistant.

Exam preparation has become another significant part of Google's education strategy. Gemini already supports full-length practice material for examinations including JEE Main and NEET UG, with material developed around vetted educational content. Students completing mock examinations can receive feedback identifying stronger areas and subjects requiring further revision.

The company has also been expanding practice options for international examinations including the SAT, ACT and GRE. The approach moves generative AI closer to adaptive test preparation, an area traditionally served by tutoring businesses, specialist educational platforms and publishers.

Google Search is being tied more closely to the same learning ecosystem. Students can request personalised quizzes across subjects and use AI-generated interactive visuals to examine difficult concepts. Search can also work with uploaded PDFs, presentation files, photographs and handwritten notes to create study documents built around the material supplied by the user.

Lens is being developed as another route into AI-assisted learning. Students can photograph a problem or study material and ask the system to explain the underlying concept, identify possible mistakes and guide them towards a solution rather than merely supplying an answer.

The strategy reflects intensifying competition among technology companies to establish generative AI as a routine educational tool. Students represent an important user group because habits developed during university years can influence which productivity and AI services they later adopt in professional life.

India has become particularly important to that contest. More than two million students in the country gained access to Google's advanced AI tools through its earlier student programme, while the 18-to-24 age group has emerged as a major component of Gemini usage.

AI costs push companies towards on-premise computing

Rising expenditure on cloud-based artificial intelligence is prompting companies in India to reassess whether renting high-performance computing capacity remains economical as AI moves from experimentation into large-scale production.

The shift is particularly visible among businesses running persistent inference, model fine-tuning, computer vision and generative AI workloads. Cloud platforms continue to offer speed, flexibility and access to advanced GPUs without heavy upfront investment, but hourly charges can become substantial when expensive accelerators remain active around the clock.

GPU prices available through the IndiaAI Compute Portal illustrate the scale of the calculation facing enterprises. On-demand access to a single Nvidia H100 SXM GPU is listed at about ₹153 an hour, while a 12-month reservation brings the rate down to around ₹117. An H200 SXM GPU is available at roughly ₹140 an hour on demand in some configurations, compared with about ₹100 under longer reservations.

More powerful systems cost considerably more. A two-GPU Nvidia B200 configuration is priced at about ₹581 an hour on demand, while an eight-GPU configuration exceeds ₹2,300 an hour. Prices vary by provider, architecture, reservation period and configuration, making headline GPU rates only one part of the overall expense.

For companies operating workloads continuously, those charges accumulate quickly. An H100 running uninterrupted at ₹153 an hour would generate compute charges of more than ₹1.3 million over a year before storage, networking, data movement and other services are considered. Larger clusters can multiply that figure rapidly.

That arithmetic is strengthening the argument for owning AI infrastructure where utilisation is predictable. Purchasing servers places GPUs directly under enterprise control and removes recurring rental charges, although companies must fund hardware, networking, electricity, cooling, maintenance and technical staff before any savings emerge.

Utilisation is therefore becoming the critical variable.

A GPU server operating only occasionally can become an expensive idle asset. Cloud infrastructure remains attractive for experimental projects, irregular training runs and companies whose AI demand changes sharply from week to week. Capacity can be increased or released within minutes without purchasing equipment that may remain unused.

The equation changes when GPUs operate continuously at high utilisation. Companies running stable production inference or regular model training can potentially spread hardware costs over several years and lower the effective cost of each computing hour.

Hardware ownership nevertheless carries risks that cloud customers largely transfer to their providers. AI processors are evolving rapidly, meaning expensive equipment can lose relative competitiveness long before it physically wears out. New generations from Nvidia, AMD and other suppliers offer higher performance, larger memory and improved energy efficiency, potentially altering the economics of systems purchased only two or three years earlier.

Software compatibility can be equally important. Nvidia's CUDA ecosystem remains deeply embedded across AI frameworks and enterprise applications. Alternative accelerators may offer attractive pricing or memory specifications, but migration can require engineering work, optimisation and testing that reduce theoretical savings.

Performance also varies significantly by workload. A cheaper GPU is not automatically less expensive if a model takes longer to produce the same number of tokens or complete the same training task. Enterprises are increasingly measuring cost per inference, cost per million tokens and useful output per watt rather than relying solely on hourly rental prices.

Cooling and power requirements present another obstacle to bringing AI infrastructure inside corporate facilities. High-density AI racks can demand far more electricity than conventional enterprise servers, with newer configurations requiring specialised liquid cooling and upgraded power distribution.

This is encouraging a third model between public cloud and equipment installed inside company offices. Enterprises can purchase or reserve dedicated GPU infrastructure housed in specialised data centres, combining greater control with professionally managed power, cooling and connectivity.

Hybrid deployment is also gaining ground. Sensitive data and predictable inference can remain on dedicated infrastructure, while cloud GPUs absorb temporary spikes, experimental workloads and unusually large training jobs.

Data governance adds another dimension. Organisations handling financial, healthcare, government or proprietary information may prefer local or dedicated systems because they offer tighter control over where data and models are processed. Cloud providers have responded with private-cloud, sovereign-cloud and dedicated infrastructure options, narrowing some of that distinction.

India's expanding shared compute infrastructure is meanwhile altering the cost equation. More than 38,000 GPUs have been empanelled under the IndiaAI Mission, with another 20,000 planned as part of efforts to broaden access to advanced computing. The programme allows eligible startups, researchers, government bodies and other users to access subsidised or competitively priced GPU capacity without making large capital investments.

Competition among cloud providers, domestic data-centre operators and specialised GPU companies is also pushing enterprises towards more granular purchasing decisions. Reserved capacity, spot pricing, smaller inference accelerators and purpose-built AI processors can reduce bills without forcing companies to abandon cloud infrastructure entirely.

AI pushes IT services beyond billable-hour model

Artificial intelligence is accelerating a fundamental change in India’s IT services industry, pushing outsourcing contracts away from billing for employee hours towards pricing based on productivity, savings and measurable business results.

Major providers including Tata Consultancy Services, Infosys, HCLTech, Wipro, Tech Mahindra and Cognizant are adjusting delivery models as corporate customers demand that AI-generated efficiency translate into lower costs. The shift threatens a decades-old model in which revenue growth was closely connected to adding engineers and charging clients for their time.

India’s technology sector is expected to cross $315 billion in revenue in FY26, with direct employment approaching six million. Yet the industry's growth strategy is increasingly based on value and specialised capabilities rather than expanding headcount. AI has moved from experimentation to industrial-scale deployment, while providers are adopting outcome-based and risk-sharing contracts as automation increases productivity.

Under traditional time-and-material contracts, customers essentially paid for the number of people assigned to a project and the hours they worked. Generative AI is weakening that relationship. Coding assistants, automated testing, software agents and AI-powered maintenance tools allow smaller teams to perform work that previously required substantially larger groups.

Customers consequently want part of those efficiency gains. Some contracts are being structured around specific targets such as faster processing, reduced technology expenditure, improved customer service or quicker software development rather than the size of the delivery team.

The transition is also changing competition. Persistent Systems and Coforge have posted considerably stronger growth than several larger rivals, benefiting from demand for rapid AI pilots and specialised services. During the April-June quarter, Persistent's dollar revenue rose about 16 per cent while Coforge recorded growth of roughly one-third. Growth among several of the largest providers was around 1-3 per cent.

Outcome pricing carries risks for suppliers. Technology companies must estimate in advance how much productivity AI can deliver, potentially exposing themselves to margin pressure when promised efficiencies fail to materialise. Contracts linked to business results can also create disputes over whether the technology provider was responsible for an outcome affected by other parts of a customer's operations.

Tech Mahindra chief executive Mohit Joshi has warned against assumptions that AI productivity will improve by 70-80 per cent over five to seven years while providers guarantee prices despite rising technology infrastructure costs. Some companies have walked away from contracts where pricing or productivity commitments were considered commercially unsustainable.

The labour impact is becoming equally significant. Entry-level software work historically provided the foundation of the industry's employment pyramid, with large numbers of graduates recruited for coding, application maintenance and testing. AI can automate growing portions of these tasks, reducing demand for some conventional junior roles while raising requirements for engineers capable of working with AI systems, cloud platforms, cybersecurity, data and industry-specific technology.

Hiring is therefore moving from volume towards skills. Campus recruitment requirements have shifted from basic coding and cloud knowledge towards AI and machine learning, advanced data analytics and DevOps, while AI and machine-learning positions command some of the industry's highest skills premiums.

That does not mean graduate recruitment is disappearing. Infosys said 10,766 fresh graduates completed training at its Mysuru education centre during FY26, with generative AI and prompt engineering incorporated into foundation training. About 84 per cent of its 328,594 employees are now AI-aware.

TCS has also intensified retraining. More than 270,000 employees had advanced AI skills by the end of FY26, compared with about one-third of that level a year earlier. The company recorded $2.3 billion in annualised AI revenue during the March quarter and is preparing thousands of forward-deployed engineers to work directly with clients on AI implementation.

The changing economics are influencing acquisitions and large contracts as well. TCS this week agreed to buy Porsche's technology consulting subsidiary MHP for an enterprise value of €320 million as part of a five-year partnership valued at €1.25 billion. The programme is designed to expand AI across engineering, manufacturing, operations and customer experience, illustrating how providers are positioning themselves deeper inside clients' businesses rather than supplying technology labour alone.

Tata Sierra.ev gains built-in 5G connectivity

Tata Motors Passenger Vehicles and Tata Communications have partnered to equip the Sierra. ev with built-in 5G cellular connectivity, strengthening the electric SUV’s role as a software-defined vehicle capable of receiving continuous digital upgrades.

The arrangement integrates Tata Communications’ MOVE Connected Vehicle Platform with Tata Motors Passenger Vehicles’ N. IO software-defined vehicle architecture. The combination is designed to provide persistent connectivity for infotainment, vehicle diagnostics, emergency functions and over-the-air software updates throughout the vehicle’s operating life.

The Sierra. ev will use the 5G connection to support faster software downloads, secure content streaming, remote vehicle support and real-time diagnostics. Owners will also be able to purchase additional digital subscription packages, creating scope for services and features to be added after the vehicle has been sold.

The technology marks a shift in how Tata Motors Passenger Vehicles is approaching vehicle development. Rather than treating software as a fixed component installed during manufacturing, the N. IO platform allows functions to evolve through updates delivered remotely. That approach brings passenger vehicles closer to smartphones and other connected devices whose capabilities can change after purchase.

Sven Patuschka, chief technology officer at Tata Motors Passenger Vehicles, said increasingly software-defined vehicles would make intuitive digital connectivity central to customer experience. He described the Sierra. ev as an important step in that transition, with the Tata Communications network providing the digital infrastructure needed for continuous innovation.

The system is also intended to support artificial intelligence-enabled applications as vehicle software becomes more sophisticated. Reliable high-bandwidth connections can allow data generated by cars to be processed more rapidly, supporting personalisation, predictive maintenance and services that depend on near-real-time communication between the vehicle and cloud infrastructure.

Safety-related functions form another part of the deployment. Embedded connectivity can maintain emergency calling services and enable remote assistance when required, while diagnostic data can help identify vehicle problems without the owner first taking the car to a workshop.

Tata Communications executive vice-president Vivek Manglik said vehicles were developing into intelligent digital ecosystems capable of running an expanding range of applications and services. He said the underlying digital infrastructure would be critical to expanding those capabilities securely while improving convenience and personalisation.

The Sierra. ev already offers the iRA. ev connected-car suite with about 70 functions and advanced 5G connectivity. Connected services include remote controls, trip-related functions and over-the-air updates. Tata Motors Passenger Vehicles is offering four years of iRA. ev services from activation, beginning with one year of its higher-tier Ultra Pack followed by three years of Essential Pack access, with options for customers to renew or upgrade.

Connectivity is increasingly becoming a competitive differentiator in the passenger-vehicle market as manufacturers move beyond conventional infotainment systems. Cars capable of downloading software remotely can receive bug fixes, revised interfaces and, where the vehicle architecture permits, additional functions without requiring physical service-centre visits.

The model also provides manufacturers with new commercial opportunities. Digital subscriptions can generate revenue after a vehicle has left the showroom, while real-time data can improve maintenance planning and help companies understand how vehicle features are being used. Such systems, however, also place greater emphasis on cybersecurity, data management and reliable telecommunications coverage.

The Sierra. ev incorporates several other technology-heavy features, including Level 2+ advanced driver-assistance functions on higher variants, digital displays, surround-view systems and app-based infotainment. Its connected platform is designed to provide the communications backbone needed for such technologies to develop further over the vehicle lifecycle.

India’s telecommunications network has expanded sharply since the introduction of commercial 5G services, giving carmakers a larger infrastructure base for connected-vehicle applications. The country had more than 1.28 billion wireless subscribers at the end of March 2026, while wireless data consumption during the January-March quarter reached nearly 77,953 petabytes.

For Tata Communications, connected vehicles represent an extension of its enterprise connectivity and Internet of Things operations. Its MOVE platform is designed to manage cellular connectivity for mobile assets across networks while giving businesses tools to control services and data connections.

US sanctions four India-based firms over Iran trade

The United States has sanctioned four India-based companies over alleged purchases and handling of Iranian petroleum and petrochemical products, extending Washington’s new economic pressure campaign against Tehran to businesses operating in India.

The companies named are Portease Partners LLP, Sadashiva Overseas Limited, PP Softtech Private Limited and Prakrutees Infra Impex India Private Limited. Three individuals associated with the businesses — Indrismiya Ashrafmiya Sheikh, Harish Ramachandra Rangi and Prashant Garg — have also been targeted.

The measures form part of “Operation Economic Outcast”, a broad sanctions drive announced by the administration of President Donald Trump to restrict revenue available to Iran and increase the cost for foreign companies continuing commercial dealings with Tehran.

Sadashiva Overseas Limited was accused of importing about $69 million worth of Iranian-origin petroleum products from several companies between February 2024 and June 2025. The transactions included purchases involving Bonjoure Commodity FZE, an entity already subject to US sanctions.

PP Softtech Private Limited allegedly imported about $25 million worth of Iranian-origin petroleum products between January 2024 and June 2025. Its director, Prashant Garg, was separately designated as part of the action.

Prakrutees Infra Impex India Private Limited was accused of importing roughly $25 million of Iranian petroleum products from several suppliers between May 2023 and February 2026.

Portease Partners LLP, described as a customs broker, allegedly facilitated multiple shipments of Iranian petrochemical products into India. Its designated partners Sheikh and Rangi were included in the sanctions action because of their roles in the company.

Washington said the companies knowingly engaged in significant transactions involving the purchase, acquisition, sale, transport or marketing of petroleum or petroleum products originating in Iran. The designations were made under existing executive authorities governing sanctions on Iran’s energy sector.

The action potentially restricts the targeted businesses’ access to the US financial system and freezes property or interests in property falling within US jurisdiction. Companies dealing with designated entities can also face heightened compliance scrutiny from international banks, insurers, shipping companies and commodity traders.

The inclusion of India-based businesses comes as Washington expands pressure beyond companies directly operating in Iran and focuses increasingly on overseas intermediaries that facilitate trade, shipping, payments and procurement.

Operation Economic Outcast has targeted close to 60 individuals, companies and vessels across several jurisdictions. The wider campaign covers Iran-linked petroleum trading, maritime networks, military procurement, missile-related activity and other channels Washington says generate or move funds for Tehran.

The administration has also warned governments and private businesses that continued economic engagement with Iran could expose them to secondary sanctions. Such measures allow Washington to penalise foreign entities even when transactions do not directly involve US companies or citizens, particularly when dealings involve designated sectors or sanctioned counterparties.

The latest strategy represents an intensification of economic pressure after months of disruption to energy markets and shipping around the Gulf. Iranian crude exports have already fallen sharply from levels recorded before the conflict, while restrictions around the Strait of Hormuz and greater scrutiny of tanker movements have complicated Tehran’s ability to reach overseas buyers.

China remains the largest destination for Iranian crude, although volumes have declined substantially under tighter enforcement and shipping constraints. Independent refiners have historically accounted for much of the trade, often using intermediaries, alternative payment arrangements and complex shipping structures.

India had largely stopped direct purchases of Iranian crude after US sanctions waivers expired in 2019, although trade in other products and humanitarian goods continued. The latest designations illustrate Washington’s increasing focus on identifying petroleum-origin transactions further down the trading chain rather than concentrating only on direct crude imports.

The measures could prompt additional compliance reviews among India-based commodity traders, customs brokers, shipping agents and financial institutions handling international payments. Businesses involved in petroleum products may face greater demands to document product origin, suppliers, beneficial ownership and payment routes before banks or insurers approve transactions.

Weather disruption forces Delhi flight diversions

Heavy rain and thunderstorms disrupted operations at Delhi’s Indira Gandhi International Airport on Monday, forcing at least 15 arriving flights to divert and causing hundreds of delays and more than two dozen cancellations as unstable weather swept across the capital.

Twelve of the diverted aircraft were sent to Jaipur, two to Chandigarh and one to Lucknow after conditions around Delhi deteriorated. International services were among the affected flights, while several aircraft approaching the airport were forced to abandon landings or alter their arrival plans as thunderstorms reduced the operational window available to pilots and air traffic controllers.

The disruption spread rapidly through the airport’s tightly packed schedule. More than 280 flights were delayed during the day and at least 20 were cancelled, with later data putting cancellations at 26. Delhi airport normally handles more than 1,300 aircraft movements daily, making weather-related interruptions capable of producing cascading delays well beyond the capital as aircraft and crews fall out of their planned rotations.

Airlines warned passengers that departures and arrivals could continue to be affected and advised travellers to verify flight status before leaving for the airport. IndiGo said schedules had been disrupted by the weather, while Air India cautioned that services operating to and from Delhi could be affected. SpiceJet also issued passenger guidance as rainfall complicated both airport operations and journeys to the terminals.

Thunderstorms, lightning and gusty winds of around 30 to 40 kmph accompanied the rain across parts of Delhi. Weather authorities issued a red alert for heavy rainfall for parts of the city during the intense spell, increasing the likelihood of temporary restrictions on flight movements when visibility, wind direction or runway conditions moved outside safe operating limits.

Rainfall varied sharply across the capital. Chhatarpur recorded 28 mm between 8.30 am and 11.30 am, while Lodi Road received 24.6 mm. Pitampura recorded 18.5 mm during the same period, with considerably lower totals at the Ridge, Mayur Vihar, Safdarjung and Ayanagar, reflecting the localised intensity typical of monsoon thunderstorms.

The aviation disruption came alongside widespread problems on Delhi’s roads. Waterlogging was reported at several locations, while traffic slowed on important routes connecting residential and commercial districts with the airport. Gurgaon and other parts of the National Capital Region were also affected, adding another layer of uncertainty for passengers attempting to reach terminals for scheduled departures.

Weather diversions are used when pilots and air traffic controllers determine that conditions at the intended airport may prevent a safe landing within operational or fuel limits. Aircraft may be placed in holding patterns while crews wait for conditions to improve, but prolonged thunderstorms can force flights towards designated alternate airports. Jaipur is frequently used as an alternative for Delhi-bound services because of its proximity and airport infrastructure.

Heavy rain alone does not necessarily halt modern airline operations, but thunderstorms present more complex hazards. Strong winds, lightning, rapidly changing visibility and wind shear can affect approaches and departures, while standing water can alter braking performance on runways. Airport capacity can therefore decline even when individual flights remain technically capable of operating, creating queues that ripple through arrival and departure schedules.

Monday’s disruptions underline Delhi airport’s exposure to seasonal weather at a time when passenger and aircraft volumes leave limited room for recovering schedules after prolonged interruptions. Flight delays can continue even after conditions improve because aircraft diverted elsewhere must refuel and reposition, crews can approach regulatory duty limits, and subsequent departures may be waiting for aircraft that have not yet arrived.

Resona deepens Tata Capital partnership for India growth

Japan’s Resona Bank has expanded its ties with Tata Capital through a business cooperation agreement aimed at helping Japanese companies secure financing, identify partners and pursue investment opportunities across India.

The non-exclusive memorandum of understanding announced on Monday builds on Resona Bank’s $20 million commitment as a limited partner in Tata Capital Growth Fund III LP. The vehicle invests in Tata Capital Growth Fund III, an India-focused growth equity fund managed by Tata Capital.

The partnership gives Resona’s corporate clients access to Tata Capital’s extensive domestic network and financial-services expertise as Japanese companies increase their exposure to manufacturing, sales, procurement and research and development operations across the country.

Tata Capital and Resona Bank will explore business matching, introductions and financing opportunities generated through their respective networks. The arrangement is designed particularly for companies assessing market entry, capacity expansion or broader commercial partnerships.

Resona’s investment strengthens the Japanese institutional presence in Tata Capital’s growth-equity franchise, which has maintained relationships with investors from Japan over successive funds. Tata Capital has previously established alliances with Japanese financial groups in areas ranging from private equity and investment banking to structured finance, leasing and cross-border transactions.

Rajiv Sabharwal, managing director and chief executive of Tata Capital, said the partnership combines Tata Capital’s understanding of the domestic market with Resona Bank’s relationships among Japanese businesses. He said Tata Capital Growth Fund III would support companies seeking to participate in the country’s expanding economy.

Mamoru Saito, executive officer at Resona Bank, described India as an increasingly important market for Japanese companies, particularly for manufacturing, sales, procurement and research and development. The agreement reflects a broader shift among Japanese financial institutions towards building local partnerships capable of supporting customers beyond traditional trade finance.

Tata Capital Growth Fund III is targeting about $250 million and is structured as a Category II alternative investment fund. It received its registration in September 2024. Its investment strategy focuses on growth-stage businesses operating across areas linked to urbanisation, strategic services and manufacturing, while also examining opportunities in healthcare, financial services, consumer businesses and technology.

The Asian Development Bank approved an investment of up to $25 million in the fund in June, underscoring growing institutional interest in mid-market companies. The fund is expected to target profitable small and medium-sized businesses capable of using additional equity capital to expand operations, create jobs and strengthen their market positions.

The Resona commitment therefore carries significance beyond its $20 million size. The Japanese bank can combine financial exposure to growing companies with commercial access for its corporate customers, potentially creating opportunities for investments, supplier relationships, technology partnerships and acquisitions.

Japanese companies have maintained a substantial presence across sectors such as automobiles, auto components, electronics, industrial machinery, chemicals and financial services. Businesses are also evaluating the country as an alternative manufacturing and sourcing base as companies diversify supply chains and seek exposure to faster-growing consumer markets.

Financial institutions have responded by broadening their role from lending towards advisory services, private equity connections, acquisition financing and local-market introductions. Resona already operates an Asian network encompassing representative offices and banking partnerships across several major economies, allowing it to support customers pursuing cross-border expansion.

The bank has relationships with multiple lenders operating in India, including Axis Bank, YES Bank and State Bank of India, as part of its overseas network. Its collaboration with Tata Capital adds access to a diversified non-bank finance platform with lending, wealth distribution, commercial finance and private-equity capabilities.

Tata Capital, the Tata group’s flagship financial-services company, has expanded rapidly across consumer and corporate credit. Its consolidated loan book stood above ₹2.2 trillion in the 2024-25 financial year, while profit after tax exceeded ₹36 billion. The company listed on domestic stock exchanges in October 2025.

Its growth-fund platform provides another channel for connecting international capital with privately held businesses. Unlike conventional lending, growth-equity investment allows companies to obtain longer-term capital without immediately increasing debt, an increasingly important option for businesses undertaking expansion, acquisitions or technological upgrades.

DP World pushes for longer Nhava Sheva tenure

DP World is seeking an extension of its concession to operate the Nhava Sheva International Container Terminal at Jawaharlal Nehru Port, reinforcing its long-term commitment to one of the country’s most important container gateways.

The Dubai-headquartered ports and logistics group has opened discussions with the Jawaharlal Nehru Port Authority over continuing operations beyond the existing concession period, which is due to expire on June 30, 2027. The request comes as DP World prepares to deploy an additional $5 billion across ports, logistics and supply-chain infrastructure in the country over the coming years.

Nhava Sheva International Container Terminal, known as NSICT, became the country’s first privately operated container terminal after a 30-year build-operate-transfer agreement was signed in July 1997. The facility has since become an established component of the container-handling network at Jawaharlal Nehru Port near Mumbai.

DP World took control of the terminal after acquiring P&O Ports in 2006. Its continued presence at Nhava Sheva now forms part of a wider strategy to integrate port terminals with rail, warehousing, freight forwarding, coastal shipping and inland logistics operations.

Discussions over the concession have gained urgency because the adjacent Nhava Sheva Gateway Terminal, also operated by DP World, has a concession running until 2031. Port authorities have examined the possibility of aligning the two concession periods so the adjoining facilities could eventually be offered as a larger combined terminal.

The proposed approach would potentially allow the two terminals to operate as a more integrated facility, offering greater berth flexibility and improving the ability to accommodate larger container vessels. Combining terminal operations could also help reduce infrastructure duplication and improve utilisation of available waterfront capacity.

A four-year extension for NSICT has previously been considered as one possible route to align its tenure with the neighbouring terminal. However, the issue involves legal and commercial questions because the original concession agreement does not provide an automatic mechanism for extending the operating period.

Commercial terms are likely to play a central role in the negotiations. The NSICT concession operates under a royalty model, with payments linked to container volumes and annual escalation provisions. Any extension would have to balance revenue expectations for the port authority against the economics of operating and upgrading an ageing terminal.

DP World’s interest in maintaining its Nhava Sheva presence comes alongside a substantial expansion of its infrastructure network. The company has already invested about $3 billion over three decades and has committed another $5 billion to strengthen integrated logistics, multimodal connectivity and maritime infrastructure.

Its network spans more than 200 locations and supports port terminals, inland container depots, rail freight services, warehouses and supply-chain operations. More than 24,000 direct and indirect jobs are linked to the company’s activities across the country.

DP World is also developing a major greenfield container terminal at Tuna Tekra in Deendayal Port in Gujarat. The project is designed with an initial annual capacity of about 2.19 million twenty-foot equivalent units and a berth of around 1,100 metres, with provision for further expansion. Completion is targeted for 2027.

The company operates container terminals at several strategic maritime locations, including Nhava Sheva and Kochi, while expanding inland freight corridors designed to connect manufacturing centres more efficiently with ports.

Jawaharlal Nehru Port remains central to that strategy because of its position serving the Mumbai metropolitan region and industrial belts across western, central and northern parts of the country. The port is connected to the Dedicated Freight Corridor and has invested in common rail infrastructure capable of handling longer and double-stack container trains.

NSICT’s original concession marked a turning point in the development of private participation at major ports. The model helped introduce private capital, equipment and operating expertise into container terminal development at a time when cargo volumes were expanding rapidly.

The port has since developed into a multi-terminal complex involving several private operators, increasing competition and overall container-handling capacity. DP World also operates the 330-metre Nhava Sheva Gateway Terminal, which has design capacity of about 800,000 TEUs annually.

Extending the older terminal’s concession would therefore have implications beyond DP World’s existing operations. A decision could influence how the port structures future tenders, combines neighbouring assets and seeks private investment while maintaining competition among terminal operators.

Rupee slips to 95.74 as market pressures persist

The rupee weakened to around 95.74 against the US dollar on Monday as subdued domestic equities, elevated crude prices and uncertainty surrounding fresh US sanctions on Iran kept currency traders cautious.

The currency closed at about 95.7450 per dollar, extending pressure seen during the previous week while remaining confined to a comparatively narrow trading band. Persistent dollar demand from importers was partly offset by foreign currency inflows and intervention by the Reserve Bank of India, limiting the extent of depreciation.

The closing level was weaker than Friday's 95.6950, when the rupee had ended almost unchanged on the day but recorded a decline of roughly 0.3% for the week. The move on Monday kept the currency within sight of the psychologically important 96-per-dollar level, which traders have watched closely during bouts of volatility linked to Middle East tensions.

The Reserve Bank of India has remained active across currency markets, with dollar sales through state-run banks helping curb speculative pressure. Its interventions have contributed to unusually low short-term volatility despite substantial swings in crude oil and other global assets.

Two-week realised volatility in the rupee has fallen below 2%, while one-month implied volatility has eased to about 4%, below its year-to-date average of roughly 5.2%. Dealers expect the currency to remain broadly within a 95.50-96.50 range in the near term unless geopolitical developments trigger a sharper move in oil or the dollar.

Domestic equity weakness added to the cautious tone. The Sensex fell about 172 points to 77,369.11, while the Nifty 50 slipped below 24,250. Investors remained wary ahead of details of Washington's planned sanctions against Tehran, even as crude prices retreated from their latest highs during Monday's session.

Brent crude fell below $93 a barrel during trading after gaining more than 5% in the previous week. Prices had climbed as negotiations involving Washington and Tehran reached an impasse and uncertainty persisted over shipments through the Strait of Hormuz.

The pullback in oil offered some relief to the rupee but did little to remove the broader risk confronting the currency. India imports close to 90% of its crude requirements, leaving the economy sensitive to sustained increases in energy costs. Higher crude prices raise the country's import bill, increase demand for dollars from refiners and can place additional pressure on inflation.

Brent has remained around the low-$90s despite disruptions surrounding Hormuz, substantially above levels that prevailed before the escalation in Middle East tensions. Market forecasts have also shifted higher, with expectations that persistent supply constraints could push Brent towards $100 a barrel later this year.

Iran-related risks have become an important driver of currency and bond markets. Washington is preparing tougher economic sanctions aimed at increasing pressure on Tehran and potentially restricting its oil trade. Iran has warned that intensified economic pressure could have consequences for energy flows from the Gulf, adding uncertainty to a region that remains crucial to global petroleum supplies.

The dollar also strengthened modestly against a basket of major currencies, rising about 0.2% and adding pressure on several Asian currencies. Regional equity markets were mostly weaker as investors reduced exposure to risk ahead of the US sanctions announcement and key economic signals from the Federal Reserve.

Capital inflows have nevertheless provided the rupee with an important cushion. Measures introduced by the RBI to strengthen external liquidity have attracted tens of billions of dollars, boosting the central bank's capacity to manage disorderly currency movements.

The central bank's large foreign exchange reserves and continuing inflows have enabled it to resist abrupt declines, while importer hedging and month-end corporate dollar requirements remain sources of downward pressure. Its interventions have also prevented the rupee from moving decisively beyond 96 per dollar during periods of heavier selling.

Private operators set for 11-airport leasing round

The government has moved closer to leasing 11 Airports Authority of India airports to private operators after a key appraisal panel gave in-principle clearance to the plan, opening the way for one of the country’s largest airport monetisation rounds.

The airports will be offered in five bundles under the public-private partnership model, pairing larger, commercially stronger facilities with smaller airports that may require financial support. The proposed concessions are expected to run for 50 years, while ownership of the airports will remain with the Airports Authority of India, or AAI.

The five proposed groups comprise Amritsar and Kangra; Varanasi, Kushinagar and Gaya; Bhubaneswar and Hubballi; Raipur and Chhatrapati Sambhajinagar, formerly Aurangabad; and Tiruchirappalli and Tirupati.

The Public Private Partnership Appraisal Committee has granted in-principle approval for the transaction, marking an important step before the government proceeds towards detailed bid documentation and the competitive auction process.

Authorities are also examining restrictions on how many bundles a single company can win. The move follows concerns that further consolidation could increase concentration in the airport sector, where a relatively small number of private groups already handle a substantial share of passenger traffic.

The proposed safeguards could become one of the most closely watched features of the auction. Adani Airport Holdings and GMR Airports operate several of the country’s largest airports, while international infrastructure companies and investment groups are also expected to study the opportunity.

Earlier preparations for the transaction had attracted interest from operators including Adani, GMR, France-based Vinci Airports and the National Investment and Infrastructure Fund. The final field of bidders will become clear only after tender conditions and qualification requirements are issued.

The government intends to use the bundling mechanism to make smaller airports commercially viable by linking them with stronger traffic-generating facilities. Revenue from the larger airport in a package could effectively support investment and operating requirements at the smaller facility, reducing the burden on AAI.

Investment requirements associated with the 11 airports have been estimated at about ₹8,622 crore as capacity expansion and infrastructure upgrades are pursued. The precise capital expenditure obligations imposed on successful bidders will depend on the concession agreements and traffic projections incorporated into the tender documents.

AAI has also been targeting substantial proceeds from airport monetisation. Earlier estimates indicated that leasing the 11 airports could generate roughly ₹6,000 crore for the authority, giving it additional resources to build and upgrade aviation infrastructure elsewhere.

The model follows previous airport privatisations under which private concessionaires assumed responsibility for operations, management and development while the underlying assets remained publicly owned.

Delhi and Mumbai were transferred to PPP operators in 2006. A later auction resulted in Ahmedabad, Lucknow, Mangaluru, Jaipur, Guwahati and Thiruvananthapuram being awarded under 50-year concessions. Those six airports were won by Adani Enterprises through competitive bidding based on per-passenger fees payable to AAI.

The forthcoming transaction differs significantly because airports are being offered in clusters rather than exclusively as individual assets. The policy is designed partly to ensure that bidders cannot concentrate solely on profitable airports while leaving lower-traffic facilities dependent on public financing.

Varanasi is among the most commercially significant assets in the programme, serving a major tourism and pilgrimage centre and Prime Minister Narendra Modi’s parliamentary constituency. Tirupati also handles substantial pilgrimage-linked passenger traffic, while Amritsar is an important international and domestic gateway for Punjab.

Kushinagar and Gaya have strategic importance for Buddhist tourism, although their traffic volumes are considerably smaller. Kangra serves the Dharamshala region, while Hubballi provides connectivity to a major commercial centre in Karnataka.

The government has previously identified 25 AAI airports for monetisation as part of its broader infrastructure asset programme. Airport leasing is intended to unlock capital from operational public assets while bringing private investment into terminals, passenger services and commercial development.

The structure has nevertheless intensified debate over competition. Finance authorities have raised concerns about the possibility of an oligopolistic market if the same groups continue acquiring multiple airports, prompting proposals to limit the number of bundles that any one bidder may secure.

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.