The Commerce Ministry is identifying professions that are less vulnerable to artificial intelligence (AI) to prepare for the fast-changing technological landscape and tap opportunities for India to emerge as a global leader, Union Minister Piyush Goyal has said.

The Commerce and Industry Minister said that the focus would be on sectors where AI could improve efficiency without necessarily replacing human workers, while leveraging India’s large pool of skilled talent, youthful population and traditional strengths.

“I got my Ministry to start working on what are the AI-proof professions. After all, we have to plan as a nation for so many millions of people. What are the sectors where AI can at max add to efficiency, not necessarily replace (human workforce),” the Minister said on Friday (October 23, 2026) while addressing Indian science and technology researchers and students in Germany.

When asked about the objective of the initiative, he said India could take global leadership in several areas, including handicrafts, handloom, culinary arts and tourism, leveraging its vast pool of skilled talent and youthful population.

In these sectors, traditional skills and human creativity remain important, he said.

“… our handicraft, handloom… no amount of AI can replace that. So there are many, many areas where I think we have some very special strengths which we will be focusing on, also simultaneously developing, to ensure that 1.4 billion people have a better life in the future,” he added.

Asked whether the government planned to offer incentives to these sectors, Mr. Goyal said support would be provided wherever necessary, but incentives would not be required for every activity.

“It’ll be a mix of several initiatives, some privately led, some government led, some people led, some opportunity led,” he said, adding the broader objective was to develop India’s strengths and create opportunities that would help improve the lives of its 1.4 billion people.

Mr. Goyal also talked about exploring ways to attract Indian researchers, engineers, and scientists working abroad to return and contribute to the country’s development.

“I’ve asked them (the researchers and students here) to tell us what they will need to consider (the idea) and how we can support more, young researchers or our talented engineers and scientific community to come and work and serve in India,” he said.

Mr. Goyal said his discussions had highlighted that attracting talent was not merely a question of financial incentives, but also of creating a supportive environment for research and innovation.

“And surprisingly, I realise it’s not just about money; it’s about the enabling environment, the support of the ecosystem. So I (will) talk to my colleagues in the government, with educational institutions, both in the government and private sector, with academia, with leading research institutions in India, and industry, to see how we can forge industry, academia, and researchers partnerships,” he said.

Referring to future global growth prospects, Mr. Goyal said developed countries would increasingly need to engage with emerging economies in Asia, Africa and Latin America, as these regions would drive future growth.

There are about 3,30,000 people of Indian origin in Germany, plus nearly 60-70,000 students.

“This large Indian student community, full of enthusiasm, seems to be very excited about taking their learnings back to India. Now we can see them not just looking for a job in Germany, but wanting to be back home to contribute to the India growth story, because that is where the action is,” Mr. Goyal told reporters.

The Minister was here to discuss ways with government and private sector leaders to boost trade and investment between India and Germany.

Published – October 24, 2026 12:36 pm IST

In July, the North Eastern Regional Power Committee (NERPC) convened a special meeting with grid operators and Power Grid Corporation of India Ltd. (PGCIL) to address an urgent issue: maintenance contracts for the region’s integrated electricity grid were nearing expiry, necessitating discussions on upgrading the systems.

The financial implications were significant. PGCIL informed the meeting that the estimated cost of upgrading the Supervisory Control and Data Acquisition/Energy Management System (SCADA/EMS) at the region’s State Load Despatch Centres had risen from ₹521 crore to about ₹688 crore. The reason: AI-driven demand was making hardware used in supervisory systems more expensive.

Escalating costs are not new to the power sector. As manufacturers phase out older components and specialists capable of maintaining legacy systems become harder to find, replacement and maintenance costs can rise sharply. However, the rapid expansion of data-centre projects is adding a new layer of pressure on power generation and transmission infrastructure.

Ramping concerns

NERPC is not the only regional body grappling with the implications of AI-driven electricity demand.

At a meeting of the Eastern Regional Power Committee’s Technical Coordination Committee last December, an Odisha representative flagged the potential “ramping impact” on States when industries such as data centres undertake “bulk collective purchase” of electricity.

Such fluctuations can strain slow-ramping generators, which cannot adjust output quickly enough to match sudden changes in demand. This could expose generators to penalties under the Deviation Settlement Mechanism (DSM) for deviations beyond their control.

AI data centres consume substantially more power than conventional facilities. Their demand patterns can also fluctuate sharply, depending on usage, especially LLMs, server distribution and the time zones of their customers.

The issue reached the Central Electricity Authority’s National Power Committee in February. Minutes of its meeting noted that “data centres typically exhibit high constant loads, with possible ramping characteristics depending on global time zones and server distribution”.

A committee constituted by Grid-India was tasked with studying data-centre electricity demand patterns, media reports suggest.

Alok Kumar, Director-General of the All-India Discoms Association (AIDA), told The Hindu that the incremental demand from data centres would remain manageable.

“An incremental demand from data centres ranging between one and two gigawatts per year will not make a huge difference,” he said, adding that “[moreover], this would be concentrated in certain cities only, like in Hyderabad, Mumbai, Chennai, Visakhapatnam and the Delhi NCR region, among others.”

However, the final leg of electricity transmission could pose a challenge. State transmission utilities would need to expand capacity quickly in locations witnessing a concentration of data-centre projects.

“State transmission utilities would have to quickly upgrade where there is demand and within time. It could be a challenge in certain urban areas because of right of way (RoW)-related issues,” Mr. Kumar said.

Nationwide deliberations

Concerns over data-centre loads and their implications for grid stability have surfaced across regional power committees.

At meetings of the Northern Regional Power Committee’s Operation Coordination Sub-Committee last year, officials repeatedly discussed monitoring data centres’ compliance with grid-connection requirements before supplying them with electricity.

In an email in February, the Northern Regional Load Despatch Centre, which oversees the region’s grid, warned that data centres could switch to backup power systems rapidly, creating a “credible risk of abrupt load rejection”.

Documents showed that Noida had 125 MW of operational data-centre capacity, with more than four times that capacity in the pipeline.

The Southern Regional Power Committee, meanwhile, has been preparing for a surge in demand from data-centre projects in Visakhapatnam (Google was not named in the discussions, it is the largest planned project in the area). Officials discussed the prospect of gigawatt-scale demand emerging in the region over the coming months.

At an April meeting, the SRPC’s Protection Coordination Sub-Committee examined the implications of data centres for transmission infrastructure. An official noted in the opening remarks that “inputs were sought from Grid-India to assess reserve requirements and system planning implications”.

Who pays for upgrades?

Mr. Kumar said the cost of upgrades should not become a broader financial burden, as transmission investments would ultimately be recovered through transmission charges.

However, he emphasised that data-centre operators must bear the costs of additional investments required to extend transmission infrastructure to their facilities. Such expenditure, he argued, should not be passed on to all consumers of the same distribution company.

Mr. Kumar pointed to Central Electricity Regulatory Commission (CERC) regulations that mandate this principle.

Line faults expose vulnerabilities

Maharashtra, home to a large share of India’s data centres, has long benefited from demand generated by the banking and financial services industry.

Yet the concentration of these facilities is also increasing the importance of grid reliability. Minutes of a Western Regional Power Committee meeting last year recorded discussions on a grid incident on June 12, when a faulty bus caused two interconnecting transformers to trip, resulting in the shedding of about 32 MW of data-centre load.

At a meeting in January, an official also warned that the proposed splitting (“line-in line-out”), of the Talegaon-Kharghar transmission line could leave the already stressed corridor vulnerable to severe overloading if another line tripped. The corridor serves an area experiencing growing data-centre development around Mumbai.

The official, from Maharashtra State Electricity Distribution Company Limited, suggested upgrading the line to 2,100 MVA.

Published – October 10, 2026 07:11 pm IST

Image used for representational purposes only.
| Photo Credit: Getty Images/iStockphoto

The story so far: Wipro recently changed how long some of its employees can remain without a billable project. Under its revised bench policy, effective October 1, employees in Band 2 and below can reportedly remain unallocated for up to 180 calendar days. For Band 3 and above, the standard period is 45 days, while employees with disabilities have a 75-day period. Employees in Band 2 and below who remain unallocated for 180 days can be separated under their employment contracts. Wipro did not respond to queries from The Hindu.

The change comes as recent industry estimates point to shorter bench periods and lower bench strength. Average bench time at large IT-services companies was estimated at 35–45 days in 2025, compared with 45–60 days in FY20 and FY21. The proportion of employees on the bench was estimated at 2–5%, down from 10–15% earlier.

So what exactly is the bench, and why has it been such an important part of the Indian IT business model?

What is the bench?

In IT services, the bench is broadly the pool of employees on a company’s payroll who are not currently assigned to a billable client project. They may be waiting for another assignment, undergoing training or working on internal initiatives. The exact definition can vary between companies.

The bench exists because companies have to balance workforce capacity against uneven project demand. A company that wins a large project needs people ready to work on it, but hiring, training and demand do not always happen at the same time. Carrying some spare capacity allows faster deployment when demand materialises, but a company continues to incur employee costs while those employees are not generating billable revenue.

That is why utilisation matters. A higher proportion of employees who are not deployed on billable work can weigh on utilisation and margins. But being on the bench does not necessarily mean being unproductive. It could mean employees may be training, working on internal projects or developing skills for another deployment.

When did the bench become part of Indian IT?

There is little evidence of a single point at which the Indian IT industry introduced a formal bench model. Company records show that unallocated capacity was already being measured and discussed by the early 2000s.

In April 2001, Infosys told analysts that the company had “slightly more than 1,000 people on the bench” at the end of the quarter. Utilisation excluding training was 73%; the remaining unbilled effort included employees on the bench as well as those on leave, management, research and development and internal information-systems projects.

Such disclosures show that unallocated capacity was already an identifiable part of workforce management at a major Indian IT-services company by the early 2000s.

Why did companies carry a bench?

A large bench was not always regarded as waste. When companies expected demand to rise, retaining trained employees allowed them to respond to new projects without having to recruit and train an entirely new workforce.

Wipro’s FY2011-12 annual report points to its investment in a “strategic bench to fuel growth”, which contributed to gross utilisation falling from 69.9% in FY11 to 68.3% in FY12.

The bench could therefore be both an investment and a cost, where companies could deliberately carry employees ahead of expected demand, accepting lower utilisation in anticipation of future projects.

Why did just-in-time hiring emerge?

The economics of carrying a bench changed when demand weakened. Spare capacity that could be deployed quickly during a growth phase became a cost when projects were delayed or failed to materialise.

Wipro’s October 2009 earnings call illustrates the shift. Then CEO Suresh Vaswani said the company’s bench had fallen from around 11,000 employees to about 7,000 and that Wipro had made its “people supply chain” tighter and more “just-in-time” in hiring and recruitment.

HCL Technologies took a similar approach during the 2008-09 downturn. Then CEO Vineet Nayar said the company had adopted “just in time hiring”, arguing that people could be hired at seven days’ notice and that there was therefore “no point of increasing your bench.”

The history is therefore not a simple journey from large benches to small ones. Companies have moved between carrying spare capacity and tightly controlling it depending on demand and utilisation.

How did companies make the bench productive?

Some companies tried to change what happened to employees between projects rather than simply reduce their numbers.

Persistent Systems said in its FY2016-17 annual report that the need to leverage employees when they were not working on customer projects had led to the “dissolution of the Bench” and the creation of TeXT, or Transforming Employee Experiences. Employees between projects could work on timed projects involving AI, machine learning, natural language processing, IoT and analytics.

Infosys launched its Zero Bench programme in 2015, creating an internal marketplace for employees awaiting longer-term assignments. In its FY2015-16 annual report, Infosys said the programme was intended to “eradicate the bench”. Within nine months, more than 12,000 jobs had been created on the marketplace, and more than 67% of employees who were on the bench had completed at least one assignment.

Eliminating the bench as an idle period is not necessarily the same as eliminating employees being between client projects. The interval can instead be filled with internal work, training or other assignments.

What is AI changing?

AI is changing what makes an employee deployable. IT-services companies are increasingly linking workforce planning to skills and reskilling. TCS’s FY2025-26 annual report says its talent model is centred on skills, roles and talent, and reports more than 2.7 lakh employees with advanced AI and machine-learning proficiency.

Wipro’s FY2025-26 filing describes a “Skill as a Currency” approach involving upskilling, reskilling and cross-skilling, while Infosys’s FY2025-26 ESG report says its AI strategy includes enabling employees to be deployed to relevant projects through upskilling.

The bench is therefore increasingly a skills-matching problem as well as a headcount problem. An employee can be available but not immediately deployable if their skills do not match the projects entering the pipeline.

Therefore, AI is not simply shrinking the bench. It is changing the definition of who is deployable. 

Who bears the risk?

A bench is a buffer against uncertain demand, but it involves a trade-off for companies. Maintaining one means carrying the cost of employees without billable assignments, while maintaining too little means risking a shortage of trained talent when demand picks up.

For employees, a shorter bench window leaves less time to find another assignment, and prolonged non-deployment can eventually put the employee’s job at risk. Skills mismatch can add to that risk, as an employee may be available but not suitable for the projects in the pipeline.

Therefore, the bench becomes a balancing act for both sides, with companies managing capacity and employees managing deployment risk.

Published – October 03, 2026 03:17 am IST

OpenAI CEO Sam Altman on Tuesday (September 29, 2026) unveiled GPT-6.1 Sol at the company’s DevDay conference in San Francisco, positioning it as a lower-cost model that delivers “very near-Astra-level intelligence at a fifth of the price” for coding and building. File
| Photo Credit: Reuters

OpenAI CEO Sam Altman on Tuesday (September 29, 2026) unveiled GPT-6.1 Sol at the company’s DevDay conference in San Francisco, positioning it as a lower-cost model that delivers “very near-Astra-level intelligence at a fifth of the price” for coding and building. File
| Photo Credit: Reuters

Representational image
| Photo Credit: Getty Images

The proposal from Beijing comes against the backdrop of China’s push to offer its AI platforms and smart manufacturing know-how to emerging countries and its trade and technology war with the U.S.
| Photo Credit: AFP

Anthropic said it had banned the relevant accounts as soon as it had detected the illegal activity. File
| Photo Credit: Reuters