Год: 2025

Quantum technologies in India are moving from academic promise to early deployment, driven by state funding, national missions, and rising startup activity.

As Karnataka commits ₹1,000 crore to its quantum mission and sets a 2035 target to become a “quantum economy”, attention is shifting beyond Bengaluru. The key question is whether the state’s tier-2 cities can participate meaningfully in this push, or whether quantum will remain concentrated in a few urban clusters.

The state has announced multiple initiatives, including an expansion of the quantum research park at IISc, indigenous hardware efforts, and partnerships between startups and universities outside the capital. However, readiness in tier-2 cities depends less on announcements and more on talent pipelines, infrastructure, and the ability to retain people once trained.

Training First Model

In Karnataka, this shift in quantum has been seen in Dharwad over the past year. In August, Bengaluru’s quantum startup QpiAI signed an MoU with IIIT Dharwad to launch the Q-Vidya eight-qubit quantum computer at the institute. The initiative aims to train researchers, provide hands-on learning to students and faculty and build academic capacity in quantum computing.

S R Mahadeva Prasanna, director at IIIT Dharwad, sees quantum computing following a familiar trajectory. “When any deep tech technology like this comes first, they will be in a few metro cities. Then, after some time, it will go into tier-2, tier-3 cities, just like the field of computing went,” he said in an exclusive interaction with AIM at the Bengaluru Tech Summit 2025.

What is different this time, he argues, is the pace. Prasanna stated that the rate at which this quantum is expanding is causing significant disruption in the field. He emphasised the need for a substantial workforce and human resources to manage this.

That urgency is shaping how IIIT Dharwad is positioning itself within Karnataka’s quantum roadmap. The institute plans to focus on capacity building before large-scale commercial activity arrives.

“We are going to train those who can go and teach the courses in their colleges,” he said. Alongside this, the institute aims to train undergraduate and postgraduate students from the region, before expanding statewide.

However, infrastructure remains a key constraint. Prasanna acknowledged that tier-2 cities still lag metros on social and professional factors.

However, he said, compared to tier-1, improvements are needed in schools, healthcare, and transportation infrastructure, which is now better but will take five more years to catch up fully.

For now, the institute is betting that opportunities can anchor talent locally. “When we create opportunities, people would like to stay back,” he said. Without that, graduates continue to migrate to Bengaluru, Pune, or Mumbai.

IIIT Dharwad hopes that training quantum manpower will eventually attract startups and industry labs to the region, reversing that flow.

Ecosystem Gaps Beyond Skills

While training momentum is evident, startups in quantum software and services are more cautious about ecosystem readiness.

“Software and services are growing and have huge potential. We are eager to deploy quantum computers and provide QCaaS and QVidya to tier-2 cities,” said Nagendra Nagaraja, founder and CEO of QpiAI, in a conversation.

Sujoy Chakravarthy, founder and CEO at Quanfluence, drew a clear distinction between education and deployment. “The maturity question is interesting,” he told AIM. In terms of universities and student training, I think the momentum is quite strong,” he said to AIM.

Chakravarthy pointed to heightened activity in smaller cities such as Chandigarh, Mandi, Kanpur, Kharagpur, Manipal, Trivandrum, and Goa. QpiAI also offers its Explorer training programme, which aims to train over 40,000 people in quantum readiness. “A lot of them are from tier-2 cities,” Nagaraja added.

However, Chakravarthy flagged structural gaps.

Chakravarthy expressed less confidence in the overall ecosystem’s preparedness, citing factors such as cross-functional talent’s willingness to relocate, the availability of capital, and ongoing infrastructure to support growth. He also noted a persistent global bias against smaller towns, which he considered to have inferior delivery capabilities.

For Quanfluence, the focus remains on training partnerships rather than physical expansion. “We don’t yet have plans to set up an office in these towns, but that could evolve over the next few months or years as the ecosystem matures,” Chakravarthy said.

This reflects a broader pattern. Even as Karnataka encourages activity beyond Bengaluru, most private investment, advanced labs, and early customers remain metro-centric. Without industry demand in tier-2 locations, trained talent often has little reason to stay.

What Quantum Readiness Actually Needs

From a large industry perspective, IBM Research India director Amith Singhee framed quantum readiness as a multidimensional challenge. Quantum computing is rapidly advancing. However, it remains in the early stages of large-scale application. Singhee identified four foundational pillars, including workforce and scaling. Besides, he also pointed out that strong R&D, industry involvement, and access to the best quantum technologies are crucial.

India has made progress on workforce exposure. IBM’s Qiskit platform and free-tier access have attracted significant participation from outside metro areas.

“The quantum advocacy program had more than 77,000 users from India,” Singhee said, adding that many were likely from tier-2 and tier-3 locations.

Yet deeper challenges remain. Singhee noted that over 90% of funding is government-sourced, which is insufficient. He emphasised that private sector involvement (e.g., finance, biotech, materials) is necessary to link quantum research to real-world applications.

Singhee also warned that research often struggles to move beyond laboratories. “We need this whole pathway from doing science to lab to market,” he said. Without test beds and collaboration structures, promising work risks staying confined to academic settings.

For tier-2 cities, he suggested hub-and-spoke models anchored by metro facilities. “Even if you set up a quantum park or a facility in a tier-1 place, you also design into that program a way for tier-2 in the region to participate,” he said, through internships, residencies, or shared infrastructure.

This approach is already visible in Karnataka. Whether it scales will depend on sustained funding and predictable policy support.

QpiAI is building infrastructure in non-metro cities. The company plans to put up data centres in Amaravati, Mangaluru, Vishakhapatnam, and the coastal areas of Maharashtra.

Meanwhile, it has acquired 10 acres in Devanahalli, rural Bangalore, to put together 100–1000 qubit machines. It would also have 15 application-specific labs connected to this massive quantum compute grid. “It’s the world’s largest quantum data centre, namely QpiAI QSC,” Nagaraja said.

As Karnataka pushes ahead with its quantum ambitions, it is not just about tier-2 cities training talent anymore. The real test is whether jobs, research opportunities, and industry confidence will follow. Without that, the state risks building a quantum workforce that still has to leave home to use its skills.

The post Is Karnataka’s Tier-2 Ready for Quantum? appeared first on Analytics India Magazine.

A public disagreement between AI researchers Yann LeCun and Demis Hassabis has reopened a long-running debate on whether human intelligence can be described as “general”.

In a recent podcast appearance, LeCun said the idea of general intelligence, when used to mean human-level intelligence, is flawed.

LeCun argues that “there is no such thing as general intelligence”, saying the term is largely used to describe human-level intelligence, which he believes is a mistake. Human intelligence, he said, is “super specialised”, shaped by evolution to handle the physical world and social interaction efficiently.

While humans navigate real-world environments and deal with other people well, LeCun pointed out that they perform poorly at many structured tasks, like chess, and are outperformed by other animals in several domains. This, he said, shows that humans are not broadly general but highly specialised.

“We think of ourselves as being general, but it’s simply an illusion because all of the problems that we can apprehend are the ones that we can think of,” LeCun said.

Hassabis responded that LeCun was conflating general intelligence with universal intelligence. “Brains are the most exquisite and complex phenomena we know of in the universe (so far), and they are in fact extremely general,” he wrote in his post on X.

He argued that while no system can escape the no free lunch theorem, a general system can still learn any computable function in principle. “In the Turing machine sense, the architecture of such a general system is capable of learning anything computable given enough time and memory,” he said, adding that human brains and AI foundation models are “approximate Turing machines”.

Hassabis also rejected the idea that human performance in narrow domains undermines generality. Referring to chess, he said it was notable that humans invented the game at all and reached elite levels of play.

LeCun later said the dispute was largely about terminology. “I object to the use of ‘general’ to designate ‘human level’ because humans are extremely specialised,” he wrote in his response.

He argued that intelligence should be judged not just by theoretical capability but by efficiency under limited resources. “For the vast majority of computational problems, [the human brain is] horribly inefficient,” he said, citing time and memory constraints in tasks such as chess.

To support his argument, LeCun used an analogy from deep learning, noting that while a simple neural network can approximate any function in theory, it becomes impractical for most real-world problems.

He also pointed to biological limits, arguing that the number of functions the human brain can represent is vanishingly small compared to the space of all possible functions. “Not only are we not general, we are [also] ridiculously specialised,” he said.

LeCun concluded by noting that humans mistake this specialisation for generality because most possible functions are incomprehensible. Quoting Albert Einstein, he wrote, “The most incomprehensible thing about the world is that the world is comprehensible.”

The post Yann LeCun, Demis Hassabis Clash Over What ‘General Intelligence’ Means appeared first on Analytics India Magazine.

When US-based global digital infrastructure company Vertiv launched its Training Academy and Technology Excellence Center in Pune last month, the move indicated the shortage of skilled service engineers at a time when infrastructure is expanding rapidly and becoming significantly more complex due to AI-driven workloads.

India’s operational data centre capacity has crossed 1.5 gigawatts (GW) and is projected to reach 4.5 GW or more by the end of the decade, supported by investments estimated at $20–25 billion. Much of this growth is being driven by hyperscale cloud adoption, data localisation requirements, and the early stages of large-scale AI deployment.

While capital expenditure and construction pipelines continue to accelerate, operators say the availability of skilled talent, particularly service engineers capable of managing high-density, AI-ready environments, has not kept pace.

Training for a new generation of infrastructure

Located at Vertiv’s Integrated Business Services (IBS) hub in Pune, the new training academy is designed to provide hands-on exposure to the systems that underpin modern data centres.

The facility features live demonstrations of cooling distribution units (CDUs), power switchgear systems, three-phase uninterruptible power supply (UPS) modules, and thermal management technologies, allowing trainees to work with production-grade equipment under simulated operating conditions.

“With the surge in AI investments in the data centre space, we expect a significant demand for a skilled, adapted, and updated workforce,” said Subhasis Majumdar, managing director, Vertiv India. He added that the academy intends to support engineers and customers working across both new and existing data centre environments.

The facility also includes a Technology Excellence Center with engineering and R&D laboratories focused on testing and validating power and thermal management solutions.

AI workloads raise the bar for service engineers

The shift from traditional enterprise IT to AI-heavy workloads has altered the operational profile of data centres. AI infrastructure typically involves GPU-dense racks, liquid cooling architectures, and power densities ranging from 100 kW to 300 kW per rack, compared to far lower levels in conventional facilities.

According to A S Rajgopal, managing director and CEO at NxtGen Cloud Technologies Private Limited, the most critical talent gap lies in managing this convergence of disciplines.

“The most critical gap today is at the intersection of high-density compute, power engineering, and AI infrastructure orchestration,” Rajgopal said. “Traditional data centre talent understands virtualisation, networking, and facilities, but AI workloads demand expertise in GPU fabric design, liquid cooling, ultra-high power racks, and AI workload scheduling.”

He added that there is also a growing shortage of professionals who can bridge physical infrastructure and AI platforms. “There is a sharp shortage of AI infrastructure architects and MLOps engineers who can bridge the gap between physical infrastructure and large-scale model training platforms,” he said, noting that the capability gap is widening faster than talent supply.

Converging physical infrastructure and digital intelligence

A similar assessment comes from Vipul Kumar, VP–edge and network at CtrlS Datacenters Limited, who said the industry is seeing a fundamental shift in the skills required to operate AI-ready facilities.

“The real challenge is building talent that can truly bridge physical infrastructure and digital intelligence,” Kumar said. “AI-ready data centres demand a shift in mindset from managing racks and uptime in isolation to understanding how compute density, thermal dynamics, energy efficiency, networking, and AI workload behaviour interact as one system.”

As AI workloads become more power- and heat-intensive, Kumar said, teams must take a holistic approach to performance optimisation, cooling strategies, energy usage, and workload orchestration. “This convergence of skills is now essential to delivering reliable, efficient, and scalable AI infrastructure,” he added.

CtrlS, he said, is building this capability through focused upskilling initiatives, cross-functional teams, and closer collaboration with technology partners to ensure its facilities are prepared for evolving AI infrastructure requirements.

Extensive retraining required

Despite India’s large pool of engineering graduates, operators say most new entrants are not immediately prepared for real-world data centre operations.

“Most engineering graduates are not job-ready for modern AI-era data centres,” Rajgopal said. Academic curricula, he noted, continue to focus on legacy IT systems and conventional cloud concepts. As a result, “companies end up spending six to 12 months, or more on retraining before graduates can handle real operational responsibility.”

Kumar echoed this view, saying that while graduates often have a strong theoretical foundation, they lack exposure to live environments. “They require immersion in real-world skills such as power systems, cooling technologies, network fabrics, risk management, security protocols, and process discipline,” he said.

This training burden has become more pronounced as data centres scale rapidly. Mumbai alone accounts for over 50% of India’s installed capacity, followed by Chennai, Delhi-NCR, and Bengaluru—regions where competition for experienced engineers is particularly intense.

According to David Yao, senior director at Vertiv’s Pune Hub, the facility combines experienced trainers with live systems and digital tools to support skill development across installation, maintenance, and continuous improvement in high-density environments.

Automation does not remove skill needs

While automation and AI-driven tools are increasingly deployed for monitoring, fault detection, and predictive maintenance, operators say these technologies do not eliminate the need for specialised talent.

“Automation and AI will reduce dependence on repetitive operational roles,” Rajgopal said. “However, this does not eliminate the talent crunch.”

Instead, demand is shifting toward advanced roles such as AIOps engineers, automation architects, cybersecurity specialists, and energy optimisation experts.

Kumar described the trend as a skills transition rather than a reduction in workforce requirements. “The talent requirement is evolving, not shrinking. The industry faces a skills transition, not a workforce reduction demand,” he said.

The post India’s Data Centre Boom Is Running Into a Talent Wall appeared first on Analytics India Magazine.

The Karnataka government is weighing steps to rein in the rapid expansion of computer science (CS) and allied engineering seats across the state, amid growing concerns that an unchecked supply of graduates in the stream could worsen unemployment among young engineers.

Speaking in the legislative council on December 16, Karnataka higher education minister M C Sudhakar acknowledged that engineering education in the state has become heavily skewed towards CS-related disciplines.

Karnataka currently has 229 engineering colleges, including 27 government engineering colleges. Together, they offer around 1.53 lakh engineering seats, the government said in response to a question raised by BJP MLC Dhananjaya Sarji.

A closer look at the distribution shows sharp imbalances, Sarji stressed.

While 27 private universities alone account for nearly 33,000 engineering seats, government engineering colleges together offer only 6,495 seats.

Even within these government institutions, only about 10% of seats are in CS-related branches, he flagged.

In contrast, Sarji highlighted that some private universities have concentrated a disproportionate number of seats in a single discipline.

In one such university, out of 4,320 engineering seats, as many as 4,020 are in computer science and related branches, he said, seeking a response from the government.

Sudhakar said the government agrees that such concentration is becoming a problem. He pointed out that while medical colleges operate under strict caps, with limits such as 250 seats per institution, engineering colleges currently do not have similar restrictions.

Referring to AICTE norms, Sudhakar said institutions have justified the expansion by citing student demand. The minister noted that concerns over excessive engineering seats are not new, recalling that the Telangana government had flagged such a trend.

Karnataka, he said, is now in the process of rationalising seat allocations, especially where a single discipline dominates intake numbers.

According to Sudhakar, out of the total 1.53 lakh engineering seats in the state, nearly one lakh are now in computer science and related fields. This, he warned, could lead to an unemployment crisis as “everyone is doing computer science”.

Using a stark analogy, the minister said the current system risks becoming one where “big fish eat small fish”, signalling that aggressive expansion by a few institutions could crowd out balance and long-term sustainability in engineering education.

Reacting to the announcement, Krishna Kumar Gowda, general secretary of Greater Bengaluru IT Companies & Industries Association (GBITCIA), acknowledged the government’s concern, saying a seat mix heavily skewed towards computer science is unsustainable for students and the broader economy.

He noted that Bengaluru and Karnataka remain leading global tech and GCC hubs, with strong future demand for computer science, AI, and data talent. Any cap, he said, must be data-driven, calibrated, and regularly reviewed to avoid future talent shortages or pushing students to other states.

“The issue is not that Karnataka has ‘too many’ computer engineers, but that CS seat growth has outpaced demand in some branches and placement capacity. Rationalisation that restores balance while protecting tech leadership is welcome,” Gowda said.

He stressed that simply reducing seats will not solve unemployment. Industry, he said, needs strong fundamentals, updated curricula, internships, and closer campus–industry collaboration to make graduates job-ready. Gowda also suggested structured consultation between the government, VTU, AICTE, and industry bodies, and a differential approach, capping low-quality programmes while allowing monitored growth in high-demand areas such as AI, robotics, and machine learning, including in tier-2 and tier-3 locations.

A few months ago, CPS Prakash, former principal of Dayananda Sagar College of Engineering, had issued a similar warning in a LinkedIn post, stating that many engineering colleges in Karnataka were adding computer science and allied branches without adequate planning. Most programmes lack trained faculty, feature outdated syllabi, and often duplicate the core CS curriculum, producing thousands of graduates with similar, misaligned skills, he had observed.

“The rapid adoption of AI is further reducing demand for human workers in tech, creating a looming mismatch between supply and demand,” Prakash wrote, adding that colleges may struggle to justify the proliferation of computer-related branches, risking high unemployment and a generation of graduates with obsolete skills, a crisis that administrators and regulators have failed to anticipate.

Neeti Sharma, CEO of TeamLease Digital, said the proposal to reduce computer science seats in Karnataka needs careful consideration. Technology, she said, is no longer limited to the IT sector, and computer science graduates are not hired only by IT companies.

While IT firms increasingly recruit from multiple engineering streams, Sharma noted that sectors such as manufacturing, BFSI, healthcare, retail, and logistics are undergoing digital transformation and actively hiring tech talent. The recent slowdown in IT hiring, she said, reflects greater selectivity rather than shrinking demand for technology skills.

“The real concern that needs to be addressed is employability, not the number of students,” Sharma said, adding that instead of cutting seats, the focus should be on improving curriculum quality, practical exposure, and industry relevance across streams.

Karnataka became India’s tech capital by building a strong and diverse talent pipeline, she said, cautioning that reducing computer science seats could weaken that advantage over time.

“The answer lies in better skills and outcomes, not fewer technology graduates.”

The head of the computer science department at a private university, on condition of anonymity, said the government’s proposal should not be seen merely as an overflow of computer science seats, but in the context of evolving global demand for technology skills.

“AI is now central across sectors, not just IT,” the HoD said, adding that opportunities exist in areas such as healthcare, construction, manufacturing and infrastructure through automation and robotics. Rather than cutting seats, they said policymakers should focus on blending AI and machine learning with core engineering disciplines to create broader employment avenues.

They also flagged the concentration of seats in a few private institutions, calling for stronger regulatory oversight.

“Approvals for intake increases and new programmes already require clearance from the Karnataka State Higher Education Council. The issue lies in enforcement,” he said.

Warning against arbitrary caps, the HoD said students determined to study AI and robotics would simply move to other states if opportunities shrink in Karnataka.

“That would hurt admissions and weaken the state’s talent pipeline. The focus should be on curriculum reform and quality, not seat reduction,” he added.

Viraj Singh Randhawa, a third-year engineering student at Manipal Institute of Technology, said he is not in favour of the proposed plan, arguing that demand for computer science remains strong.

He said imposing a cap on computer science seats should not extend to private institutions and, if implemented at all, be limited to government colleges.

Randhawa added that if the government moves to cap computer science intake, it should simultaneously increase seats in allied disciplines such as AI, data science and cybersecurity.

If expanding intake in these areas is not feasible, he said the state should consider increasing the number of engineering colleges, before imposing any cap on computer science seats.

The post Will Capping CS-Related Engg Seats in Karnataka Resolve Employment Concerns? appeared first on Analytics India Magazine.

As demand for large language models and real-time AI applications accelerates, the constraint is no longer just chips or capital, but electricity. Across the country, data centre operators are prioritising access to power over geography, snapping up any parcel of land that can support energy-hungry compute. Questions of connectivity and architecture are dealt with later.

This scramble is quietly reshaping how AI systems will be built and deployed. Instead of a few massive data centres clustered in predictable hubs, AI capacity is spreading across smaller, distributed locations, from modular facilities to telecom sites and enterprise edges.

This global push to scale AI has created what Shekar Ayyar, CEO of hyperscale networking software firm Arrcus, described as “a literal land grab.” Companies are acquiring any parcel of land that comes with assured electricity, and worrying about the rest later.

Ayyar said that these consolidated points do not have enough power to support all AI requirements. This scramble, he argued, is reshaping how AI infrastructure will be built. Instead of a few massive, centralised data centres, AI capacity is spreading across modular facilities, cell towers, and enterprise edges.

That shift is pushing networking, often treated as plumbing, back to the centre of the AI conversation.

Why Networking Lags Behind Compute

Ayyar traces the current inflection point to his years at VMware, where he saw computing move from fixed, purpose-built servers to virtualised, software-defined infrastructure. “In the network industry, that opportunity is happening right now,” he said.

“Networking is still about two, three decades behind where the compute transformation happened.”

Most enterprise and telecom networks still rely on rigid hardware boxes bought for specific functions from large vendors. That model, Ayyar argued, is incompatible with the pace of AI-led change.

“You cannot put a rigid infrastructure and hope that it will serve your purpose for the next decade or two,” he said, adding, “You have to have a software-driven infrastructure.” Arrcus positions itself as a horizontal software layer that can run across different types of networking hardware, allowing operators to adapt their networks as applications change.

Ayyar suggested that Arrcus can be applied across landscapes like a paintbrush to support all the needed applications, drawing an analogy to VMware’s early pitch.

The Innovator’s Dilemma

That approach also shapes how Arrcus sees competition. Incumbent vendors, Ayyar said, are constrained by their existing business models.

“It’s one of these innovator’s dilemma problems,” he said. “These are large, publicly traded companies. They just can’t go and say, we’re moving all that into a software-driven model architecture.”

While other startups are also working on disaggregated networks, Ayyar argued that Arrcus differentiates itself through the breadth of its production deployments, from low-end switching to high-end routing, across telecom, data centre, and enterprise environments.

The company also works across semiconductor ecosystems rather than tying itself to one.

Referring to partners such as Broadcom and NVIDIA, he said, “Our strength is that we work with both companies and frankly, other companies too.”

Opportunities in India & the IPO

The land grab narrative also ties into why Arrcus is doubling down on India. Historically, India served primarily as a technology development base. Ayyar believes that has changed.

“This is a very opportunistic time for anybody that has technology that will be instrumental in the creation, deployment and consumption of AI,” he said, adding that India is now both a producer and consumer of next-generation infrastructure.

He compared the moment to the early IT boom. “It is paramount for people to determine what’s the right skill set, and what’s the right talent pool, and what the right capability set is that needs to be built in a large population like India,” he said.

Arrcus plans to double its India headcount, expand beyond a single centre in Bengaluru, and push India towards contributing “at least sort of 10% of our total overall global book”.

The company currently works with 15–20 large customers globally and is shifting from a largely direct sales model to one driven by partnerships, including one with Fujitsu. Arrcus has formed a strategic partnership with Fujitsu to build next-generation network infrastructure for AI.

Its investor list includes Lightspeed, General Catalyst, SoftBank, Aramco, NVIDIA, Hitachi, Fujitsu and Samsung. “These are not just investors,” Ayyar said. “They should actually help us with go-to-market.”

That growth trajectory is setting the stage for a public listing, which Ayyar expects to be ready by 2027, he said.

The post This Firm Wants to be the ‘Next Big Disruptor’ in Networking appeared first on Analytics India Magazine.

Taiwanese electronics manufacturer Hon Hai Precision Industry (Foxconn) hired nearly 30,000 workers at its new iPhone assembly facility near Bengaluru over the last nine months, marking one of the fastest factory ramp-ups ever seen in India, according to a report by The Economic Times.

The move underscores Apple’s strategy to diversify its manufacturing footprint beyond China, with India emerging as a key hub.

Located in Devanahalli on a 300-acre campus, the facility began test production in April/May this year with iPhone 16 models and is now assembling the latest iPhone 17 Pro Max devices, the report said. More than 80% of the output is being exported, reinforcing India’s growing role in Apple’s global manufacturing network.

The plant’s workforce is notable for its demographic profile, with around 80% women, most of them first-time workers aged between 19 and 24. Foxconn has constructed six large dormitories to house employees, several of which are already operational, with more under development. At peak capacity next year, the facility is expected to employ up to 50,000 people.

With further expansion planned, the Devanahalli campus is projected to house more women workers at a single location than any other government or private establishment in the country. Employees have migrated from neighbouring states, and the site is expected to evolve into a mini township with residential, medical, educational, and recreational infrastructure.

Workers receive free accommodation, subsidised meals, and earn an average monthly salary of about ₹18,000—among the highest for women in blue-collar manufacturing roles.

Foxconn is investing close to ₹20,000 crore in the project, which is set to become India’s largest factory by employment and production capacity once fully operational. The plant is expected to eventually host up to a dozen iPhone assembly lines, compared with around four currently, and will surpass Foxconn’s existing iPhone facility in Tamil Nadu.

The expansion has been supported by India’s production-linked incentive (PLI) scheme for large-scale electronics manufacturing, launched in 2021, as Apple steadily shifts a larger share of iPhone manufacturing to India amid geopolitical uncertainties. All iPhone 17 models are now assembled in India and exported globally.

Apple’s India operations are backed by a supply chain of nearly 45 companies across component manufacturing, sub-assembly, and logistics. New recruits at the Devanahalli plant undergo six weeks of on-the-job training before joining production, as Apple and its partners work to build skills and deepen the local manufacturing ecosystem.While India is positioning itself as an all-rounder in manufacturing with campaigns like ‘Made In India’, it’s also projecting itself as an alternative to China. Bengaluru itself, also known as the Silicon Valley of India, has been using this to its advantage.

The post Foxconn hires 30,000 workers at Bengaluru iPhone plant in record ramp-up: Report appeared first on Analytics India Magazine.

The Karnataka government has moved to deepen its role in startup financing by encouraging state public sector enterprises (PSEs) to invest in government-backed venture capital funds, such as Karnataka Information Technology Venture Capital Fund (KITVEN). The government aims to unlock more patient capital for deep tech and frontier startups.

Announcing the move, Karnataka’s minister for electronics, IT/BT and biotechnology Priyank Kharge said the decision was aimed at strengthening institutional support for innovation, while expanding the pool of long-term capital available to early-stage companies.

“Karnataka government has taken an important step to deepen institutional support for startups and deep tech innovation,” Kharge said in a statement.

However, experts see the policy as an extension, rather than a fundamental shift in the state’s approach to startup funding.

Sankarshan Basu, professor of finance at IIM Bangalore, said the initiative does not amount to the state transforming its public enterprises into venture capitalists.

“The idea of Karnataka to encourage the state public sector enterprises to invest in government-backed venture capital fund – KITVEN – is probably not a direct shift from passive state support to active financial participation,” Basu said.

Focus on Innovation Growth Story

Kharge said that state PSEs would now be encouraged to invest in funds like KITVEN, which has been backing startups across IT, biotechnology, semiconductors, AVGC and other emerging sectors for over two decades.

According to the minister, KITVEN currently has a cumulative corpus of nearly ₹190 crore, with investments across 55 startups and exits that have delivered “strong returns.”

He highlighted dedicated vehicles, such as KITVEN Fund 5 and the Beyond Bengaluru Cluster Seed Fund, which are designed to channel capital into emerging startup hubs outside Bengaluru.

“This move will unlock long-term capital for early-stage companies working in deep tech and other frontier technologies, while also allowing profitable PSUs to participate in Karnataka’s innovation growth story,” Kharge said.

Prof. Basu, instead, described it as a complementary move, noting that “given the nature and scale of the investments required, no one channel will be adequate.”

‘One More Formal Avenue’

Basu added that public-sector participation in innovation funding was not new in principle.

“It has always been a vehicle of investment with adequate appetite and expertise to invest in new and innovative activities, making this one more formal avenue for the same,” he said.

From the venture capital community, views are cautiously optimistic, with emphasis on governance and operational autonomy.

Vishnu Das, principal at deep tech-focused venture firm Celesta, said access to PSE capital could strengthen KITVEN’s ability to back high-risk, early-stage companies, provided it comes without restrictive conditions.

“Having worked with KITVEN and seen how they operate, I think having access to PSE capital is a good thing,” Das said, adding that KITVEN has shown willingness to “take first cheque risks for companies especially in deep tech.”

However, he cautioned that the terms attached to such capital would be critical.

Safeguards Would be Essential

“It’s critical that these PSEs have reasonable expectations with respect to returns, exit timelines, etc. Moreover, they should not be involved in the investment decision-making, which should be left to the IC and investment team of KITVEN,” Das said, warning that excessive governance and administrative burdens could undermine early-stage investing speed.

Das also pointed to structural constraints within public-sector entities. “In my experience, PSEs have the ability, but not the risk appetite and decision-making speed,” he said, citing accountability pressures associated with deploying taxpayer capital.

Despite these challenges, Das said PSE participation could help crowd in private capital rather than distort early-stage markets. Beyond funding, he noted, PSEs could serve as large-scale test beds and early customers for deep tech startups—an advantage he described as “more valuable than capital.”

To make the model work, Das said safeguards would be essential, including minimal additional administrative burden, no overreach in investment committee processes, and limited constraints on which technologies or companies can be funded. He pointed to the Centre’s Research, Development and Innovation (RDI) framework as a reference for intent, if not a direct template.

As Karnataka seeks to scale its startup ecosystem beyond Bengaluru, and into capital-intensive sectors like semiconductors and deep tech, the success of this initiative may hinge less on the availability of public money and more on how independently and efficiently it is deployed.
Government data shows that the Karnataka Innovation and Technology Venture Fund’s latest vehicle, KITVEN Fund-5, is a SEBI-registered Category I Alternative Investment Fund with a target corpus of ₹100 crore, focused on backing startups in disruptive technologies such as artificial intelligence, machine learning, MedTech and electric vehicles.

Investments are typically made at initial cheque sizes of ₹2–3 crore per company, and the fund also prioritises ventures from tier-2/3 cities and women entrepreneurs.

Global Shift Towards Public-Private Capital

Responding to the government’s decision, Mir Amjad Husain, chief innovation officer at The National Institute of Engineering, Mysuru, said in his Linkedin post that the decision would catalyse long‑term growth and create valuable opportunities for the ecosystem.

Industry observers say that government-linked capital playing a role in venture capital is well established globally and in India, even if Karnataka’s specific PSU-to-VC model is locally novel.

In parts of Europe, government-backed entities like the European Investment Fund, Bpifrance and KfW Capital are among the most active limited partners in venture funds, boosting tech and deep tech ventures.

China’s state-owned VC firms such as Fortune Venture Capital and Shanghai Venture Capital Company have been investing commercially for decades. Beijing recently announced a national venture capital guidance fund to catalyse private investment in frontier technologies.

In India, central mechanisms such as SIDBI’s fund of funds for startups (FFS) have deployed over ₹9,400 crore by mid-2022 into alternative investment funds. Deep tech sector leaders have publicly called for stronger institutional LP participation to scale domestic venture capital.

These trends reflect a broad shift toward blended public-private capital models in innovation financing around the world.

The post Why Karnataka Should Tread with Caution While Encouraging PSUs to Fund Startups appeared first on Analytics India Magazine.

In 2025, AI became the defining force across technology sectors, driving new investments, reshaping business models and shifting how organisations build, deploy and scale digital systems. The impact was visible across Indian IT, global startups, Big Tech and global capability centres (GCCs), which recalibrated their plans to accelerate adoption.

Indian IT

This year marked a structural pivot for Indian IT, driven less by optimism and more by measurable AI adoption across enterprises. A joint EY-CII report showed that 47% of Indian enterprises now run multiple GenAI use-cases in production.

This surge changed the $264-billion Indian IT industry, pushing companies to update their services with more automation, cloud tools and enterprise AI.

TCS led the transition with a headline-making $6.5-billion commitment to AI-ready data centres, the first large-scale infrastructure bet by any Indian IT giant.

The IT company reached $1.5 billion in annualised revenue from AI-related services. CEO K Krithivasan told analysts at the company’s Analyst Day 2025 event that this shift from digital to AI is a “huge opportunity” and a “civilizational change” in how enterprises operate.

Infosys, Wipro and HCLTech followed suit by embedding GenAI across delivery lines, from modernisation and cybersecurity to industry-specific platforms.

Infosys secured a landmark $1.6 billion contract from the UK’s National Health Service Business Services Authority, signed in October. Another highlight was HCLTech’s partnership with OpenAI to drive enterprise AI.

Quarterly results showed that AI and cloud projects helped the big IT firms grow steadily, even though they hired fewer people and focused more on bringing in specialists with domain and AI skills.

AI Startups

According to a report by Second Talent, AI startups raised more than $89 billion this year, accounting for 34% of all venture capital, while a separate report by CB Insights shows that they are on track to secure more than half of total annual VC funding for the first time in 2025.

OpenAI announced in March that it raised $40 billion in a funding round, making it the largest private tech funding round ever. The company launched GPT 5.2 and a dedicated Sora app. Meanwhile, ChatGPT turned three in November.

Big funding rounds such as Anysphere’s $2.3 billion and Mistral’s €1.7 billion made it clear that investors now prefer companies building core infrastructure and developer tools instead of flashy AI ideas.

Anthropic added to the momentum with a massive $13 billion Series F round that valued the company at $183 billion. Databricks, the data and AI leader, closed a $4 billion Series L round. This move pushed its valuation past $134 billion.

Anthropic pushed deeper into coding-led AI with the launch of Claude Opus 4.5. Even early-stage players saw remarkable traction, with the Stockholm-based startup Lovable raising $330 million in a Series B funding round at a $6.6 billion valuation.

Big Tech

Big Tech spent this year in an all-out sprint to secure leadership in AI, announcing massive investments in both infrastructure and next-generation models.

Google expects full-year 2025 capital expenditures between $91 billion and $93 billion, Meta allocated between $64 billion and $72 billion, with plans for further increases, and Microsoft spent about $80 billion on AI cloud workloads and data centres.

Amazon led with an estimated $100 billion to $125 billion investment focused on AI capabilities and cloud infrastructure. The cloud giant announced plans to pour up to $50 billion into expanding AI and supercomputing systems for US government customers on AWS.

Meanwhile, Oracle struck one of the biggest cloud-computing deals on record, securing a $300 billion contract with OpenAI that will run for five years from 2027.

At the same time, India is witnessing a boom in data centres. Industry projections suggest the demand is real. India’s data centre demand is expected to surge from 1.3 GW in FY2025 to between 4.7 GW and 5.7 GW by FY2030, attracting a significant influx of investment from both domestic and global players.

Amazon has committed to investing over $35 billion in India by 2030, while Microsoft has pledged $17.5 billion over the next four years to expand its cloud and AI infrastructure in the country.

On the model front, Google pushed aggressively with Gemini, rolling out Gemini 2.0, 2.5 and eventually Gemini 3.0, alongside new AI-native tools such as Antigravity and deeper integration across Cloud and Workspace. The company spent the year clawing back dominance with a calmer, more mature AI strategy. This year, Nano Banana Pro was the company’s most popular image generation model.

Microsoft strengthened its lead in enterprise AI by creating a unified CoreAI engineering group and embedding AI across its cloud, developer stack and productivity platforms. The company also introduced its in-house speech model, MAI-Voice-1, and began public testing of its large language model MAI-1-preview, signalling its ambition to build specialised AI systems.

NVIDIA remained the sector’s growth engine as demand for its chips surged, boosting its valuation toward the $5-trillion mark. At the same time, AMD gained ground with new AI accelerators positioned to challenge its dominance.

GCCs

GCCs had a standout year as multinational companies increasingly turned their India hubs into centres of innovation rather than routine back-office operations. India now hosts more than 1,700 GCCs, with over 100 new centres added in the past two years, according to a Zinnov report.

The firms doubled down on engineering, AI, product development and data-led decision-making. A sharp shift towards AI was visible everywhere. According to a report by EY, about 58% of GCCs reported active investment in agentic AI, while more than 80% placed generative AI on their near-term roadmap.

Alongside IT and engineering, GCCs are now applying AI across customer service, finance, cybersecurity, operations and analytics, supported by dedicated innovation teams and internal AI centres of excellence.

Major GCC hubs include Bengaluru, Hyderabad, Pune, Chennai, Mumbai and the NCR. The industry is expected to grow to $105 billion by 2030, with nearly 2,400 centres employing more than 2.8 million people, further reinforcing India’s position as a leading destination for global enterprise operations.

Talent needs have shifted as well, with companies prioritising data science, cloud engineering, AI/ML and domain expertise over traditional back-office roles.

The post AI in 2025: A Complete Breakdown of Trends in Indian IT, Startups, GCCs and Big Tech appeared first on Analytics India Magazine.