One argument often used to quell concerns about the rising energy and resource demand of data centres is that artificial intelligence (AI) models will need less in the future as they improve and become more efficient.
But this seemingly logical thinking is a trap, according to a new United Nations report that quantifies the environmental costs of AI.
The report estimates that by 2030, AI’s energy use could double to consume 3 per cent of the world’s electricity, produce emissions to equal the UK, and deplete more water for cooling than the annual drinking water needs of the global population.
It also anticipates that the use of AI will follow an economic principle known as the “Jevons paradox”, which predicts that when technological improvements increase the efficiency of a resource, it leads to a rise, rather than a fall, in the total consumption of that resource.
The paradox is named after economist William Stanley Jevons, who observed this effect with the use of coal in 19th-century England. Efficiency gains did not reduce overall consumption. Instead, the lower costs resulted in expanded use and higher overall demand.
As AI models become cheaper and more attractive, the report expects this to encourage new uses and higher volumes of use, eroding and possibly erasing any savings from efficiency advances.
To avoid falling into this trap, it lays out a roadmap for responsible AI use based on guiding principles of transparency, efficiency by design, equity and justice, lifecycle responsibility, global cooperation, and sustainable use.
Last year, data centres already consumed as much electricity as Saudi Arabia, which ranks as the world’s 11th largest electricity consumer.
If electricity use doubles as projected by 2030, the associated carbon footprint would require 6.7 billion trees grown over ten years to offset this demand.
Data centres would also require 9.3 trillion litres of water and land nearly ten times the size of Mexico City.
Beyond resource use, the report also underscores the structural inequity at the heart of the AI boom, with only 32 nations hosting AI-specific cloud infrastructure and 90 per cent of that capacity located in the US and China.
It warns of a widening digital divide between nations that build and control AI systems and those that consume them, with the latter often bearing a disproportionate environmental burden caused by mineral extraction and e-waste.
Two main forces shape AI’s operational footprint: how much we use it and how we use it.
This involves all tasks AI models perform, from text and code generation to image and video. Each of these tasks requires different levels of computational effort.
The model choice also matters as each AI system performs these tasks with distinct energy and environmental costs.
The report argues responsible AI requires full value-chain governance, from mineral sourcing to recycling and safe disposal.
It calls for a twinning of capability and environmental stewardship – thinking about both what AI can do for us and the protection of the natural environment.
This would mean making environmental disclosures a routine part of AI development, at both the model and task level, and incorporating projected AI demand in climate and energy planning.
Responsible AI is crucial as countries are promoting and adopting AI across government and the public sector.
In Aotearoa New Zealand, the government has launched a national AI strategy and a public service AI framework.
While the framework was informed by the OECD’s values-based AI principles, including inclusive and sustainable development, there is no requirement for environmental disclosures and no regulator compiling energy use or emissions.
Likewise in Australia, improving public services is part of the national AI plan. For example, the National Film and Sound Archive of Australia has created Bowerbird, a machine learning-enabled mass audio and video transcription engine, to document material. The Department of Veteran’s Affairs has developed a proof-of-concept tool to see whether AI can help speed up the processing of claims.
Both countries take a deliberate “light touch” and principles-based regulatory approach to AI. But this approach risks overlooking the growing environmental cost of AI that can’t be solved by improving it.
The natural environment is foundational to the economy, culture and wellbeing. It should be at the centre of our thinking. It’s time to rethink the AI innovation playbook and shift focus toward a sustainable tech future.
This article is republished from The Conversation under a Creative Commons license. Read the original article.
By Asmi Bhatia and Alex Gabriel Simon
The overarching narrative is that Indian equity markets missed out on the global artificial intelligence boom. But a look under the hood reveals a slew of smaller firms winning from trillions of dollars being spent on AI capacity.
The poster child for this rally is Sterlite Technologies Ltd., the optical-fiber maker owned by the Vedanta Group, which has surged more than 530 per cent this year. It got a $1.1 billion multi-year contract from a US-based hyperscaler last month. Its competitor, HFCL Ltd., has jumped 191 per cent while MTAR Technologies Ltd., which makes precision cooling and power components, has more than tripled.
Since every AI query runs through power-hungry data centers, which require immense electricity and cooling, old-economy industrial firms have transformed into India’s hottest market play. In Mumbai dealing rooms, it’s called the ‘AI capex trade.’
“We may be on the wrong end of the AI trade, but we could be on the right side of the AI capex trade,” said R. Sivakumar, chief investment officer at Axis Mutual Fund. “One could consider companies benefiting from data centers and the entire value chain associated with this capex.”
Great to meet @gautam_adani in Ahmedabad this morning and build on our existing partnership with the Adani group. As India fast emerges as a leading innovation hub for @Uber, we are setting up our first data center in the country with the Adani Group to test and deploy our tech.… pic.twitter.com/jll5NcPT4l
— dara khosrowshahi (@dkhos) May 13, 2026
A Reliance Industries Ltd. joint venture signed an $11 billion pact to build local data centers last year, while AdaniConnex Pvt. has partnerships with Google as well as Uber Technologies Inc. to help build their data centers.
‘Picks and Shovels’
“The most attractive exposure is in the industrial supply chain — the ‘picks and shovels’ that build, power, and cool these facilities,” Nomura Holdings Inc. analysts led by Akash Gupta wrote in a June 2 report.
Also, a two-to-four year lead time in supplying some components has “created an enviable seller’s market with multi-year backlogs,” Nomura analysts wrote, adding that orders secured now will bring revenue between 2027 and 2029.
Foreign investors are already piling in. Shareholding of foreign funds in industrials rose to 14 per cent as of end-March, the highest in two years, according to Elara Capital (India) Pvt., even as global funds remain record sellers of Indian stocks.
On a top-down basis, India is one of the worst-performing markets globally as it lacks pure-play AI firms and semiconductor makers that are turbocharging Taiwanese and South Korean equities. But the global obsession with generative AI is boosting those that keep these hyperscalers running, such as Hitachi Energy India Ltd., ABB India Ltd. and Cummins India Ltd.
“The rally in companies like Sterlite and MTAR is driven by the market’s growing conviction that AI is creating a multi-year infrastructure capex cycle, not just a software opportunity,” according to Angel One.
Mahesh Viswanathan, chief executive officer of Finolex Cables Ltd. said in an earnings call last month that this was “the right time to be in this industry.” Finolex’s have surged nearly 36 per cent this year.
The market is rewarding companies with visible AI-linked earnings rather than just thematic exposure, according to Angel One. Also, the biggest near-term risk is valuation as share rallies have left “no room for execution disappointments,” the brokerage added.
For instance, Anant Raj Ltd., the only listed pure-play data center firm, has gained just about 8 per cent this year. Meanwhile, Sterlite is trading at about 70 times its 12-month forward earnings, compared to NSE 500’s 19 times.
But no market watcher is downplaying this opportunity.
“Data center capex has emerged as the single largest contemporary industrial investment cycle,” Nomura analysts wrote. It’s “larger than the global wireless 4G roll out, the post-2008 LNG build-out, or the early-2010s shale boom.”
At Google I/O 2026, the company introduced a significant change to Search. It unveiled a new AI-powered search experience that allows users to search using text, images, files, videos and even Chrome tabs. While Search continues to display traditional web results, Google now prioritises AI-powered answers that appear before conventional search listings. As a result, traditional web results have been pushed further down the page, reducing their visibility.
The move triggered a backlash, with some users shifting to alternative search engines such as DuckDuckGo. According to TechCrunch, critics argued that AI-powered search could undermine the open web, while others raised concerns about inaccuracies in AI-generated responses. There was also a section of users who simply did not want AI integrated into their search experience.
DuckDuckGo, which allows users to opt out of AI-powered search features entirely, saw app installs in the US rise 18.1 per cent week-on-week between May 20 and May 25, according to the report. The growth continued for six consecutive days and peaked at 30.5 per cent on May 25.
The search engine’s AI-free search page, which disables AI-generated answers and images by default, recorded average week-on-week growth of 22.7 per cent and peaked at 27.7 per cent on May 24.
Against this backdrop, the UK’s Competition and Markets Authority (CMA) instructed Google on June 3 to allow publishers to opt out of having their content used in AI-powered Search features if they choose.
In response, Google announced it is testing a new Search Console control in the UK. The feature will allow publishers to prevent their content from appearing in AI-powered Search experiences such as AI Overviews, AI Mode and AI-generated Discover results, while continuing to remain visible in traditional Google Search results.
The company said opting out will not affect search rankings. Google is also introducing new reporting tools that will give publishers greater visibility into how their content is being used within AI-powered Search experiences.
The move comes at a time when publishers around the world are grappling with declining search referrals, falling page views and growing concerns that AI-generated answers are consuming the very content on which they depend.
Before examining the significance of Google’s latest move, it is important to understand how search evolved from a discovery engine into an answer engine.
Traditional search engines were built around discovery.
A user searching for “best budget smartphone”, “how to file taxes” or “what caused inflation” would receive a list of links. Search engines helped users find information, but the actual consumption of that information happened on individual websites.
Products such as Google’s AI Overviews and AI Mode change that dynamic. They synthesise information from multiple sources and present it as a single answer directly within the search interface.
Instead of browsing several websites, users increasingly receive what appears to be a complete response immediately.
This has transformed search from a navigation tool into an answer engine.
Google’s own figures illustrate how quickly users have adopted the format. The company says AI Overviews now reaches more than 2.5 billion monthly users, while AI Mode has crossed one billion monthly users.
The appeal is obvious. Users save time, avoid opening multiple tabs and receive information in a conversational format. For routine informational queries, AI-generated summaries often satisfy user intent without requiring additional research.
But convenience for users has created a growing challenge for publishers.
The shift has given rise to what many publishers describe as the “zero-click internet” — a web where information is consumed but the original source is often bypassed.
Publishers’ concerns are not simply that AI summarises information. Search engines have displayed snippets for years.
The concern is that AI-generated answers are now comprehensive enough that users never leave the search page to visit the original source.
As a result, websites do not receive traditional page views or traffic unless users specifically click through to the source.
Researchers studying the impact of Google’s AI Overviews have already found evidence that this shift is reducing visits to source websites.
One study, titled “Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia”, estimated that exposure to AI Overviews reduced Wikipedia traffic by around 15 per cent on average.
According to Similarweb, zero-click searches now account for roughly 60 per cent of all Google queries. For news-related searches, that figure reportedly rose to 69 per cent following the launch of AI Overviews.
A report by The Next Web stated that UK publisher DMG Media recorded traffic declines of up to 89 per cent for certain queries.
Another study, titled “Measuring Google AI Overviews: Activation, Source Quality, Claim Fidelity, and Publisher Impact”, found that more than half the pages cited in AI Overviews contained display advertising. This suggests publishers may lose advertising revenue when users consume information within Google’s interface rather than visiting the original source.
The issue extends beyond traffic.
At stake is the business model that sustains much of the web.
Most publishers, particularly independent ones, depend on advertising revenue generated when readers visit their websites. If AI-generated answers increasingly satisfy users without requiring a click-through, publishers lose both readership and revenue opportunities.
Meanwhile, Google continues to monetise the search experience through advertising displayed alongside AI-generated responses.
The impact of AI Search would be easier for publishers to absorb if traffic could shift elsewhere.
The problem is that Google’s dominance leaves few realistic alternatives.
The debate surrounding AI-powered search would look very different if users had multiple search engines of similar scale. In reality, they do not.
Google remains the internet’s primary gateway to information.
According to Sensor Tower’s State of Web 2026 report, Google was the world’s most visited website in 2025, recording approximately 240 billion visits. No other search engine featured among the 20 most visited websites globally.
The picture is similar in the UK and India.
According to a report by Republic World, UK regulators noted that Google accounts for more than 90 per cent of general search queries in the country.
That dominance was one of the main reasons the CMA intervened and required Google to provide publishers with greater control over AI Search usage.
Historically, publishers have faced an uncomfortable reality. If they wanted visibility in Google Search, they generally had to accept Google’s terms regarding indexing and content presentation.
The arrival of AI Overviews intensified that tension.
A Wall Street Journal report noted that while content from news publishers was helping power AI-generated answers, the need for users to visit the original websites was falling significantly.
The significance of Google’s announcement is not that publishers will suddenly abandon AI Search. Many may not.
Instead, the importance lies in restoring leverage.
Until now, publishers effectively faced an all-or-nothing choice. They could allow their content to be used in AI-powered Search experiences or risk losing visibility within Google’s ecosystem.
The new controls separate those decisions.
Publishers can remain discoverable through traditional Search while gaining greater control over how their content is used in AI-generated experiences.
However, if publishers want to return to a model that relies heavily on traffic-driven advertising revenue, a substantial portion of the industry may need to opt out collectively.
The CMA said the measure is intended to strengthen publishers’ bargaining position and provide greater control over content usage.
This could create several outcomes:
In other words, the announcement may not immediately restore lost traffic, but it does represent a step towards rebalancing the relationship between platforms and publishers.
Google is not the only company building answer-first search experiences.
Microsoft Bing Copilot, Perplexity and several emerging AI search platforms also rely heavily on web content to generate responses.
If Google expands this feature globally after testing, it could establish a precedent and accelerate broader industry standardisation.
If regulators begin treating publisher control as a baseline requirement, competing AI search providers may face pressure to offer similar mechanisms.
Even if publisher opt-outs become widespread, they are unlikely to reverse the broader transformation taking place across the web.
Search behaviour itself is changing.
Users increasingly expect direct answers rather than lists of links. AI systems are becoming better at synthesising information, handling follow-up questions and delivering personalised responses.
AI-generated search experiences are appearing for a growing share of queries, particularly informational searches where a concise answer can satisfy intent.
That means publishers may need to rethink long-standing assumptions about audience acquisition.
For years, success depended on ranking highly in search results. In the AI era, visibility alone may no longer guarantee traffic.
Instead, publishers may increasingly focus on building direct relationships with audiences through subscriptions, newsletters, apps, memberships and communities.
Google’s new opt-out control does not solve the fundamental tension between AI convenience and publisher economics.
Users are unlikely to abandon the speed and convenience of AI-generated answers, while publishers remain dependent on traffic and engagement to sustain their businesses.
What the new policy does provide is a degree of control that many publishers argue they lost as AI-powered Search expanded.
Whether that translates into better traffic, stronger negotiating power or new commercial arrangements remains to be seen.
But for an industry grappling with the rise of the zero-click web, it represents one of the first meaningful attempts to restore balance between the platforms that distribute information and the organisations that create it.
Apple’s antitrust case in India has reached a stage where the outcome may extend far beyond financial penalties, potentially reshaping how the iPhone ecosystem operates in the country.
The case began in 2021 after a complaint by non-profit Together We Fight Society, and later drew support from industry bodies such as the Alliance of Digital India Foundation and companies including Match Group. It focuses on Apple’s App Store practices, particularly whether its rules around app distribution and in-app payments restrict competition and limit market access for developers.
The case has now moved into a more decisive phase. After months of delays and legal challenges over the scope of financial disclosures, Apple has agreed to submit data related to its India business, allowing the regulator to assess its economic footprint and determine potential penalties.
While the possibility of a large fine has drawn attention, the deeper issue lies in whether the CCI will impose structural remedies that alter Apple’s control over app distribution and payments.
The allegations against Apple centre on the design of its iOS ecosystem. Developers have argued that Apple imposes unfair conditions by mandating the use of its App Store for distribution and its in-app purchase system for digital transactions, effectively tying the two together. Restrictions on steering users to alternative payment methods or offering competing billing systems have also been flagged as limiting competition.
The regulator has also examined specific provisions within Apple’s App Store guidelines that go beyond simple platform control. These include rules that prohibit developers from directing users to alternative payment mechanisms within apps, effectively blocking links or prompts that could route transactions outside Apple’s system.
In addition, the requirement that developers use Apple’s in-app purchase system for digital goods ties app distribution to Apple’s own payment infrastructure. The Commission has noted that this combination of restrictions limits the ability of developers to choose competing payment providers and may amount to a tying arrangement that restricts competition.
The CCI’s probe has examined whether this amounts to abuse of dominance by denying market access to competing payment processors and limiting the ability of developers to operate freely within the ecosystem.
The Commission has also departed from Apple’s argument that the relevant market should be defined as the broader smartphone segment, where its share remains limited. Instead, it has taken a narrower view centred on the iOS ecosystem, identifying the “market for app stores for iOS in India” as the relevant market for assessing Apple’s conduct. Within this framework, the App Store effectively becomes the only gateway for developers to reach iPhone and iPad users, leading the Commission to form a prima facie view that Apple holds a monopoly position in this segment.
Apple has consistently pushed back against this framing. In its submissions, the company has argued that it operates in a highly competitive smartphone market in India, where its share remains small relative to Android. It has also maintained that its App Store policies are designed to ensure security, privacy and a consistent user experience, and that commissions reflect the value of its platform, tools and global reach.
However, the regulator’s approach indicates that the assessment is shifting from device-level competition to ecosystem-level control.
The current scrutiny is not the first time Apple’s business practices have come under the CCI’s lens. More than a decade ago, in a 2011 case, the regulator examined similar concerns around restrictions within Apple’s ecosystem.
In that matter, Apple was accused of entering into arrangements with telecom operators such as Airtel and Vodafone that effectively locked iPhones to specific networks, while also restricting users to applications approved through its App Store.
The Director General’s investigation found that the arrangement had elements of a tie-in, particularly in how the device and network services were linked, and acknowledged that Apple exercised control over which applications could run on its devices.
However, the case did not result in any finding of anti-competitive harm. A key reason was Apple’s limited presence in India at the time. According to data cited in the investigation, Apple’s share of the smartphone market was around 1.5 per cent in 2008, less than 1 per cent in 2009 and 2010, and about 2.4 per cent in 2011, based on IDC estimates referenced in the DG report.
On this basis, the regulator concluded that while certain practices could restrict user choice, they did not have an appreciable adverse effect on competition given the company’s small footprint in the overall market.
Apple’s defence in that case also centred on market structure. The company and its partners argued that the Indian telecom and handset markets were highly competitive, that agreements with operators were non-exclusive, and that any restrictions were either temporary or consistent with global industry practices. They also pointed out that iPhones were eventually made available in an unlocked form and that consumers had multiple alternatives in both devices and services.
More than a decade later, the same arguments are being revisited in a very different market environment.
Apple’s share of India’s smartphone market has risen to around 9 per cent in the first quarter of CY 2026, according to Counterpoint Research, driven by strong demand in the premium segment and aggressive financing schemes.
The scale of Apple’s ecosystem in India has also expanded sharply. A study published by the company last year estimated that the App Store ecosystem facilitated Rs 44,447 crore in billings and sales to developers in India in 2024. It also stated that app downloads in India more than tripled from 2019 to 2024 and developer earnings from Indian users rose over fivefold, indicating a sharp increase in both usage and monetisation on iOS.
As for the overall app ecosystem (Android and iOS), data from Sensor Tower shows that in-app purchase revenue in India has grown from about $520 million in 2021 to over $1 billion in 2025, and is projected to reach $1.25 billion in 2026. In the first quarter of 2026 alone, India generated more than $300 million in in-app revenue, with non-gaming apps contributing over $200 million and growing 44 per cent year-on-year.
While Apple is still behind Android players in overall volume, its presence in the high-value segment has expanded, and its user base has grown steadily.
This transformation is central to the CCI’s current case. The earlier investigation acknowledged restrictive elements in Apple’s ecosystem but found them economically insignificant. Today, those same elements are tied to a rapidly expanding revenue stream that directly affects developers, payment providers and digital businesses.
In effect, the regulatory question has shifted from whether Apple imposes restrictions to whether those restrictions now matter at scale.
The issues being examined in India closely mirror developments in other jurisdictions, where regulators have taken a more interventionist approach to platform markets.
In the European Union, the Digital Markets Act has classified companies like Apple as gatekeepers and imposed obligations aimed at opening up their ecosystems. Apple has been required to allow alternative app distribution channels, enable third-party app stores and provide developers with the ability to use external payment systems.
Japan has also pushed for changes, particularly around payments, leading Apple to allow certain third-party billing options in specific app categories.
These measures reflect a broader shift in how competition authorities are approaching digital ecosystems, moving from case-by-case penalties to structural remedies that reduce platform control over distribution and monetisation.
The outcome of the CCI’s case is likely to hinge not just on the findings of abuse, but on how far the regulator is willing to intervene in reshaping Apple’s ecosystem in India.
The most immediate area of focus is expected to be payments. Given that the case centres heavily on Apple’s in-app billing system, allowing alternative payment mechanisms or easing restrictions on how developers process transactions appears to be the most direct and implementable remedy.
Changes to commission structures or greater flexibility in pricing could also come into play, particularly as the regulator examines whether Apple’s fees are disproportionate in the Indian context.
The question of third-party app stores, however, presents a more complex challenge. While global precedent, particularly in the European Union, shows that opening up app distribution is a key regulatory tool, the Indian case has so far been more tightly framed around payments and market access within the existing App Store framework. Any move to allow alternative app stores would mark a more fundamental shift and may not be the immediate priority.
Analysts say this distinction is important in understanding how the case could unfold. “While Apple sharing financial data with the CCI is an important step in the investigation, it does not necessarily mean India will follow the EU’s path and mandate third-party app stores,” said Sanyam Chaurasia, principal analyst at Omdia. “Indian regulators are likely to focus more broadly on whether App Store policies restrict competition, developer choice, or monetisation opportunities.”
He added that Apple continues to view India as an ecosystem expansion market, with a focus on moving users towards higher-value devices and deepening adoption across its product portfolio. “If regulatory changes emerge, the biggest immediate beneficiaries would likely be Indian developers through greater flexibility and improved economics, while the impact on consumers and Apple’s broader ecosystem strategy would be gradual rather than disruptive given iOS’s relatively small installed base in India,” Chaurasia said.
This suggests that while the direction of regulatory scrutiny is aligned with global trends, the pace and depth of intervention in India could be more measured. The case may not immediately force a structural opening of iOS, but it does increase the likelihood of incremental changes that weaken Apple’s control over payments and developer monetisation.
ChatGPT Plus, Gemini Advanced, Claude Pro and Copilot Pro all promise access to more capable models, faster performance and advanced features, reflecting a broader industry belief that users will eventually pay for AI in the same way they pay for streaming, cloud storage or productivity software.
According to McKinsey’s report, titled “The Economic Potential of Generative AI”, generative AI could add between $2.6 trillion and $4.4 trillion in annual economic value globally, highlighting why companies are betting consumers will eventually pay for AI.
Yet adoption alone does not ensure subscription growth. According to Deloitte’s 2026 Digital Media Trends report, consumers are becoming more selective about digital subscriptions and reassessing which services are worth paying for.
Millions of users already pay for advanced AI tools, but whether AI becomes an essential service or remains a premium offering remains unclear. The question is critical for technology companies. As companies spend billions of dollars building and operating increasingly powerful models, the challenge is no longer simply getting people to use AI. It is determining how they will ultimately pay for it.
If AI companies appear confident that users will eventually pay for artificial intelligence, it is because the economics of the technology leave them with little choice.
According to Gartner’s Forecast: AI Spending, Worldwide, 2024–2029, global spending on AI is expected to reach $2.59 trillion in 2026, representing a 47 per cent increase from the previous year. The figure underscores the scale of investment flowing into the sector, from data centres and specialised chips to model development and software integration.
Building and operating advanced AI systems remains one of the most capital-intensive activities in the technology industry, making sustainable revenue streams increasingly important for companies competing in the space.
The spending boom is also entering a new phase. According to Gartner, much of the initial wave of AI investment was driven by technology vendors and hyperscalers building the infrastructure needed to support generative AI. The next stage is expected to be led by enterprises as they move from experimentation to implementation.
As businesses begin embedding AI more deeply into workflows, customer service, software development and productivity tools, demand for more capable AI systems is likely to expand beyond technology enthusiasts and early adopters.
That shift has important implications for consumers. If premium AI tools become standard across workplaces, access to advanced models could increasingly be viewed as a productivity advantage rather than a discretionary expense. The divide may no longer be between users and non-users of AI, but between those with access to more capable systems and those relying solely on free alternatives.
The industry is moving beyond the initial excitement surrounding chatbots and image generators into a phase where customers, whether individuals or enterprises, are asking a more practical question: Does the technology deliver enough value to justify ongoing spending?
The answer may ultimately determine whether AI subscriptions become a mainstream category of consumer spending or remain a premium offering for a narrower group of users.
The industry’s confidence in paid AI services is not based solely on hype. It is also rooted in a simple belief: if AI can save users enough time, many will be willing to pay for it.
According to McKinsey’s report, much of generative AI’s economic impact is expected to come from areas such as customer operations, software engineering, marketing and research — functions where employees spend large parts of their day working with information, content and communication.
This helps explain why AI subscriptions have found their strongest audience among professionals. For software developers, consultants, researchers, marketers and content creators, AI is increasingly becoming a work tool rather than a novelty.
The more a job depends on writing, analysing information, coding or problem-solving, the greater the potential value AI can deliver, according to McKinsey.
The consultancy’s research also highlights how much time knowledge workers spend searching for information, reviewing documents and performing repetitive tasks. Generative AI has the potential to reduce some of that workload, allowing workers to focus on higher-value activities.
For an individual user, the calculation can be relatively straightforward. If a paid AI subscription saves several hours each month, the cost may be easier to justify.
McKinsey also points to early evidence of productivity gains within businesses. In one example cited by the firm, a company with 5,000 customer service agents recorded a 14 per cent increase in issue resolution after deploying generative AI tools.
While such results will vary across industries and use cases, they help explain why businesses are increasingly willing to invest in AI technologies.
For a growing group of users, therefore, the debate is no longer whether AI is useful. The more relevant question is whether the productivity gains are large enough to justify a monthly subscription fee.
For professionals who rely on AI regularly, the answer may already be yes.
If the productivity case explains why some users are willing to pay for AI, the consumer reality is more complicated.
AI subscriptions are entering a market where consumers are already managing a growing number of recurring payments. Streaming platforms, music services, cloud storage, gaming memberships and productivity software all compete for a share of monthly spending.
For AI companies, the challenge is not simply convincing users that AI is useful, but persuading them that it deserves a place alongside subscriptions they already pay for.
According to Deloitte’s 2026 Digital Media Trends report, consumers are becoming increasingly selective about recurring digital expenses as subscription costs continue to rise. The report highlights a growing sense of subscription fatigue, with many users reassessing which services provide enough value to justify monthly payments.
Some key findings include:
This matters because AI is arriving as a new subscription category rather than replacing an existing one.
Unlike internet connectivity or mobile services, which have become essential household expenses, AI is still competing for discretionary spending.
For many consumers, the question is not whether AI is helpful, but whether it is helpful enough to warrant another monthly payment.
India’s AI subscription challenge
The challenge becomes even more pronounced in India.
While digital adoption has expanded rapidly, Indian consumers have traditionally favoured affordable and ad-supported digital services.
According to the FICCI-EY Media & Entertainment Report 2026, digital subscription revenues in India grew 60 per cent in 2025 to Rs 163 billion, while paid video subscriptions rose to 216 million.
The growth suggests Indian consumers are willing to pay for digital services when they perceive clear value.
The report also noted that digital advertising revenue increased 26 per cent during the year, indicating that both advertising-supported and subscription-based models continue to coexist.
That dynamic could shape the adoption of paid AI services as well.
While professionals, developers and knowledge workers may find sufficient value in premium AI subscriptions, mainstream users may be more inclined to rely on free offerings unless the benefits become significantly more compelling.
Deloitte’s findings indicate that subscription growth increasingly depends on value rather than availability. For AI companies, success will hinge on whether premium services can offer benefits compelling enough to justify a recurring monthly payment.
The contrast between power users and casual users is already becoming apparent. Professionals may see AI subscriptions as worth paying for, but many casual users remain satisfied with free tools.
Who actually pays for AI?
So far, the conversation has largely focused on whether consumers are willing to pay directly for AI tools such as ChatGPT Plus, Gemini Advanced or Claude Pro.
But the bigger shift may be happening elsewhere.
Rather than convincing users to subscribe to AI as a standalone product, technology companies are increasingly embedding AI into products people already use — and, in some cases, already pay for.
Microsoft offers one of the clearest examples of this strategy. The company has integrated Copilot across Microsoft 365, bringing AI features to applications such as Word, Excel, Outlook and PowerPoint.
According to CNBC, Microsoft had 89 million consumer Microsoft 365 subscribers as of mid-2025.
For many of these users, AI is no longer a separate purchase decision. It has become part of a broader productivity package.
Google is pursuing a similar strategy. Gemini is increasingly being integrated across Gmail, Docs, Search and Android services.
Apple, meanwhile, is embedding Apple Intelligence into its ecosystem of devices and operating systems.
The same trend is beginning to emerge in India.
Smartphone brands are marketing AI-powered features as part of the device experience rather than as separate paid services. Features such as AI-assisted photo editing, real-time translation, call summaries and writing tools are increasingly bundled into premium smartphones.
Consumers may not think of themselves as paying for AI, but AI is becoming one of the reasons they upgrade devices or remain within a particular ecosystem.
In that sense, the future of AI monetisation may resemble what happened with other technologies:
AI may ultimately follow the same path.
Three users, three different futures
Findings from McKinsey, Deloitte and Gartner suggest the future of AI subscriptions is unlikely to follow a single path. Instead, adoption and willingness to pay will vary significantly across different categories of users.
The power user
This group includes developers, consultants, researchers, analysts, educators and content creators whose work depends heavily on information and productivity.
According to McKinsey’s report, generative AI could significantly boost labour productivity across knowledge-intensive occupations.
For these users, AI is not merely a convenience. It is increasingly becoming a professional tool that saves time and improves output.
For a software developer generating code, a consultant preparing research briefs or a journalist analysing large volumes of information, the value proposition is relatively straightforward.
If AI saves several hours each week, the subscription cost becomes easier to justify.
The second group uses AI occasionally rather than daily. These users may ask AI to rewrite an email, suggest a travel itinerary, summarise an article or answer a quick question.
For them, free versions of AI services are often sufficient.
This is where Deloitte’s findings become relevant. As subscription fatigue grows, consumers are becoming increasingly selective about recurring expenses.
Unless premium AI features deliver clear and consistent benefits, many casual users may see little reason to upgrade.
These are users who gain access to AI through products they already pay for rather than through dedicated AI subscriptions.
They may encounter AI through a Microsoft 365 plan, a Google Workspace subscription, a premium smartphone or a bundled digital service.
In India, this model may prove particularly effective. Consumers have historically responded well to bundled offerings, whether through telecom plans, streaming packages or device ecosystems.
As AI becomes integrated into smartphones, productivity software and cloud services, many users may end up paying for AI indirectly without ever signing up for a dedicated AI subscription.
That possibility reinforces a broader point.
The future of AI may not be defined by how many people subscribe directly to ChatGPT Plus or Gemini Advanced. Instead, it may be shaped by how successfully technology companies weave AI into products consumers already use.
If that happens, the question of whether people will pay for AI becomes somewhat irrelevant. They already will be — just not through a separate line item on their monthly bill.
The future of AI is unlikely to be defined by standalone subscriptions alone. Instead, AI may increasingly be woven into the software, devices and digital services consumers already use, much like cloud storage, GPS and video conferencing before it.
Xiaomi has launched the Xiaomi 17T smartphone in India. The new T-series smartphone comes with a Leica-tuned camera system, a 6.59-inch AMOLED display, a 6,500mAh battery, and Xiaomi HyperOS 3. The company said the device is aimed at users looking for a combination of photography features, multimedia capabilities, battery life, and AI-powered tools. The Xiaomi 17T will be offered in two storage variants and will be available through online and offline retail channels starting June 10.
According to Xiaomi, the Xiaomi 17T is available in the following configurations:
The Xiaomi 17T will go on sale on June 10 via MI’s website, ecommerce platform Amazon, and Xiaomi retail stores.
The Xiaomi 17T sports a 6.59-inch 1.5K AMOLED display with a peak brightness of up to 3,500 nits and a 120Hz refresh rate. The display boasts support for HDR10+ and Dolby Vision viewing.
The smartphone is powered by the MediaTek Dimensity 8500 Ultra chip, paired with 12GB of LPDDR5X RAM and up to 512GB of UFS 4.1 storage. The Xiaomi 17T runs Xiaomi HyperOS 3 and includes AI-powered features such as Xiaomi HyperAI, Google Gemini integration, and more.
The Xiaomi 17T features a Leica-powered triple rear camera setup led by a 50-megapixel primary camera. The smartphone also includes a 50MP 5x periscope telephoto camera and a 12MP ultra-wide camera. Xiaomi has also introduced Leica Live Moment, a feature said to be designed to capture a sequence of moments rather than a single frame.
The device packs a 6,500mAh battery. According to Xiaomi, the battery is designed to provide all-day usage across productivity, entertainment, and content creation tasks. It comes with support for 67W HyperCharge wired charging.
