iWorld
Feb-18: Mobile broadband numbers increase as wired internet subscribers decline
BENGALURU: The total number of broadband internet connections have increased by about eight per cent in the calendar year 2018 (year started 1 January 2018, CY-2018) until 28 February 2018 (Feb-18) as per Telecom Regulatory Authority of India (TRAI) data. The period under review in this paper is the period between 1 January 2018 and 28 February 2018. CY-2017 closed with 362.87 million (36.287 crore) broadband connections as on 31 December 2017. The total number of broadband connections in Feb-18 was 392.06 million (39.206 crore). Broadband internet growth in the country was driven by mobile (phones and dongles) internet services which had about 8.5 per cent subscriber growth and closed February 2018 with 373.94 million (37.384 crore) subscribers. TRAI defines broadband internet speed as download speeds equal to or exceeding 512 kbps. TRAI data has been rounded off to the nearest 10,000, hence the accuracy of this report is limited to that extent.
During the period under consideration, wired internet subscriber numbers declined 0.8 per cent to 17.72 million (1.772 crore) from the 17.86 million (1.786 crore) subscribers reported at the end of December 2017 or as at 1 January 2018. Fixed wireless (WiFi, Wi-Max, point-to-point radio and VSAT) subscriber numbers also declined 9.1 per cent during the period to 0.4 million (0.04 crore) from 0.44 million (0.044 crore).
Among the top five internet players, Indian telecom major Bharti Airtel (Airtel) showed the highest growth rate during the period under review at about 13 per cent. However, in absolute numbers, it was Mukesh Ambani’s biggest startup in the world – Reliance Jio Infocomm or Jio that added the most number of subscribers in the two months of the current year at 17.04 million or 1.704 crore. During the period, Airtel added 9.15 million (0.915 crore) broadband internet subscribers.
The top five service providers in India as on 28 February 2018 constituted 94.99 per cent market share of the total broadband subscribers. These service providers were Jio (177.13 million, 17.713 crore), Airtel (80.24 million 8.024 crore), Vodafone (55.54 million, 5.554 crore), Idea Cellular or Idea (38.52 million, 3.852 crore) and BSNL (21.00 million, 2.1 crore).
While the first five players saw a growth of subscribers during the period under review, the government-owned BSNL or Bharat Sanchar Nigam Limited has been losing them. BSNL had 21.95 (2.195 million) broadband subscribers and lost about 0.95 million (0.095 crore) subscribers or de-grew by over four per cent.
Top five wireless broadband internet players
As mentioned above, broadband wireless aka mobile internet players have been the broadband internet subscriber numbers’ growth drivers. As on 28 February 2018, the top five wireless broadband service providers were Jio (177.13 million, 17.713 crore), Airtel (78.07 million, 7.807 crore), Vodafone (55.54 million, 5.554 crore), Idea (38.52 million 3.852 crore) and BSNL (11.71 million, 1.171 crore).
Here also, Airtel has reported the largest growth in percentage terms – it grew by about 13 per cent, while Jio had the highest growth in absolute numbers – Jio grew by 17.04 million (1.704 crore) during the first two months of 2018. BSNL has been bleeding wireless broadband internet players during this period. It lost about 0.86 million (0.086 crore) subscribers or de-grew by approximately nine per cent.
Top five wired broadband internet players
As has also been mentioned above, wired internet subscriber numbers have declined during the first two months of 2018. As on 28 February 2018, the top five wired broadband service providers were BSNL (9.30 million, 0.93 crore), Airtel (2.17 million, 0.217 crore), Atria
Convergence Technologies or ACT (1.30 million, 0.13 crore), MTNL (0.88 million, 0.088 crore) and Hathway Cable & Datacom (0.75 million, 0.075 crore).
The top five wired internet players in the made up over 81 per cent of the total wired internet subscribers in India. Their share has grown during the period under consideration despite a slight drop in share in Jan-18. Except for BSNL and the other government-owned player Mahanagar Telecom Nigam Limited or MTNL, the other three players among the top five have grown the number of wired broadband internet subscribers. BSNL lost about 80,000 subscribers while MTNL lost about 30,000 subscribers during Jan-Feb 2018. The other three players among the top five have added about 60,000 subscribers (added about 20,000 subscribers each) during the period under review.
Among the other wired broadband internet players, besides the five mentioned above, are television multi system operators (MSOs) and local cable TV operators (LCOs). The numbers provided by TRAI indicate that while during the period under review, the top five wired broadband internet players lost about 50,000 subscribers, the total number of wired broadband internet subscribers fell by about 0.14 million or 140,000. This means that the other players have lost about 90,000 subscribers.
Also Read :
Jio continues leading broadband subs addition while wireline internet loses subs in Oct
TRAI bats for converged regulator & renaming of NTP’18
SC could take up TRAI-Star case on tariff regulations
eNews
India uses ChatGPT for technical tasks nearly 4 times the world average: OpenAI
From classrooms to code, India’s AI use is increasingly skill-driven and youth-led.
MUMBAI: If code is the new currency, India is already minting it by the million prompts. In the world’s largest democracy, artificial intelligence is no longer a distant abstraction or a boardroom buzzword. It is a daily companion, drafting emails in Hyderabad, debugging code in Bengaluru, polishing essays in Delhi, and fielding life advice in towns far beyond the metros. Fresh data from OpenAI’s “Signals” initiative offers a rare, granular glimpse into how India is using ChatGPT, and the numbers suggest the country is not just adopting AI; it is actively shaping its use.
India is one of the largest markets globally for ChatGPT’s weekly active users and ranks among the top five countries for API usage. With OpenAI’s global consumer base exceeding 800 million users, most of them on free tiers, the dataset captures adoption patterns that go far beyond enterprise subscriptions.
Indian users, notably, are punching above their weight when it comes to advanced capabilities. Among ChatGPT Plus and Pro subscribers, usage of the data analysis tool is roughly four times above the global median. Use of Codex, OpenAI’s coding platform, is about three times above the median. Indians are nearly three times more likely than the global median to ask coding-related questions and almost twice as likely to seek help on education and learning.
This matters because it signals something economists call a shrinking “capability overhang”, which is the gap between what AI tools can do and how fully users exploit them. In India, that gap appears to be narrowing rapidly.
The geography of this coding intensity tracks the country’s technology hubs. Telangana, which is home to Hyderabad, ranks first in usage of OpenAI’s coding capabilities. Karnataka, home to Bengaluru, follows in second place, while Tamil Nadu comes third. In other words, the prompt traffic mirrors the tech corridors.
Nearly two-thirds of consumer ChatGPT messages in India are now non-work related, while slightly over one-third are tied to work. That marks a significant shift. In earlier phases of adoption, work was the dominant use case. It was only in early 2025 that non-work messages overtook professional use, and the divergence widened throughout the year.
Even so, India remains slightly above the global average in work-related usage. Around 35 per cent of consumer messages in India relate to work, compared with roughly 30 per cent globally.
At work, the emphasis is squarely task-oriented. Around 45 per cent of work-related conversations fall into “doing” behaviours such as drafting documents, transforming text, and completing tasks, compared with a much smaller share in non-work contexts. Technical help and writing dominate. In offices across the country, ChatGPT functions as a digital co-pilot that debugs code, polishes presentations, and unblocks stalled workflows.
Outside work, the tone shifts. Over 35 per cent of non-work messages revolve around practical guidance, which includes everyday advice and how-to queries. Roughly 20 per cent relate to seeking information. Nearly one-fifth involve writing tasks such as drafting or editing. Self-expression and learning loom large. In personal life, Indians appear to use AI less as an executor and more as an explainer, sounding board, and study partner.
India’s demographic dividend is clearly reflected in its AI habits.
Users aged 18 to 24 now account for just under half of all ChatGPT messages sent in the country. They surpassed the 25 to 34 age group in mid-2024 and have held the lead ever since. Globally, the 18 to 24 cohort accounts for about one-third of messages; in India, the share is markedly higher.
Combined, users aged 18 to 34 generate roughly 80 per cent of total consumer ChatGPT messages in India. Given that around 40 per cent of India’s population is under 25, the youth skew is unsurprising, but its implications are profound. Education-related queries, early-career problem-solving, and skills development are likely to dominate near-term AI impacts.
Usage patterns also differ by age. The 18 to 24 cohort accounts for a near majority of messages seeking practical guidance, technical help, and self-expression. Meanwhile, the 24 to 34 group sends a slightly higher share of multimedia and technical help queries relative to its overall share of usage.
If AI norms are being written in real time, it is young Indians who are holding the pen.
OpenAI does not collect gender data, but inferred patterns based on typically masculine and feminine first names reveal a measurable gap in India. A little under 60 per cent of users have typically masculine names, and just over 40 per cent have typically feminine names. This skew is more pronounced than the global average.
Worldwide, users with typically feminine names now account for slightly more than half of all messages. This shift occurred only in the summer of 2025, when feminine-name usage overtook masculine-name usage globally. In India, the gap persists, although it has been narrowing over the past year.
There are also topical differences. Users with typically feminine names are more likely to send messages related to self-expression, practical guidance, and writing. Those with typically masculine names lean more towards seeking information and technical help.
The data does not capture motivations, but it does highlight where inclusion efforts and digital literacy initiatives could focus if AI is to broaden opportunity rather than deepen divides.
The consumer story aligns with India’s broader AI momentum. The country ranks second globally in AI skills penetration and has one of the fastest-growing AI talent pools. It accounts for 9.2 per cent of global AI publications in computer science as of 2023, which represents a substantial contribution to research output.
At the same time, investment in AI data centres and digital public infrastructure is expanding, promising to knit together datasets and resources at scale. Enterprise adoption is also robust, which suggests that consumer experimentation is unfolding alongside institutional integration.
OpenAI’s “Signals” project is built with aggregated, privacy-preserving data and released with a time lag. It aims to provide a durable measurement layer for the AI era. The idea is not to track individuals, but to surface patterns such as where adoption is accelerating, who is using the tools, and what they are actually doing.
In a country as vast and varied as India, such evidence is more than academic. It shapes decisions about workforce training, small business support, education policy, and safeguards.
For now, the numbers paint a picture of a nation that is not merely consuming AI, but conversing with it in an energetic, experimental, and increasingly skilful manner. In India, the future of work and learning is not being downloaded. It is being drafted, debugged, and rewritten in real time.
