India’s Telegram Ban Was Temporary. The Power Behind It Is Not.

Apar Gupta is a fellow at Tech Policy Press.

On the afternoon of June 16, India’s Ministry of Electronics and Information Technology ordered Telegram switched off across the whole of the country within one hour. The block was to run until June 22. Telegram was told to disable its message editing feature until June 30. The Department of Telecommunications was to instruct app stores and internet providers to pull the app, and the matter was to reach a committee within 48 hours.

The order carried a final instruction, under Rule 16 of the blocking rules, that "no part of this communication should be made public." Public notification was instead provided on the morning of June 16 by the National Testing Agency (NTA), which reasoned that the ban would be temporary. It was tied to one event, the re-examination of the National Eligibility-cum-Entrance Test, the medical-college entrance exam that about 2.2 million students were due to take onJune 21. As per the government justification — which was then subsequently accepted by the High Court of Delhi — this was a narrow, time-bound measure in response to an emergency.

I litigate blocking and shutdown cases, and I have read several blocking orders and challenged them in legal proceedings. The Telegram ban is a dangerous escalation by the government — one that should worry anyone who uses a smartphone in India, not mainly because of what it did to one app for six days, but because of the legal precedent cemented by it.

The timeline for blocking

While blocking orders are shrouded in secrecy, the judgement of the Delhi High Court describes a rapid process. On May 21, the NTA wrote to the Ministry of Electronics and IT (MEITY), as it has the formal authority to issue blocking orders. Acting on it, MeitY issued a notice on June 1 and met Telegram and NTA on June 3 (paras 5–6). On June 9, the Ministry sent Telegram a list of URLs, channels, accounts and bots which resulted in the platform disabling 900 of the 1,300 links. It is pertinent to note that Telegram is used by approximately 150 million accounts in India and, since 2024, has received the highest number of law enforcement requests globally from the country.

Despite these compliance actions, MeitY passed the interim order on June 16, which resulted in the entire platform being blocked and Telegram challenging the move in the Delhi High Court the next day. A committee under Rule 7 heard the parties on June 17, and a final order confirmed the block. The judgment was reserved on June 18 and delivered on June 19. The re-exam was on June 21, so the block was to be lifted on June 22.

It is important to consider that this timeline runs straight through the court's summer recess. This is an annual break from the normal functioning of the Delhi High Court when it rises for about a month and the Chief Justice nominates a handful of judges to take urgent matters. They sit both singly and in division benches on fixed days, and a party with something pressing files a form for urgent mentioning and asks to be listed.

The case was heard by Honorable Justice Tejas Karia, sitting as Vacation Judge. Many legal observers were left wondering whether a constitutional question about the State's power to switch off an entire platform used by 150 million people being argued and decided inside a vacation window—against a clock the government itself had set—was the best legal strategy. However, such comment fails to recognize the massive harm to free expression as much as the commercial interest of any technology service provider due to the rapid attrition of users following even short service disruptions.

Telegram suffers a defeat in Delhi High Court

The statutory power to issue blocking directions originates from Section 69A of the Information Technology Act, 2000 which lets MEITY block "public access of any information through any computer resource" where the Ministry is satisfied that this is necessary and in the interest of, among other things, public order or to prevent incitement to a cognizable offense. An intermediary that disobeys can be jailed for up to seven years. The procedure is in the 2009 Blocking Rules, which provide a hearing for the affected party before a decision (Rule 8), an emergency route that skips that hearing where no delay is acceptable (Rule 9), and the confidentiality clause the order used (Rule 16). Above these rules sits the constitutionally protected fundamental right to freedom of speech and expression under Article 19(1)(a) and the restrictions that are contained under Article 19(2).

Telegram argued that the blocking orders were beyond the statutory power for blocking content, as it permits blocking specific information and not "a blanket restriction on an entire intermediary platform." The company further submitted that it ignored the more than 150 million lawful users who rely on Telegram, including students and teachers sharing study material. Telegram relied on Anuradha Bhasin v. Union of India, (2020) 3 SCC 637, which requires the least restrictive measure. As Reuters reported, in its petition, which the government kept off the public record, Telegram argued that the order rested on the "impermissible premise that misuse by a subset of users justifies blocking of an entire platform."

The governing authority on website blocking powers in India beyond the statute is contained in the case of Shreya Singhal v. Union of India, (2015) 5 SCC 1, where the Supreme Court upheld 69A precisely because it read the power as narrow, exercised through reasoned written orders open to challenge. However, over time, the Union Government has sought to apply this power expansively, repeatedly arguing that the nature of digital communications—due to their speed and scale—requires a lower constitutional threshold for judicial review. While this argument was rejected in Shreya Singhal, this was a central feature of the government’s defense, which premised its legal arguments on Telegram's architecture, specifically its cloud storage, large public channels, bot ecosystem, usernames in place of phone numbers, and editable and self-destructing messages. Layered on top of this argument, the government held that "information," as defined in the Information Technology Act, 2000 includes not only specific posts and files but "the underlying technological components" through which content moves, so that an application is itself blockable information. The phrase "any information," as the court reasoned, is "sufficiently broad to encompass all information hosted by an intermediary where the scale of harm so warrants." The Attorney General added a sharp proposition by submitting that an entity that is by its own design "structurally incapable of ensuring accountability" cannot invoke proportionality, and a "profit-driven commercial platform" cannot selectively rely on it. This inverts the constitutional doctrine of proportionality which requires the least restrictive measure and provides for a complete bypass to its application.

The court upheld the Telegram block on both questions it framed. On reasons, it held that the emergency made the reasons sufficient and that the later final order could fortify them. It also held that, as per its reading of the power of the Government under the Information Technology Act, 2000, there was "no reason to exclude an application or platform" from the meaning of "information," so MeitY could block Telegram itself. When applying the proportionality doctrine, it accepted the architecture argument almost whole, found narrower takedowns useless against mirror channels and bots, and rested on the limited duration of the orders, which in its view made them narrowly tailored. Doing so it dismissed Telegram's petition and curiously, the Shreya Singhal judgment that saved this very power of blocking by reading it narrowly, appears nowhere in the decision.

What the future holds

The conspiracy theory going around is that the real winner in this series of events is Meta, on the premise that WhatsApp will gain users due to the service disruption of Telegram for a week. Telegram founder Pavel Durov spent the week backing the theory, telling X he “wouldn’t be surprised if” Reliance, which he thinks is part-owned by Meta, joined WhatsApp in lobbying for the ban. (Reliance Jio denied it. Durov had possibly mixed up Reliance Communications, whose network code he kept waving around, with the unrelated Reliance Industries).

But in the long term, the order doesn't give any single company an advantage. It hands over power to the government, and that power points back at everyone, including Meta. Here, the actual blocking of an entire platform is not necessary, and its value rests in the Ministry setting terms for day-to-day cases of censorship. It includes using this precedent as a threat to secure compliance with broad, political censorship of thousands of accounts or platform changes, such as its traceability demand.

This is the point where platform-wide blocking stops being a one-off concern and becomes a negotiation ploy where platforms and user rights play to a handicap. Amber Sinha set this out at length in Tech Policy Press, and I said as much the day after the order came out, next to IFF's statement of concern. So expect a more confident Ministry across every platform, reaching for whatever excuse sounds best that week ranging from morality, national security, public panic, and now — said with a straight face — the integrity of a medical entrance exam. The pending appeal of X Corp v. Union of India is testing the Sahyog portal and Section 79(3)(b). Each one asked how hard a platform could be pushed. After Telegram, the answer is that it can be switched off completely. Sinha adds the grim part. With a full shutdown now a real threat, platforms won't wait to be asked. They'll over-comply, ahead of time, with demands they would once have fought.

The bigger shift is in how courts look at any of this. Judicial review is meant to protect speech in India. In practice, however, it checks an application of the principles of natural justice and whether the process was legal. In Tata Cellular v. Union of India, (1994) 6 SCC 651, the Court said as much — it looks at how a decision was made, not whether it was right, and won't replace the official's judgment with its own. So the government can limit a right as long as it produces the right paperwork. The core of that paperwork is a reasoned, written order and this is neither novel and sits within the everyday practice of an Indian constitutional court. For instance, when faced with unregulated phone tapping in PUCL v. Union of India, (1997) 1 SCC 301, the Court didn't strike the power down. It wrote a set of procedural guidelines and left it there. In Shreya Singhal, it upheld Section 69A because of its safeguards. Even Anuradha Bhasin v. Union of India, (2020) 3 SCC 637, for all the praise its four-part proportionality test gets, ended by sending the Kashmir shutdown orders back to executive review committees instead of quashing them.

This judgment is the latest of that kind. Its proportionality finding rests almost entirely on the block being temporary. It does little to ask whether something narrower would have worked. And it should have asked, because the NTA's own Director-General, Abhishek Singh, admitted there had been no leak at all, only fake messages spreading panic about the re-examination. Students were on VPNs and other apps within a day.

Now for the part that has really changed. If courts only checked the form, then the real limit on censorship was the bureaucracy's own incompetence. In the past, it has been argued that government malevolence is limited only by its competence. Put simply, the sloth of a government office is the best democratic safeguard. That comfort is gone. The orders that used to arrive half-finished now arrive polished, technically literate, and written in the style judges like. It logs a meeting with the company on June 3, recites "exhausted alternatives," and counts channels and bots reaching about 146,000 accounts. It even came with expert backing from two of the country's most senior technologists. IIT Madras's Director, V. Kamakoti, who also sits on the National Security Advisory Board, ran a live demonstration purporting to show how a PDF posted on Telegram at a fixed time could later be swapped for a different file while the original timestamp stayed frozen, the better to fake proof of a "pre-exam" leak. He declined to reveal the method, calling it responsible disclosure. IIT Kanpur's Director, Manindra Agrawal, pressed the same case on X. He said Telegram lets a user edit a post "without reflecting that edit has been done," and "no other major social media channel has this."

Both claims were promptly contested by the two teenagers who had exposed the CBSE on-screen-marking mess weeks earlier. Nisarga Adhikary, a nineteen-year-old whom Agrawal's own institute had just hired into its C3iHub lab, replied that Telegram's client is open source, posted the exact code that records an edit's date and time, and noted that a platform-wide block wasn't even technically possible as proxy routes straight around it. Sarthak Sidhant, the other teenager, who had testified before a parliamentary panel, posted a screenshot of an edited Telegram message wearing a plain "edited" tag, captioned, with due ceremony, as fact-checking the IIT-K director with a "meow meow" message.

The irony was that Telegram's own founder had already conceded the narrow point the directors were overstating. Durov said the platform had removed hundreds of offending channels and was "making the 'edited' label more visible to prevent backdating scams," an admission the government promptly read back to the court as proof the problem was real.

And now the part I'm warned not to speculate about—so I will. The same models that can help a citizen draft a writ petition or an RTI (which gave rise to the pop-up campaign by the Cockroach Janta Party or helped the teenagers that exposed the CBSE OSM scandal) can just as easily help a Ministry write a cleaner, well reasoned blocking order, faster, and harder to pick apart. This is part of the "AI accelerant" of India's censorship machine. Hence, when constitutional courts will merely check the form and trust the substance, without undertaking a substantive examination of the proportionality or reasonableness of the censorship.

So, hello, Claude. Can you help me censor 150 million people today?