The Delhi High Court has done something India's content-blocking regime rarely encounters: it asked whether the punishment fit the post.
On April 6, 2026, Justice Purushaindra Kumar Kaurav directed X to restore two satirical accounts — @DrNimoYadav, run by Prateek Sharma, and @Nehr_who, run by Kumar Nayan — that the platform had withheld in India since March 18 under a Section 69A order from the Ministry of Electronics and Information Technology (MeitY). The government had alleged the accounts used AI-manipulated images and video to push false narratives defaming Prime Minister Narendra Modi. The court did not bless the contested posts: it left the specific flagged tweets withheld pending review and ordered the operators to appear before MeitY's inter-ministerial committee. But it refused to let an entire account — and every lawful post on it — stay dark over a handful of items.
A narrow ruling with a large principle
The distinction the court drew is the whole story. India's blocking machinery had escalated from withholding individual posts to suspending whole accounts; the March 18 order reportedly named 12 of them. Once an account is withheld, everything on it disappears, including speech the government never alleged was unlawful. Sharma's counsel made the practical point bluntly: if the offending tweets are identified, delete those tweets and restore the account — there is no need to erase the rest. X's counsel added the structural risk: if an account is deleted wholesale and a court later finds the posts were protected speech, there is no way to restore what was lost.
That is proportionality reasoning, and it matters far beyond two parody handles.
The government's case, fairly stated
Start with the strongest version of MeitY's position. Synthetic media is not a hypothetical threat. AI tools now produce convincing fake images and audio of public figures cheaply and at scale, and a coordinated network pushing fabricated content about a sitting head of government during a polarized period is a legitimate target of state concern. Section 69A is constitutionally sound — the Supreme Court said so in Shreya Singhal v. Union of India (2015), upholding it even as it struck down the vague Section 66A. And MeitY officials were quick to note that the Delhi High Court did not invalidate anything: the flagged posts stay down, the committee process continues, and the order, they argued, "validated" the blocking framework. On their reading, the system worked.
Why blanket blocking still fails the test
The trouble is that Shreya Singhal upheld Section 69A precisely because of its safeguards, not despite their absence. The Court relied on the requirement that blocking orders record reasons in writing, that the 2009 Blocking Rules provide a hearing and a review committee, and that orders remain open to challenge. Proportionality — using the least restrictive means that achieves the aim — is the connective tissue of that holding. Withholding an entire account to suppress a few posts is the opposite of least-restrictive. It sweeps in protected satire, commentary, and the audience's right to receive it, in service of removing content that could be targeted directly.
Satire compounds the point. Parody of the powerful sits at the core of the expression Article 19(1)(a) protects; labelling it a "false narrative" does not convert criticism into a security threat. When the remedy for an allegedly manipulated post is to silence a creator's entire feed, the chilling effect lands on exactly the kind of speech a democracy should tolerate most.
The larger machine the ruling does not reach
The Delhi High Court's intervention is welcome, but it is a single proportionality check on a rapidly expanding apparatus. The scale is real: the government blocked more than 28,000 URLs in 2024 under Section 69A, with X and Facebook absorbing the largest share. And the legal architecture is widening. In X Corp v. Union of India (September 24, 2025), the Karnataka High Court upheld the Sahyog portal and the use of Section 79(3)(b) read with Rule 3(1)(d) as a takedown route "distinct and separate" from Section 69A — and held that X, as a foreign company, cannot invoke Article 19 at all. Critics, including the Software Freedom Law Centre, warn this builds a parallel channel that routes around the very Shreya Singhal safeguards that made 69A constitutional. X is appealing.
That is the worry the Delhi order throws into relief. If proportionality must be litigated account by account while a second, lighter-touch pipeline scales through a police-facing portal, the safeguards become optional in practice.
What proportionate regulation looks like
None of this requires India to tolerate genuine, harmful deepfakes. It requires the state to act at the level of the harm. Flag the specific post, give the originator the written reasons and hearing the Blocking Rules already promise, target content rather than identities, and reserve account-level suspension for cases where the account itself — not a few of its posts — is the unlawful instrument. That is not a loophole for bad actors; it is the difference between a blocking power and a blunt instrument.
The Delhi High Court reminded the system of that difference. The question now is whether the rule survives contact with the machinery being built around it.