UAPA bail jurisprudence: why bail is so hard, and how the SC is reshaping it

TL;DR: The Unlawful Activities (Prevention) Act is built to deny bail. Section 43D(5) tells a court it cannot release an accused if there are reasonable grounds to believe the accusation is prima facie true, and NIA v. Zahoor Ahmad Shah Watali (2019) read that bar so strictly that, for years, an undertrial could sit in jail with almost no exit. Then Union of India v. K.A. Najeeb (2021) reopened a door: Article 21 and the right to a speedy trial can override the statutory bar when incarceration drags on and trial is nowhere in sight. Since then the Supreme Court has granted bail on delay and weak-evidence grounds in Vernon Gonsalves (2023), Shoma Sen (2024), Sheikh Javed Iqbal (2024) and Jalaluddin Khan (2024), even as Gurwinder Singh (2024) reasserted the “jail is the rule” line. The result is a statute designed to deny liberty, and a constitutional court increasingly unwilling to let it. The 5 January 2026 Delhi-riots bail batch, where five accused walked but Umar Khalid and Sharjeel Imam did not, shows how unsettled the balance still is.

On this page

What the UAPA actually is

The Unlawful Activities (Prevention) Act, 1967 is India’s principal anti-terror law. It started life as a modest statute aimed at associations that questioned the country’s sovereignty and integrity. Over six decades, and especially after the repeal of TADA and POTA, it grew into the wide instrument it is today. Amendments in 2004, 2008 (after the Mumbai attacks) and 2019 expanded it to cover terrorist acts, terrorist organisations, funding of terror, and, since 2019, the power to designate individuals (not just groups) as terrorists.

For an ordinary reader the important point is structural. The UAPA is not a single offence. It is a framework that layers terror-specific offences (Chapters IV and VI of the Act) on top of, or alongside, the ordinary penal code. When a person is booked under these chapters, a special set of procedural rules kicks in, and the most consequential of them is the bail rule.

This is where the UAPA stops behaving like ordinary criminal law. Under the general scheme, a person accused of a crime is presumed innocent and is usually entitled to bail unless the State shows good reason to keep them in. We have written about that baseline in our piece on why “bail is the rule, jail is the exception”. The UAPA inverts that presumption for the offences it covers. The law tells the judge, in effect, that the default is custody, and the accused has to climb out.

It helps to see how the statute reached this point. The 1967 Act was a response to a constitutional amendment that allowed Parliament to impose reasonable restrictions on the freedoms of speech and association in the interest of the sovereignty and integrity of India. For decades it was a relatively narrow law about banning associations. The terror-specific architecture arrived later, borrowed largely from the two laws India had let lapse. The Terrorist and Disruptive Activities (Prevention) Act, TADA, expired in 1995 after sustained criticism of its misuse. The Prevention of Terrorism Act, POTA, was repealed in 2004. But the State did not abandon the tools those laws contained; it folded them into the UAPA. The 2004 amendment imported terrorist-act offences. The 2008 amendment, passed in the weeks after the Mumbai attacks, tightened the procedural provisions, including the bail rule and the extended custody timelines. The 2019 amendment added the power to designate individuals as terrorists, not merely organisations.

The point of that history is simple. The UAPA carries the DNA of laws that were withdrawn precisely because they were thought too harsh. Their harshest features, the reverse presumptions and the restricted bail, survived inside the new statute. So when courts grapple with Section 43D(5) today, they are wrestling with a design choice made deliberately: the law is meant to be hard on the accused at the front end, on the theory that terror offences justify a heavier hand than ordinary crime.

The numbers explain why this matters so much. Government data placed before Parliament shows that of 10,440 people arrested under the UAPA between 2019 and 2023, only 335 were convicted, a conviction rate of about 3.2 percent (Deccan Herald). In 2023 alone, 2,914 people were arrested and only 118 were convicted (Maktoob Media). The Supreme Court itself has flagged that the all-India conviction rate sits below 6 percent, and in Jammu and Kashmir below 1 percent (LiveLaw). Put bluntly, the overwhelming majority of people held under the UAPA are eventually acquitted. The punishment, very often, is the process itself, and the centre of that process is the denial of bail. A law designed to be hard at the front end, combined with a near-empty record of convictions at the back end, produces a system where years of custody land overwhelmingly on people the courts will never find guilty.

The Section 43D(5) bail bar

The bail bar lives in Section 43D(5) of the UAPA. Stripped of its proviso, it says that a person accused of an offence under Chapters IV and VI of the Act shall not be released on bail if the court, on a perusal of the case diary or the report filed under Section 173 of the Code of Criminal Procedure, is of the opinion that there are reasonable grounds for believing that the accusation against such person is prima facie true.

Read that again, because the wording does a lot of work. The judge does not ask “has the State proved guilt.” The judge asks whether the accusation is “prima facie true.” And the threshold is framed as a negative: bail shall not be granted if those reasonable grounds exist. So the burden, in practice, sits on the accused to show the accusation is not even prima facie true, before charges are framed and before any evidence is tested.

Two other features make the bar bite harder.

First, the time the State gets to investigate. Under ordinary law, if the police do not file a charge sheet within 60 or 90 days, the accused becomes entitled to “default bail.” The UAPA doubles the upper limit to 180 days (SCC Online). A court can extend even that on a public prosecutor’s report. So the accused can spend half a year in custody before the State has to commit to a case on paper.

Second, the bar is not absolute on its own terms. A line of judgments has clarified that Section 43D(5) is attracted only if the chargesheet actually reveals a prima facie case. If it does not, the embargo does not apply at all (LiveLaw). That qualification matters, and we will return to it, because the whole modern story of UAPA bail is courts probing what “prima facie true” can be made to mean.

The Watali standard: how the bar became a wall

A statutory bar is only as strict as the way courts read it. In National Investigation Agency v. Zahoor Ahmad Shah Watali, decided on 2 April 2019, the Supreme Court read it about as strictly as it could be read.

Watali, accused of funding secessionist activity in Jammu and Kashmir, had been granted bail by the Delhi High Court. The Supreme Court reversed that and restored the trial court’s refusal. In doing so, the Bench laid down how a court must approach the prima facie inquiry under Section 43D(5), and the directions became the controlling standard for years.

The core holdings, as summarised in legal commentary, were these (CaseMine):

  • The “degree of satisfaction” a court needs under the UAPA is lighter than the satisfaction needed to convict. The court only has to see whether the accusation is prima facie true, not whether it is proved.
  • At the bail stage, the court must take the material in the chargesheet at face value. It cannot weigh the evidence, test its admissibility, or judge whether it would survive cross-examination at trial.
  • The court is not to conduct a “mini trial.” It looks at the totality of the material and asks whether, taken as a whole, it supports the accusation.

The practical effect was severe. If the court must accept the State’s material as it stands, without asking whether it is reliable or admissible, then almost any chargesheet that alleges the ingredients of a terror offence will clear the prima facie bar. The accused had no real way to attack the foundation of the case at the bail stage. Critics described the consequence as a near-automatic denial of bail, and the line “the court must accept the NIA’s version of events” became shorthand for the problem (The Wire).

There is a subtle logical trap inside the Watali approach, and it is worth naming because every later judgment circles back to it. If a court must accept the chargesheet material at face value and may not test whether it is admissible or reliable, then the prima facie inquiry collapses into a reading-comprehension exercise. The judge reads the document, sees that it alleges the ingredients of a terror offence, and the bar applies. The accused cannot say “that confession was coerced,” or “that witness is unreliable,” or “that recovery was planted,” because all of those are merits questions reserved for trial. The accused is left arguing about the meaning of words on a page the State drafted. That is why Watali, more than any single later case, defined the experience of being denied UAPA bail.

Watali did not change the statute. It changed how the statute felt to a person sitting in jail. The bar became a wall.

The Najeeb breakthrough: Article 21 enters the room

The wall had one crack, and the Supreme Court found it in Union of India v. K.A. Najeeb, decided on 1 February 2021 and reported at (2021) 3 SCC 713.

Najeeb was an accused in a 2010 case where a professor’s hand was chopped off in Kerala. He was arrested in 2015 after years as an absconder, charged under the UAPA, and he remained in custody as his co-accused were tried and the trial against him stalled. The Kerala High Court granted him bail on the ground of prolonged incarceration. The Union challenged it.

The Supreme Court dismissed the Union’s appeal and kept Najeeb out on bail. The reasoning is the hinge on which all later UAPA bail jurisprudence turns (Supreme Court Observer):

  • Section 43D(5) and the constitutional right under Article 21 operate in different planes. The statutory bar restricts a court’s ordinary power to grant bail. It does not, and cannot, take away the constitutional court’s power to protect fundamental rights.
  • Where an undertrial has suffered a long incarceration and there is no realistic prospect of the trial finishing in a reasonable time, the right to a speedy trial under Article 21 is engaged. At that point, the statutory restriction “would melt down.”
  • This is not a finding on the merits of the accusation. The court is not saying the case is weak. It is saying that indefinite pre-trial detention, with no trial in sight, is itself unconstitutional.

The Oxford Human Rights Hub described Najeeb as reading Article 21 protection into “the stringent UAPA bail jurisprudence,” precisely because it refused to let a statutory bar swallow a constitutional guarantee (OHRH).

Najeeb did not overrule Watali. The two now sit side by side. Watali governs the prima facie inquiry. Najeeb governs what happens when the State takes too long. The two doctrines pull in opposite directions, and the gap between them is where most UAPA bail fights now happen.

The Article 21 delay doctrine, case by case

After Najeeb, the question was no longer “is the case prima facie true.” For long-held undertrials, the question became “how long is too long, and how empty must the trial calendar be.” The Supreme Court has answered that question incrementally, case by case.

Vernon Gonsalves v. State of Maharashtra (28 July 2023). Two Bhima Koregaon accused, jailed nearly five years, were granted bail by a Bench of Justices Aniruddha Bose and Sudhanshu Dhulia. The Court did two things at once. On delay, it held that five years of incarceration with a grave charge but a distant trial justified bail. On the merits, and this was the bolder move, it actually looked at the probative value of the material, and held that mere possession of literature said to propagate violence does not by itself amount to a “terrorist act” under Section 15 of the UAPA (Bar and Bench). Commentators read Vernon as quietly recovering the basics that Watali had pushed out: a court can, at the bail stage, ask whether the material even discloses an offence (Indian Constitutional Law and Philosophy).

Shoma Kanti Sen v. State of Maharashtra (April 2024). Another Bhima Koregaon accused, an academic over 66 years old, jailed almost six years with charges not yet framed. The Bench of Justices Aniruddha Bose and Augustine George Masih granted bail and, importantly, held that the Section 43D(5) bar did not even apply because there was no reasonable ground to believe the accusations were prima facie true (SCC Online). The Court observed that any deprivation of liberty breaches Article 21 unless it is reasonable, follows a fair procedure, and is proportionate. Shoma Sen blended both doctrines: weak prima facie case and long custody.

Sheikh Javed Iqbal v. State of Uttar Pradesh (18 July 2024). This one rested almost entirely on delay. The accused had been in custody more than nine years; charges were framed, but only two prosecution witnesses had been examined (LegalOnus). The Court made a sharp point that pushes back directly on Watali-style reasoning: the more serious the offence, the greater the duty to conclude the trial quickly. Seriousness of the charge cannot be used to make trial delays constitutionally acceptable. The judgment also took care to distinguish Gurwinder Singh, holding it is wrong to say bail can never be granted under a particular statute (LiveLaw).

Jalaluddin Khan v. Union of India (13 August 2024). A retired police constable, accused of renting an upper floor of his house to alleged PFI members, was booked under Sections 13, 18, 18A and 20 of the UAPA. Both the special court and the High Court refused bail. A Bench of Justices Abhay S. Oka and Augustine George Masih granted it, and used the case to restate the principle plainly: “bail is the rule and jail is an exception” applies even under the UAPA, with a modification, so long as the statutory conditions are met (Maktoob Media). Justice Oka warned that “if courts start denying bail in deserving cases, it will violate rights guaranteed under Article 21.” The Court also rapped the investigation for relying on distorted evidence (Lawyer E-News).

Notice the pattern. Each judgment leans on a slightly different lever. Vernon and Shoma Sen reopen the merits door that Watali shut. Sheikh Javed Iqbal turns the gravity argument on its head. Jalaluddin Khan reasserts the constitutional baseline. Together they form a body of law that says the bar is real but not insurmountable.

Recent grants: a quick reference table

A side-by-side reading makes the trend visible. All of these are Supreme Court bail outcomes under the UAPA.

CaseYearCustody at bailPrimary groundOutcome
NIA v. Zahoor Ahmad Shah Watali2019-Prima facie case accepted at face valueBail refused
Union of India v. K.A. Najeeb2021~5 yearsArticle 21, delayed trialBail upheld
Vernon Gonsalves v. State of Maharashtra2023~5 yearsDelay plus no “terrorist act” on the materialBail granted
Shoma Kanti Sen v. State of Maharashtra2024~6 yearsNo prima facie case plus long custodyBail granted
Sheikh Javed Iqbal v. State of U.P.2024>9 yearsDelay alone; gravity cannot justify itBail granted
Jalaluddin Khan v. Union of India2024~1 yearWeak case; “bail is the rule” restatedBail granted
Gurwinder Singh v. State of Punjab2024-Prima facie case found trueBail refused

The arc runs from 2019 to 2024, and the timeline below tracks how the doctrine shifted.

PhasePeriodWhat the Court was saying
The wall2019 onwardsTake the chargesheet at face value; no mini-trial; bar bites hard (Watali)
The crack2021Article 21 can override the bar when trial is delayed (Najeeb)
The widening2023 to 2024Delay grounds plus a cautious return to probing the material (Vernon, Shoma Sen, Sheikh Javed Iqbal, Jalaluddin Khan)
The reassertion2024When the case is genuinely prima facie true, jail remains the rule (Gurwinder Singh)

Gurwinder Singh sits deliberately at the end. It is the reminder that none of this is a free pass.

Bail versus the bar: the tension nobody has resolved

Gurwinder Singh v. State of Punjab, decided in February 2024 (2024 INSC 92), is the counterweight to everything above. A Bench of Justices M.M. Sundresh and Aravind Kumar refused bail to an accused alleged to have received funds from a proscribed organisation. The Court said, in terms, that the conventional idea that “bail is the rule, jail is the exception” finds no place when dealing with bail under the UAPA. Under this statute, the Bench observed, “jail is the rule and bail an exception” (Verdictum).

So within the same year, the Supreme Court said both things. Jalaluddin Khan said bail is the rule even under the UAPA. Gurwinder Singh said jail is the rule under the UAPA. Civil liberties groups asked, with some justification, whether the Court was setting the clock back (PUCL).

The contradiction is more apparent than real, and understanding why is the key to the whole subject. The two lines are answering two different questions:

  • When the case against you is prima facie true and the trial is moving, Gurwinder Singh governs. The bar applies, and jail is the rule. This is the Watali plane.
  • When the case is weak, or when you have already spent years inside with no trial in sight, Najeeb and its progeny govern. Article 21 overrides the bar, and bail becomes available. This is the constitutional plane.

The unresolved problem is that nobody has fixed the dial. How many years of custody trigger the Article 21 override? Five, as in Najeeb and Vernon? Six, as in Shoma Sen? Or must it reach nine, as in Sheikh Javed Iqbal? How empty must the trial calendar be? The Court decides case by case, which gives it flexibility but leaves the accused, and the lawyer drafting the bail application, without a bright line. Reading these judgments closely is its own skill, and our guide on how to read a judgment is a useful companion when you are trying to work out which plane your case sits on.

There is a second source of discord, and it sits inside the Supreme Court’s own bench structure. UAPA bail is decided overwhelmingly by two-judge Benches. Watali was a two-judge Bench. Najeeb was a two-judge Bench. Vernon, Shoma Sen, Sheikh Javed Iqbal, Jalaluddin Khan and Gurwinder Singh were all two-judge Benches. When two such Benches disagree, neither can formally overrule the other, and the result is a body of law that reads less like a settled rule and more like a running argument. The Court has, in places, tried to impose order on itself, with one Bench censuring another for not following the Najeeb line (The Federal). But until a larger Bench is constituted to reconcile the planes, the doctrine will keep oscillating between “bail is the rule” and “jail is the rule” depending on which two judges hear the matter.

High Courts have not waited for that reconciliation. Several have shown that bail under the UAPA is possible even on the merits, applying the qualification that the Section 43D(5) bar is attracted only when the chargesheet discloses a prima facie case (The Quint). That is the doctrinal opening Vernon and Shoma Sen widened at the apex level. The lesson for a practitioner is that the merits route and the delay route are both live, and a well-built application argues both.

The practical reality for an accused

Doctrine is one thing. What actually happens to a person booked under the UAPA is another, and it is grimmer than the case law suggests.

Consider the sequence. You are arrested under Chapters IV or VI of the Act. The 180-day clock starts, so you may sit for half a year before a chargesheet is even filed. When the chargesheet comes, Section 43D(5) is triggered, and under Watali the court reads it at face value. Your first bail application, filed early, almost certainly fails, because there has been no time to build the delay argument and the material has not been tested. The special court rejects it. The High Court rejects it. By the time you reach the Supreme Court, years have passed.

This is why almost every successful UAPA bail order you read about involves an accused who has been inside for half a decade or more. The system, in effect, requires the prisoner to earn the Article 21 argument by serving the time. Vernon Gonsalves and Arun Ferreira walked after five years. Shoma Sen after six. Sheikh Javed Iqbal after nine. The conviction data quoted earlier makes the cruelty plain: with a conviction rate around 3 percent, most of these people were never going to be convicted. They served years of a sentence that, legally, was never imposed.

There is a documented pattern here that researchers have flagged. A widely cited study found that the vast majority of UAPA accused are jailed for long periods and eventually acquitted (The News Minute). When the Supreme Court itself describes the conviction rate as “extremely poor” (The Tribune), it is acknowledging that the front-loaded denial of bail has become the real penalty.

It is worth being precise about why this happens, because it is not simply that judges are unsympathetic. The machinery is stacked at each stage. Investigation under the UAPA is slow by design, with the 180-day window and frequent extensions. Trials are slower still, because terror cases involve voluminous material, protected witnesses, electronic evidence and special-court calendars that are perpetually overloaded. The appellate ladder, from special court to High Court to Supreme Court, takes years to climb, and each rung tends to defer to the one below on the prima facie question. So the delay that eventually unlocks the Article 21 argument is not an accident of any single court; it is the cumulative output of a system that moves slowly at every joint. The bitter irony is that the slower the trial, the stronger the bail argument becomes, which means the State’s own inability to prosecute quickly is what eventually frees the accused, long after the cost of custody has already been paid.

For anyone weighing how a stringent special statute treats liberty, the UAPA does not stand alone. The same structural fight, statutory bar versus Article 21, plays out under the NDPS Act, which we cover in NDPS bail: national sovereignty against personal liberty. And the procedural machinery around arrest and bail itself has shifted with the new criminal codes, which we map in the new criminal laws: BNS, BNSS and BSA explained.

The 2026 Delhi-riots batch: the doctrine under stress

The most recent stress test came on 5 January 2026. A Division Bench of Justices Aravind Kumar and N.V. Anjaria decided the bail pleas of seven people accused under the UAPA in the 2020 North East Delhi riots conspiracy case. The Court split the batch. It granted bail to five accused, including Gulfisha Fatima, Shifa ur Rehman, Meeran Haider, Saleem Khan and Shadab Ahmad, treating them as alleged local facilitators. It refused bail to Umar Khalid and Sharjeel Imam, describing them as the alleged “architects” of the conspiracy (NewsOnAir).

The split is the doctrine in miniature. For the five, the Court was willing to treat the case as weak enough, or the custody long enough, to release them. For the two it labelled architects, it found a prima facie case under Watali and applied the Gurwinder Singh logic. Both men had by then spent more than five years in custody without trial.

The human reactions were sharp and attributable. Umar Khalid’s father said simply, “It is unfortunate.” Aakar Patel, Chair of the Board at Amnesty International India, said: “While we welcome the court’s decision to grant bail to their co-accused, it is shameful that Umar Khalid and Sharjeel Imam continue to be denied bail” (Amnesty International). Amnesty noted the two had been “detained for more than five years without trial.”

The story did not end there, and this is the part that shows the system arguing with itself. On 18 May, the Supreme Court expressed strong disapproval of the very January 2026 judgment that had denied bail, observing that the two-judge Bench had failed to properly follow a binding three-judge Bench ruling on UAPA bail (SiasatDaily). One part of the Court was effectively telling another that it had got the Najeeb line wrong. As of mid-2026, fresh bail pleas were being heard. The matter remains live.

That internal disagreement is the honest summary of where UAPA bail stands. The constitutional plane and the statutory plane have not been reconciled. Two Benches can look at the same five years of custody and reach opposite results, and a third can step in to say one of them misread the law.

What to do and what to expect

If you, or someone you act for, faces UAPA charges, the case law translates into a few practical realities. None of this is legal advice for a specific case; it is the shape of the terrain.

Expect the first bail application to be uphill. Filed early, before custody has run long and before the material can be attacked, it faces the full force of Watali. Special courts and High Courts rarely grant UAPA bail early. Plan for that, do not be surprised by it.

Build two arguments in parallel. The first attacks the prima facie case itself, the Shoma Sen and Vernon line: does the chargesheet, even taken at face value, disclose the ingredients of a terror offence? The second is the Article 21 delay argument, the Najeeb and Sheikh Javed Iqbal line: how long has the accused been in custody, how many witnesses have been examined, and is there any realistic timeline for the trial to finish? The second argument strengthens with every passing month.

Document the delay relentlessly. The Article 21 argument is built from the trial court record. The number of witnesses examined, the adjournments, the dates charges were framed or not framed, the projected length of the trial. Sheikh Javed Iqbal succeeded because the record showed nine years of custody and two witnesses. That kind of record is the argument.

Understand which plane you are on. If the evidence is genuinely strong and the trial is moving, Gurwinder Singh is the realistic frame, and the fight is about delay over time. If the evidence is thin, Vernon and Shoma Sen give a route to attack the prima facie case directly. Reading the chargesheet against the statutory ingredients is the first task, and it rewards care.

Expect strict conditions if bail is granted. The recent grants came tied to passport surrender, restrictions on movement, single mobile connections kept active, and in Shoma Sen’s case a direction to keep the phone’s GPS on and paired with an NIA officer. Bail under the UAPA is rarely clean release. It is conditional liberty.

The larger expectation should be sober. The Supreme Court is reshaping UAPA bail, and the direction since Najeeb is unmistakably toward treating prolonged pre-trial detention as a constitutional wrong. But the reshaping is slow, fact-bound, and contested even within the Court. For the person in the cell, the relief, when it comes, often comes years late.

Niyam helps you read the statute and the judgments the way a court does, mapping Section 43D(5) against Watali, Najeeb and the 2024 to 2026 line so you can see which plane a case sits on. Start for ₹100 and put the full UAPA bail record at your fingertips.

Frequently asked questions

Why is bail under the UAPA so much harder than ordinary bail?

Because Section 43D(5) inverts the usual presumption. In ordinary criminal law, bail is the rule and the State must justify keeping you in. Under the UAPA, for offences in Chapters IV and VI, the court is barred from granting bail if it finds reasonable grounds to believe the accusation is prima facie true. After Watali (2019), courts read that material at face value without testing it, which made the bar very hard to clear early in a case.

Did K.A. Najeeb overrule Watali?

No. The two coexist. Watali governs how a court assesses whether the accusation is prima facie true. Najeeb (2021) says that even where the bar applies, a constitutional court can still grant bail under Article 21 when an undertrial has been in custody for a long time with no realistic prospect of a speedy trial. They operate on different planes: one statutory, one constitutional.

How long must someone be in custody before the Article 21 delay argument works?

There is no fixed number. The Supreme Court has granted bail at roughly five years (Najeeb, Vernon Gonsalves), six years (Shoma Sen) and nine years (Sheikh Javed Iqbal). What matters is the combination of long incarceration and a trial that is going nowhere, judged on the facts of each case. That uncertainty is one of the open problems in the doctrine.

Can bail be granted on the merits, not just on delay?

Yes, and this is the more recent shift. In Vernon Gonsalves (2023) and Shoma Sen (2024), the Court looked at the actual material and held that it did not disclose a terror offence, so the Section 43D(5) bar did not apply. The accepted position is that the bar is attracted only if the chargesheet reveals a prima facie case; if it does not, the embargo does not operate at all.

What did Gurwinder Singh decide, and does it contradict Jalaluddin Khan?

Gurwinder Singh (2024) refused bail and said that under the UAPA, “jail is the rule and bail an exception.” Jalaluddin Khan (2024) granted bail and said “bail is the rule” applies even under the UAPA. They are not truly contradictory. Gurwinder Singh was a case where the Court found the accusation genuinely prima facie true; Jalaluddin Khan was a case where the Court found the case weak and the investigation flawed. Same statute, different facts, different planes.

What happened with Umar Khalid and Sharjeel Imam in 2026?

On 5 January 2026, the Supreme Court granted bail to five co-accused in the Delhi riots conspiracy case but refused it to Umar Khalid and Sharjeel Imam, calling them alleged “architects” of the conspiracy, despite both having spent over five years in custody. In May 2026 a different Bench criticised that January order for failing to follow a binding larger-Bench ruling on UAPA bail, and fresh pleas were being heard. The matter remains unresolved.

Are there always conditions attached to UAPA bail?

In practice, almost always. Recent grants came with passport surrender, movement restrictions within a State, a single active mobile connection, and, in Shoma Sen, an order to keep the phone’s GPS on and paired with an investigating officer. UAPA bail is conditional liberty, not unconditional release. For how bail conditions and anticipatory bail work more generally under the new codes, see our guide on anticipatory bail and BNSS Section 482.