The Delhi High Court has declined to grant regular bail to Umar Khalid and Sharjeel Imam in the larger-conspiracy case arising from the February 2020 North-East Delhi riots, holding that the Court could not dilute or modify the specific liberty conditions imposed by the Supreme Court while rejecting their bail pleas in January 2026. A Division Bench of Justice Prathiba M. Singh and Justice Dinesh Bhatt dismissed their appeals against the July 4, 2026 order of the Special NIA Court, bringing the question of prolonged incarceration under the Unlawful Activities (Prevention) Act, 1967 once again into sharp judicial focus.
The order is legally significant because the High Court was not considering the bail applications in an entirely open field. The Supreme Court had already examined the prosecution material against Khalid and Imam in its January 5, 2026 judgment and, while granting bail to several other accused, declined relief to the two appellants. At the same time, the Supreme Court left them liberty to renew their bail applications upon the completion of examination of the protected witnesses relied upon by the prosecution or upon the expiry of one year from the date of its order, whichever occurred earlier. The Delhi High Court treated that specific direction as binding and concluded that judicial discipline prevented it from effectively rewriting the Supreme Court’s terms.
The litigation must therefore be understood against two distinct layers of judicial reasoning. The first concerns the stringent statutory threshold under Section 43D(5) of the UAPA and the Supreme Court’s assessment of the material against Khalid and Imam. The second concerns the constitutional question of prolonged pre-trial incarceration under Article 21 and the extent to which constitutional courts can relax statutory restrictions on bail where the trial itself is substantially delayed. The latest Delhi High Court order principally turns on the first layer as modified by the Supreme Court’s specific directions, even though the second continues to remain an important constitutional issue in the background.
The underlying FIR is FIR No. 59/2020, investigated by the Delhi Police Special Cell, and concerns the prosecution theory of a larger conspiracy behind the communal violence that occurred in North-East Delhi in February 2020. The accused face multiple provisions of the Indian Penal Code along with offences under the UAPA. The allegations are contested by the accused and remain allegations to be tested through the criminal process; the present proceedings concern bail and do not constitute a determination of guilt.
The procedural history has become particularly important because the case has now remained pending for several years. Khalid was arrested in September 2020, while Imam has also remained in custody in connection with the proceedings. Their latest applications were premised substantially upon the continuing passage of time and the fact that the criminal trial had not yet reached the stage of examination of prosecution witnesses. The High Court recorded that arguments on charge had concluded on July 13, 2026, but the framing of charges and commencement of evidence remained important stages in the larger procedural timeline.
The appellants placed three principal developments before the High Court as circumstances warranting reconsideration of their bail. First, reliance was placed upon the Supreme Court’s subsequent judgment in Syed Iftekhar Andrabi v. National Investigation Agency, which dealt with the interface between Section 43D(5) of the UAPA and Article 21. Second, reliance was placed upon interim bail granted by the Supreme Court to co-accused Tasleem Ahmed and Khalid Saifi on grounds connected with the delay in the trial. Third, the appellants relied upon the further period of incarceration following the Supreme Court’s January judgment.
The Bench, however, distinguished these developments from the position specifically recorded by the Supreme Court in relation to Khalid and Imam. The Court noted that Syed Iftekhar Andrabi arose from a different prosecution and therefore did not directly reconsider the findings concerning the present appellants. Similarly, the interim bail granted to Tasleem Ahmed and Khalid Saifi could not, in the High Court’s view, establish parity because the roles attributed to those accused were materially different from the roles attributed to Khalid and Imam in the prosecution case.
This distinction between parity in law and parity in factual role is important in UAPA bail jurisprudence. Bail cannot ordinarily be granted merely because another accused in the same FIR has received relief. Courts examine the specific material and role attributed to each accused, particularly when Section 43D(5) applies. The January 2026 Supreme Court ruling itself adopted a differentiated approach: while five accused were granted bail, Khalid and Imam were denied relief because the Court assessed the prosecution material against them differently. The subsequent Delhi High Court order has essentially preserved that distinction.
The January Supreme Court judgment is therefore central to understanding the present result. The Court had examined the prosecution’s allegations concerning the alleged organisation and coordination of the events and held that Khalid and Imam occupied a position distinct from several other accused. It referred to material which, at the prima facie stage, was said to place them at a higher level in the alleged conspiracy. That assessment was made specifically for determining the statutory bail threshold and must not be confused with a final finding after trial.
Section 43D(5) of the UAPA creates an unusually stringent restriction on bail. Where, upon perusal of the case diary or the report under Section 173 of the Code of Criminal Procedure, the Court is of the opinion that there are reasonable grounds for believing that the accusation is prima facie true, the statutory framework substantially restricts the grant of bail. This makes the bail stage materially different from an ordinary criminal prosecution. The Court is not conducting a full trial, but it must nevertheless undertake a judicial assessment of the prosecution material sufficient to determine whether the statutory bar is attracted.
The constitutional complication arises because this statutory restriction does not exist in isolation from Article 21. The Supreme Court’s three-Judge Bench decision in Union of India v. K.A. Najeeb recognised that constitutional courts retain the power to grant bail where prolonged incarceration and the absence of a reasonably foreseeable conclusion to the trial make continued detention inconsistent with the fundamental guarantee of personal liberty. The Court had distinguished the statutory restrictions on bail from the constitutional obligation to protect liberty where the criminal process itself threatens to become punitive by reason of excessive delay.
It is precisely this relationship between K.A. Najeeb and the later UAPA bail jurisprudence that has generated continuing judicial discussion. In May 2026, a Supreme Court Bench considering Syed Iftekhar Andrabi expressed concern about smaller Benches progressively weakening the constitutional force of a larger-Bench decision without expressly overruling it. The Court specifically identified the interface between Section 43D(5) and Article 21 as an issue requiring consideration by a larger Bench.
The May judgment is particularly relevant because the Supreme Court observed that K.A. Najeeb had critically examined the earlier decision in Zahoor Ahmad Shah Watali and held that the latter could not be understood as a precedent for indefinitely denying bail to an undertrial where the trial has no foreseeable end. The reference to a larger Bench demonstrates that the precise relationship between statutory restrictions and constitutional liberty remains an evolving question rather than a completely settled formula.
Yet the Delhi High Court’s difficulty was narrower. Whatever the larger constitutional debate may ultimately produce, the Supreme Court had already issued a case-specific direction concerning Khalid and Imam. The High Court therefore treated itself as bound not merely by general observations concerning UAPA bail, but by the specific procedural liberty granted to these two accused. The distinction is important: a lower court cannot treat a subsequent legal development in another case as an implied modification of an operative direction issued by the Supreme Court in the very case before it.
The Bench had, in fact, explored whether the appellants would be willing to wait until the first week of January 2027 before pressing the appeals. The counsel appearing for the appellants declined that course and sought adjudication of the appeals at the present stage. The High Court consequently proceeded to examine the fresh grounds rather than disposing of the matter merely on the basis that the specified period had not elapsed.
This aspect makes the order more nuanced than a simple statement that the High Court considered itself powerless to examine bail. The Court did consider the new circumstances. What it ultimately held was that those circumstances did not permit it to circumvent or modify the specific conditions imposed by the Supreme Court. In other words, the High Court’s reasoning was rooted in the hierarchy of courts and the doctrine of judicial discipline rather than a conclusion that prolonged incarceration can never constitute a ground for bail under the UAPA.
The reliance upon the interim bail granted to Tasleem Ahmed and Khalid Saifi presented another important issue. Interim bail based upon delay or individual circumstances does not automatically establish a general precedent that every accused in the same case is entitled to identical relief. Bail remains substantially role-sensitive. The prosecution’s case against each accused must be examined independently, particularly in a conspiracy prosecution where the alleged contribution of different participants may operate at different levels. The High Court therefore declined to treat the relief granted to the two co-accused as creating an automatic parity claim for Khalid and Imam.
The High Court also referred to its earlier decision concerning Athar Khan, another accused in the same FIR, whose bail was rejected after the Court examined the role attributed to him. That judgment is itself pending consideration before the Supreme Court, where notice was issued in August 2026. The reference illustrates the unusually interconnected nature of bail jurisprudence arising out of this single FIR: decisions concerning one accused can generate arguments about parity, precedent and changed circumstances for others, while differences in individual roles may simultaneously prevent identical treatment.
There is, however, a broader constitutional concern that cannot be overlooked. By September 2026, the case has entered its seventh year from the commencement of the relevant criminal proceedings, while the accused continue to remain undertrial. The Supreme Court itself has recognised in K.A. Najeeb that the UAPA’s stringent bail framework cannot be interpreted as permitting incarceration without meaningful regard to Article 21. The right to a speedy trial is not a procedural luxury; it forms part of the constitutional guarantee of life and personal liberty.
The difficulty lies in translating that principle into a workable judicial test. A court cannot grant bail merely because a prosecution under a special statute has taken time. Complex conspiracy prosecutions involving numerous accused, voluminous evidence, protected witnesses and multiple connected proceedings inevitably require greater time than ordinary criminal trials. At the same time, complexity cannot become an indefinite justification for pre-trial incarceration. The constitutional question is therefore one of proportionality between the legitimate needs of prosecution and the individual’s continuing deprivation of liberty.
The Supreme Court’s January 2026 judgment attempted to address this tension by distinguishing between the alleged roles of different accused and by directing expedition of the trial. For Khalid and Imam, it nevertheless prescribed a specific point at which bail could be renewed. The Delhi High Court’s present order effectively treats that mechanism as binding until the conditions specified by the Supreme Court are satisfied or altered by the Supreme Court itself.
This brings the matter back to the doctrine of judicial discipline. A High Court is constitutionally subordinate to the Supreme Court, and Article 141 gives binding force to the law declared by the Supreme Court. More fundamentally, the administration of justice requires that a lower court should not indirectly neutralise a specific order of the apex court by re-characterising a subsequent judgment in another matter as a change in law. If a party believes that the Supreme Court’s earlier order is inconsistent with a later constitutional interpretation, the appropriate course is to seek clarification, review or reconsideration before the Supreme Court.
That does not mean that judicial discipline is a barrier to identifying constitutional inconsistency. The May 2026 reference itself demonstrates that Supreme Court Benches can recognise tensions in precedent and place them before a larger Bench. The institutional distinction is that such reconsideration must occur through the appropriate judicial channel. A coordinate or subordinate court cannot independently rewrite a binding direction.
The latest ruling consequently leaves an unresolved but increasingly important constitutional question in the background: how should courts balance a stringent statutory bail bar against Article 21 when an undertrial has remained incarcerated for years and the trial has not reached a meaningful evidentiary stage? The answer may ultimately emerge from the larger Bench considering the perceived conflict in the UAPA bail jurisprudence. Until then, the Supreme Court’s case-specific directions continue to govern Khalid and Imam.
The case also demonstrates why bail jurisprudence under special criminal statutes cannot be reduced to the slogan that “bail is the rule, jail is the exception” without engaging with the statutory text. That principle remains important, but Parliament has deliberately imposed a higher threshold in UAPA prosecutions. Constitutional courts must therefore operate simultaneously within two legal commands: the statutory restriction under Section 43D(5) and the constitutional prohibition against unjustified or disproportionate deprivation of liberty.
Conversely, the existence of a stringent statutory threshold should not convert the bail stage into a premature trial. Section 43D(5) requires a prima facie assessment, not a final determination of guilt. The prosecution’s material must be assessed sufficiently to decide whether the statutory bar operates, but the ultimate credibility of witnesses, admissibility and probative value of evidence, and the competing factual narratives remain matters for trial. The latest Delhi High Court order does not determine the guilt or innocence of either accused.
The distinction is particularly important in a case involving allegations of conspiracy. Conspiracy cases often rely upon a combination of communications, meetings, statements, conduct and circumstantial material. The legal challenge at trial will ultimately concern whether those individual pieces establish the necessary agreement and participation beyond reasonable doubt. Bail adjudication operates at a different evidentiary level and should not be mistaken for final adjudication.
The refusal of bail also does not extinguish the constitutional concern over the pace of the trial. The High Court has recorded that arguments on charge concluded in July 2026, while the prosecution and defence remain engaged in the larger process. The Supreme Court’s earlier direction for expeditious proceedings consequently assumes continuing importance. The legitimacy of prolonged detention under a stringent anti-terror statute is inseparable from the State’s ability to bring the accused to trial within a reasonable period.
Ultimately, the September 30 order is less a fresh pronouncement on the merits of the Delhi riots prosecution than a reaffirmation of the hierarchical structure of judicial decision-making. The Delhi High Court has not held that prolonged incarceration can never justify bail under the UAPA, nor has it declared Article 21 subordinate to Section 43D(5). It has held that, in the peculiar circumstances of these two cases, it cannot disregard or modify the specific directions issued by the Supreme Court concerning when their bail pleas could be renewed.
The next significant development may therefore come not from another routine bail application, but from the larger constitutional question already placed before the Supreme Court concerning the relationship between K.A. Najeeb, the UAPA’s statutory bail restrictions and subsequent decisions applying them. Until that question is authoritatively resolved, the present order illustrates a fundamental principle of appellate hierarchy: constitutional liberty remains the governing value, but the manner in which that value is judicially reconsidered must itself respect the discipline of precedent, institutional hierarchy and the binding force of orders passed by the Supreme Court.
