New York City Federal Criminal Defense Lawyers

Federal Criminal Defense in New York City

The case is not decided at trial. In the Southern and Eastern Districts of New York, the case is decided in the months before the indictment is unsealed, in a series of conversations that most defendants do not know are occurring until the conversations have already concluded. A target letter arrives, or a subpoena, or a visit from agents who present themselves as merely curious, and the period that follows is the only period in which the defense possesses genuine advantage. Once the grand jury returns its indictment, the architecture of the prosecution is complete. What remains is a question of degree.

Most firms that advertise federal criminal defense in this city describe the process as it appears after charges have been filed. They catalogue the offenses. They explain the guidelines range. They reassure the reader that trial remains an option. The reassurance is accurate, and it does not address the problem. Federal prosecutors in Manhattan and Brooklyn have already assembled wiretaps, financial records, cooperator testimony, and the kind of digital evidence that no individual can prevent from being generated, all organized into a case they would not have brought if they were not prepared to see it through.

The Proffer and Its Consequences

The proffer session, sometimes referred to as a “queen for a day” agreement, is the mechanism through which a defendant offers information to the government in exchange for the possibility of cooperation credit. The word “possibility” carries the weight of the entire arrangement. A proffer agreement is not immunity. It is a limited waiver: the government agrees not to use the defendant’s own statements in its direct case at trial, though it retains the right to use those statements for impeachment, to derive investigative leads from them, and to introduce them if the defendant later offers testimony or argument inconsistent with what was said in the room.

In the Southern District, prosecutors expect cooperators to disclose all criminal conduct across the entirety of their lives. The SDNY cooperation framework, which has operated this way for decades, requires the defendant to plead guilty to the conduct revealed during cooperation, and to do so before any cooperation credit has been earned. The defendant is, in effect, confessing completely, with no assurance that the confession will be rewarded. The government retains sole discretion over whether to file the 5K1.1 motion for a downward departure under the Sentencing Guidelines. If it declines, the defendant has already entered a guilty plea, already disclosed everything, and already forfeited the ability to present defenses that would contradict the proffer statements. The framework rewards candor. It also punishes it, in cases where honest disclosure reveals conduct the defendant could otherwise have concealed.

The question is not whether to cooperate. The question is whether what you know is worth what you will be required to admit.

A client called on a Wednesday last spring, six days after receiving a target letter from an AUSA in the securities fraud unit. The client had spoken to no one. That restraint, which felt to the client like paralysis, was the single best decision of the case. We reviewed the target letter, assessed the likely scope of the investigation based on the statutes cited and the records already subpoenaed, and determined that the client’s exposure was limited to conduct that overlapped with a codefendant’s scheme but did not originate with our client. We presented that analysis to the government before the grand jury convened. No indictment followed.

Not every pre-indictment intervention produces that result. In most of the cases we have observed, though the sample is not large enough to constitute proof, the government proceeds regardless of what defense counsel presents. But the intervention itself is only possible when the client has not already destroyed its value by speaking to agents, producing documents without counsel, or cooperating with parallel civil inquiries that feed directly into the criminal investigation. The civil inquiry and the criminal investigation are, in practice, the same investigation. Securities regulators and federal prosecutors share information, share witnesses, and share strategy. A deposition given in a civil enforcement action becomes material in the criminal file, and the defendant who answered those questions under oath without understanding that a parallel criminal inquiry was underway has already constrained the defense in ways that are difficult to undo once counsel is retained.

And the timing of cooperation, if cooperation is the correct decision, determines its value almost entirely. Information provided before the grand jury convenes is worth more than information provided after indictment, which is worth more than information provided after a codefendant has already told the government the same facts. By the time most defendants retain counsel and begin to consider their options, the period of maximum value has already begun to contract.

What Changed in November 2025

The 2025 amendments to the Federal Sentencing Guidelines, which took effect on November 1, eliminated the formal departure framework that had governed federal sentencing since United States v. Booker rendered the guidelines advisory in 2005. The three-step sentencing process has been simplified. Courts now calculate the guidelines range and then consider variances under 18 U.S.C. § 3553(a) without the intermediate step of assessing guideline-specific departures.

For defendants in drug trafficking cases, the amendments introduced a cap on the offense base level for individuals who qualify for a mitigating role adjustment. The practical effect is that low-level participants, the drivers and couriers and lookouts who are drawn into conspiracy charges carrying mandatory minimums, may now receive sentences that bear some closer relation to their actual involvement. Whether this holds in practice depends on the judge assigned to the case. The cap is a ceiling, not a floor.

The amendments also resolved a circuit split concerning whether a traffic stop constitutes an “intervening arrest” for purposes of criminal history calculation. It does not. This is a narrow point, but for a defendant whose criminal history category determines whether the guidelines range begins at forty-six months or fifty-seven, narrow points are the only kind that matter.

The HALT Fentanyl Act, which permanently scheduled fentanyl-related substances under Schedule I, has prompted the Sentencing Commission to propose further amendments to the drug quantity tables for the 2026 cycle. The proposed changes would address sentencing enhancements specific to fentanyl offenses, a subject on which the Commission has received considerable public comment. The comment period closed in February.

The Two Courthouses

The Southern District of New York, headquartered at 500 Pearl Street in Manhattan, and the Eastern District of New York, at 225 Cadman Plaza East in Brooklyn, prosecute federal crimes in the same city under the same statutes with measurably different institutional cultures. The distinction is, if we are being precise, not a matter of competence. Both offices employ prosecutors who are among the most capable in the federal system. The distinction is one of emphasis and appetite.

SDNY has constructed its identity around white collar prosecution. Securities fraud, public corruption, and the kind of complex financial scheme that consumes years of investigation before the first subpoena is served: these are the cases that define the office and that its prosecutors select for career advancement. The Assistant United States Attorneys in SDNY come from the same pool of elite law school graduates who might otherwise have entered private practice at firms that bill at rates exceeding a thousand dollars per hour, and the institutional incentive structure rewards the prosecution of high-profile, complex cases that generate professional capital for the individual attorneys who bring them. EDNY handles a broader docket. Organized crime, drug trafficking, firearms offenses, and international smuggling constitute a larger share of its caseload, and the prosecutorial culture is more workmanlike, less invested in the profile of the case and more invested in its resolution.

The judge assignment is random in both districts. One cannot select the judge. The wheel turns, a name appears, and the sentencing tendencies, evidentiary preferences, and procedural habits of that particular judge become the conditions under which the entire case will proceed. Some judges in SDNY sentence fraud defendants well above the guidelines range. Some judges in EDNY grant suppression motions that other judges in the same building would deny without hesitation. The variation is substantial.

I am less certain about the sentencing differential between the two districts than the published comparisons suggest. The numbers indicate that SDNY imposes longer sentences for comparable white collar offenses, but the comparison is difficult to construct cleanly because SDNY selects different cases, with different loss amounts and different procedural histories. The raw data does not control for these variables, and I would not want a client to choose between districts on the basis of an aggregate statistic that may obscure more than it reveals.

Digital Evidence and the Fourth Amendment

The Second Circuit held in United States v. Harry that continuous pole camera surveillance of a building for fifty days does not constitute a search under the Fourth Amendment. The implications for defense practice in these districts are immediate: law enforcement can monitor a location indefinitely, without a warrant, using technology that records every person who enters and leaves, every vehicle that arrives, every pattern of activity that, over weeks, assembles into a portrait of association and habit that no single observation could have produced. The pole camera functions the way a landlord’s ledger once functioned in a building where every tenant owed something: technically a record of comings and goings, operationally a map of vulnerability.

Carpenter v. United States established that cell site location information is protected by the Fourth Amendment and requires a warrant. But Carpenter left open the boundaries of its own reasoning, and the Supreme Court has now agreed to hear a case involving geofence warrants, which compel technology companies to identify every device present within a defined geographic area during a specified time period. The Fifth Circuit declared such warrants unconstitutional. The Fourth Circuit reached a different conclusion. Whether a geofence warrant constitutes the kind of general warrant the Fourth Amendment was drafted to prohibit, or whether it is a permissible tool subject to the particularity requirement, is a question that will reshape federal criminal practice once the Court resolves it.

For a defendant in the Southern or Eastern District, the practical question is less doctrinal and more immediate: what does the government already possess? The answer, in most cases, is considerable. Location data, communications metadata, financial transaction records, surveillance footage from both government cameras and private systems obtained by subpoena. The volume of digital evidence in a modern federal investigation has altered the relationship between prosecution and defense in ways that the suppression doctrine has not yet addressed. The motion to suppress remains the primary tool for challenging evidence obtained through government overreach, though the doctrine that governs its success was designed for an era in which overreach looked different.

Whether the courts intended this asymmetry is a question worth asking, though not one that yields a practical answer for the client sitting across the conference table.

Supervised Release Under the Current Framework

The 2025 Guidelines restructured the framework for supervised release, the period of government oversight that follows incarceration. Courts are now required to make individualized determinations regarding whether to impose supervised release at all, and if so, for how long and under what conditions. The previous default, which presumed the maximum statutory term for certain offenses, has been removed.

A new provision, U.S.S.G. §5D1.4, encourages courts to reassess conditions early in the supervision term and to consider terminating supervision after one year if the defendant’s compliance and risk profile support that outcome. The Guidelines now formally distinguish supervised release from probation, recognizing that the former is intended to serve rehabilitative rather than punitive ends. The conditions imposed during supervised release, including drug testing, travel restrictions, employment reporting requirements, and for certain offenses restrictions on internet access, function in ways that the defendant experiences as punishment regardless of how they are classified. The legal distinction is noted. The lived experience is something else.

A Consultation and Its Purpose

In 2019, before the most recent wave of sentencing reform and before digital evidence had assumed the role it now occupies in federal prosecution, a defendant facing federal charges in New York City confronted a system that was already formidable. The system has not become less so. The conviction rate in federal court remains above ninety percent. The investigative resources available to the government remain, for practical purposes, without limit. The sentencing exposure for most federal offenses remains severe enough that the distance between a favorable and unfavorable outcome is measured in years of a person’s life.

The federal criminal process rewards preparation that occurs before the process becomes visible to the defendant. The target letter, the agent’s visit, the parallel civil inquiry that shifts in tone: these are the moments at which the case’s direction is established. What follows is a function of what was or was not accomplished in those early weeks.

A first consultation is not a commitment. It assumes no obligation and costs nothing. It is a diagnostic: an assessment of exposure, an evaluation of the government’s likely theory, and a determination of whether intervention at the current stage can alter what comes next. One calls because the alternative, which is to wait, is to begin the defense after the room for defense has already narrowed, and the federal system does not announce when that narrowing has occurred.