On 3 September 2026, the leader of Guyana’s political opposition walked into a magistrates’ court in Georgetown to fight for the right to stay in his own country. Azruddin Mohamed, the wealthy businessman who finished second in Guyana’s 2025 presidential election, is wanted by the United States on gold-smuggling, tax-evasion, money-laundering and fraud charges. A magistrate is now holding six full days of hearings, spread over two weeks, to decide whether he can be committed for surrender to Washington. One week earlier, the Caribbean Court of Justice cleared the first legal roadblock by confirming that Guyana’s home affairs minister had acted within his authority when he signed off on the case. The defence has one card left on the table: a direct attack on the validity of the extradition treaty between Guyana and the United States, an argument the magistrate has deferred but not dismissed. For an international audience, the stakes go well beyond one Caribbean politician. This is a live demonstration of how a common-law committal court tests a superpower’s extradition file page by page, and it invites the mirror question a French lawyer asks instinctively: if the same file landed on a Paris desk tomorrow, with the same charges, the same sanctions history and the same political profile of the requested person, would France surrender him? The answer runs through twelve articles of the French Code of Criminal Procedure and three recent rulings of the Cour de cassation, and it is far less automatic than Washington’s complainants tend to assume.
I. Can the United States extradite Guyana’s opposition leader from Georgetown?
A. What are the US fraud and gold-smuggling charges against Azruddin Mohamed?
The American file against the Mohamed family has been built in public for more than two years. In June 2024, the US Treasury’s Office of Foreign Assets Control designated Nazar Mohamed, his son Azruddin Mohamed, their company Mohamed’s Enterprise and a Guyanese government official, Mae Thomas, under Executive Order 13818, which implements the Global Magnitsky sanctions programme against corruption. The Treasury’s case was that Mohamed’s Enterprise, one of Guyana’s largest gold exporters, had evaded Guyana’s tax on gold exports by under-declaring shipments to the Guyanese authorities, had bribed customs officials to falsify documents, and had defrauded the Guyanese state of tax revenue. Between 2019 and 2023 alone, the scale alleged was vast, and Caribbean reporting on the US charging documents speaks of more than 10,000 kilograms of gold exported between 2017 and June 2024 with export taxes and royalties evaded on the order of 50 million US dollars, moved through falsified customs seals and duplicate export licences. When Azruddin Mohamed launched his We Invest in Nationhood party and its manifesto in 2025, he told journalists that once in office the United States would work with him and lift the measures against him. He finished runner-up in the presidential election of September 2025, and the sanctions stayed exactly where they were.
The criminal case followed the sanctions logic to its natural end. The Associated Press reported on 1 September 2026 that Mohamed is accused of fraudulently undervaluing gold exports to evade more than 50 million dollars in taxes, alongside alleged money laundering and wire and mail fraud. Those four labels matter enormously to an extradition lawyer, because each of them has to survive a double-criminality test in the requested state: the conduct must be punishable on both sides of the ocean, and at a sufficient level of seriousness, before any judge can commit anyone for surrender. Gold-export fraud sounds exotic, but its legal skeleton is ordinary: false customs declarations, understated export values, tax evasion, the laundering of the resulting proceeds, and the use of electronic communications and the postal system to carry the scheme through. Every common-law extradition system, Guyana’s included, will ask the same first question about that skeleton. Is there evidence capable of justifying committal, as if the conduct had happened at home? The Georgetown hearings are that question, argued over six full court days, and the answer will decide whether the file ever reaches the minister who signs the surrender order.
There is a second dimension Washington’s prosecutors cannot switch off. Azruddin Mohamed is not an anonymous businessman. He is the most visible opposition figure in a petro-state whose oil wealth has turned every election into a geopolitical event, and whose 2025 campaign explicitly promised voters that American sanctions would disappear under his presidency. Guyanese commentators close to the government have already framed that narrative as an admission, arguing that Mohamed cannot credibly claim political persecution while explaining his troubles by pointing to American sanctions and an international press investigation. The defence will frame it the other way: a prosecution that began with sanctions designations in 2024, escalated into criminal charges in 2026, and targets the man who came second in a presidential election carries, at minimum, the appearance of a political trajectory. Extradition courts do not decide who is right about the politics. They decide whether the politics infect the request. That distinction, between the motive of a prosecution and the political-offence character of the acts, is the hinge on which the Georgetown treaty challenge turns, and it would be the hinge of a French examination too, as the second half of this article shows.
B. Where do the Georgetown extradition hearings stand in September 2026?
The procedure has already produced one final answer and two live questions. The final answer came from the Caribbean Court of Justice, Guyana’s apex court, which ruled on 1 September 2026 that the home affairs minister had acted within his power when he issued the authority to proceed that opened the extradition case. Mohamed’s lawyers had argued that the minister had no lawful authority to launch the process; the Court rejected that submission and kept the case alive. That ruling is the Guyanese equivalent of what French lawyers call the regularity of the government’s referral: before any judge examines the merits, the court checks that the executive instrument that set the machine in motion was lawfully signed. In Georgetown that check is now over, and the government won it.
The live questions are being argued before Chief Magistrate Judy Latchman in the Georgetown magistrates’ court, where hearings resumed on 3 September 2026. State prosecutor Terrence Williams is leading the evidence for the requesting state, and the court has fixed six all-day sittings running into the following week, a timetable that signals a document-heavy file: customs records, export licences, banking trails and the US charging instruments, all of which must be translated into evidence admissible under Guyanese committal rules. The defence, led by attorney Roysdale Forde, opened with the boldest available move, a submission that there is no valid extradition treaty between Guyana and the United States at all. If that argument succeeded, the entire case would collapse regardless of the strength of the fraud evidence, because extradition without a treaty basis is, in most systems, legally impossible. The magistrate did not throw the argument out. She deferred her ruling on the treaty point and ordered the hearings to continue on the evidence, a case-management choice that keeps both tracks alive: the court hears the fraud file now and decides the treaty objection with the benefit of the complete record.
Behind that scheduling decision sits a second defensive wall. The defence is also relying on an order from Guyana’s Court of Appeal, obtained earlier in the litigation, whose precise scope is now contested between the parties. Procedural layering of this kind is familiar to French extradition practitioners: in Paris, a requested person can challenge the referral before the chambre de l’instruction, seek release, appeal to the Cour de cassation on points of form, and then attack the surrender decree before the Conseil d’Etat, with each court examining a different slice of the file. The Georgetown equivalent plays out across the magistrate, the Court of Appeal and the Caribbean Court of Justice, and each layer buys time while testing a different lock. Time matters here for a structural reason. Extradition committal is not a trial of guilt; the magistrate does not decide whether Mohamed smuggled gold. She decides whether the American evidence, taken at its highest, discloses conduct that Guyanese law treats as extraditable crime, whether the documents are in order, and whether any bar to surrender, from treaty invalidity to political character, blocks the road to the minister’s desk. If she commits him, the file returns to the executive for the surrender decision, which is itself reviewable. If she discharges him, the United States can in most systems ask for the decision to be revisited only through narrow appellate routes. Six days of hearings for that binary outcome is not delay. It is the system working at full pressure, and it explains why the week of 8 September is the decisive week of the case.
One further feature of the Georgetown calendar deserves attention from international readers. The Chief Magistrate is hearing the case in open court, day by day, while the requested person remains one of the country’s best-known public figures and the requesting state is the hemisphere’s dominant power. That configuration, a small-state committing court testing a superpower’s file in public, is exactly the scenario in which extradition law shows its purpose. The treaties were written for ordinary fugitives, but their safeguards, sufficiency of evidence, double criminality, treaty validity, political-offence protection, were designed for cases where power is unequal and the requested person is prominent. Whether Georgetown’s safeguards hold or bend under the weight of this file will be known within days. But the French comparison casts the same events in a sharper light, because French law subjects every one of those safeguards to a judicial examination that is, in several respects, stricter than the common-law committal model. That is the subject of the second part of this analysis.
II. Would France surrender a suspect in Mohamed’s position?
A. Would French courts extradite for foreign tax fraud and money laundering?
Imagine the file transferred to Paris: a foreign government asks France to surrender a resident opposition politician for undervaluing gold exports, evading export taxes, laundering the proceeds and using electronic communications to run the scheme. French law would begin with a question Georgetown never asks, because the answer there is assumed. Under article 696 of the Code of Criminal Procedure, which provides that in the absence of a treaty stipulation to the contrary the conditions, procedure and effects of extradition are set by the Code, and under article 696-2, which states that “Le gouvernement français peut remettre, sur leur demande, aux gouvernements étrangers, toute personne n’ayant pas la nationalité française”, the first lock is nationality. France does not extradite its own nationals, full stop, and article 696-4 repeats the bar in categorical terms: “L’extradition n’est pas accordée : 1° Lorsque la personne réclamée a la nationalité française”. Had Mohamed been a French citizen found in Paris, the American request would end at that line, whatever the evidence. Guyana, like most common-law states, has no equivalent absolute shield for its nationals, which is why the Georgetown court can commit a Guyanese citizen for surrender abroad. That single difference explains more about the two systems than any other: France treats nationality as a jurisdictional wall, while Guyana treats it as irrelevant to committability.
The second lock is the seriousness and double-criminality threshold, and here the Mohamed charges would almost certainly pass, but not without a fight over the tax component. Article 696-3 provides that extraditable acts include all acts punished with criminal penalties by the requesting state’s law, and correctional acts “quand le maximum de la peine d’emprisonnement encourue, aux termes de cette loi, est égal ou supérieur à deux ans”, adding the decisive sentence: “En aucun cas l’extradition n’est accordée par le gouvernement français si le fait n’est pas puni par la loi française d’une peine criminelle ou correctionnelle.” Money laundering, wire fraud and mail fraud all have direct French equivalents punished well above the threshold, so double criminality would be satisfied for the core of the file. The tax-evasion limb is historically the delicate one in Franco-American practice, because fiscal offences were long treated with suspicion by extradition law, but modern French case law accepts surrender for tax fraud where the conduct is punishable under French tax-criminal provisions, and the treaty practice between France and the United States has covered fiscal matters for decades. A French chambre de l’instruction would therefore dissect the American file offence by offence, granting extradition for the counts that mirror French offences and stripping out any count that fails the mirror test, a granularity the Georgetown committal, which tends to treat the file as a block, rarely achieves.
The third lock is the one Mohamed’s Paris lawyers would build their entire defence around: the political dimension. Article 696-4 provides that extradition is refused “2° Lorsque le crime ou le délit à un caractère politique ou lorsqu’il résulte des circonstances que l’extradition est demandée dans un but politique”. French courts interpret that second branch, extradition requested for a political purpose, strictly but seriously: the requested person must show that the prosecution is instrumentalised to neutralise a political opponent, not merely that he happens to be one. An opposition leader indicted for common financial crime two years after being sanctioned, in the middle of an electoral cycle in which he promised voters the sanctions would vanish, would give a French court exactly the factual matrix in which that argument is pleaded, with press coverage, sanctions chronology and electoral calendar exhibited in support. The argument would probably fail on these facts, because the fraud file predates the candidacy and rests on customs documents rather than speeches, and French courts consistently hold that common-law financial crime does not become political because its author enters politics. But it would be heard, examined, and answered in a reasoned opinion, and the Cour de cassation would control the reasoning. The same article offers neighbouring shields worth noting for completeness: no surrender where the acts were committed on French territory, where they have already been finally tried in France, where limitation has run under either law, where the requesting state’s penalty offends French public policy, or where the foreign court lacks fundamental defence guarantees, each of which the chambre de l’instruction must verify of its own motion.
Procedure would then supply the fourth lock, and it is here that the French system shows its most distinctive feature: the requested person is brought “dans les quarante-huit heures devant le procureur général territorialement compétent” under article 696-10, informed in a language he understands, given immediate access to counsel, and told expressly that he may consent or oppose surrender and that “elle a la faculté de renoncer à la règle de la spécialité”. If he opposes, article 696-15 seizes the chambre de l’instruction without delay, and that court delivers a reasoned opinion within one month, an opinion that “est défavorable si la cour estime que les conditions légales ne sont pas remplies ou qu’il y a une erreur évidente”. A negative opinion does not automatically free the person where the prosecutor appeals, but the Cour de cassation polices the effects precisely: on 9 August 2023, in a case where Ukraine sought a man for aggravated embezzlement and money laundering and the Chambéry chambre de l’instruction had issued a negative opinion, the criminal chamber had to decide whether judicial supervision survived the opinion (Cass. crim., 9 August 2023, no. 23-83.192), holding the line that “Par arrêt du 30 mars 2023, la chambre de l’instruction a émis un avis défavorable à son extradition” carries defined procedural consequences the lower court could not improvise around. Cassation review itself is narrow by design, since “Le pourvoi formé contre un avis de la chambre de l’instruction ne peut être fondé que sur des vices de forme de nature à priver cet avis des conditions essentielles de son existence légale”, which means the legal debate must be won before the chambre de l’instruction, not above it. Add the documentary lock of article 696-8, under which every request “est adressée au gouvernement français par voie diplomatique” with the conviction, committal order or arrest warrant plus the applicable foreign statutes, and the release valve of article 696-24, which frees the provisionally arrested person when, thirty days after arrest, none of those documents has reached Paris, and the picture is complete: a French Mohamed would face a slower, more fragmented, more judicialised obstacle course than the Georgetown committal, with a reasoned opinion at its centre and cassation control at its edge.
B. How would France test the fairness of the US trial and sentence?
Suppose the charges pass every lock in the previous section. French law still refuses to surrender a person into a trial that would be fundamentally unfair or a sentence that would be manifestly excessive, and the two controlling precedents are less than three years old. The first is the judgment of 28 March 2023 concerning an Italian request (Cass. crim., 28 March 2023, no. 22-84.382), in which the criminal chamber restated that a requested state party to the European Convention on Human Rights “a l’obligation de s’assurer que celle-ci ne sera pas exposée, dans l’Etat requérant, à un déni de justice flagrant”, a flagrant denial of justice that can result, the Court specified, from the impossibility of obtaining a fresh judicial ruling on the merits after a conviction in absentia. The chambre de l’instruction seized of such a complaint must therefore verify concretely whether the requested person will be retried, and an opinion that skips that verification lacks the essential conditions of its legal existence. Applied to a file like Mohamed’s, where no in-absentia conviction is alleged and the United States affords jury trial with appellate review, the flagrant-denial argument would almost certainly fail, but the method matters more than the outcome: France does not take the requesting state’s justice system on trust, not even America’s, and it demands proof, count by count, that a retrial or a full first trial with defence rights awaits the person surrendered.
The second precedent is even closer to financial-crime extradition, because it arose from a South Korean request built on capital-markets and economic-crime charges carrying life imprisonment (Cass. crim., 6 May 2025, no. 24-85.773). The Court held that “l’extradition doit être refusée lorsque la personne concernée démontre qu’elle est exposée à un risque réel de voir prononcer contre elle une peine manifestement disproportionnée dans l’Etat requérant”, while stressing that the notion requires strict interpretation and can succeed only in very exceptional cases, and that the mere fact that the foreign sentence would be heavier than the French one is legally inoperative. US federal fraud sentencing, with its loss-amount driven guidelines, routinely produces headline figures that shock European observers: a 50-million-dollar loss amount can mechanically generate a guidelines range measured in decades, even for a first offender, before the judge’s discretion trims it. A French chambre de l’instruction would therefore ask the precise question the 2025 judgment commands: considering the nature and gravity of the acts and the conditions under which the foreign court will sentence, is the realistic exposure manifestly disproportionate, or merely heavier than at home? For Mohamed-scale allegations, the honest answer is that the test would likely be survived, because large-scale customs fraud draws severe penalties in France too, but the defence would be entitled to force the court to reason it through on the sentencing memoranda, not on impressions. Note the symmetry with another recent cross-border file our firm has analysed, the Sheikh Hasina extradition case, where the sentence awaiting the requested person, there a death sentence, made the disproportionality and public-policy analysis the centre of the French test rather than a footnote.
Two further French safeguards complete the trial-fairness examination. The specialty rule, of which the requested person must be expressly warned under article 696-26, provides that “Il l’informe également qu’elle peut renoncer à la règle de la spécialité”, which means, in substance, that a surrendered person may only be tried for the facts the surrender covered, unless he waives the protection. If Washington obtained Mohamed for customs fraud and then wanted to add sanctions-evasion or foreign-bribery counts arising from the same enterprise, France’s consent would be needed first, and Paris can say no. And detention pending surrender is itself judicialised: where the prosecutor will not leave the person free, article 696-11 sends him before the first president of the court of appeal for committal under extradition warrant or, where appearance is guaranteed, for judicial supervision or electronic monitoring, while article 696-19 lets him apply for release at any time, with the chambre de l’instruction ruling within twenty days, fifteen where the application comes within forty-eight hours of committal. A French Mohamed would therefore litigate his liberty in parallel with his surrender, exactly as the Ukrainian businessman did in Chambéry in 2023, and each ruling would be reasoned, appealable and published. The Georgetown magistrate controls the evidence; the Paris chambre controls the evidence, the liberties, the specialty perimeter and the sentence preview, all at once.
Conclusion
The Georgetown hearings will answer three questions in sequence, and international observers should watch them in order. First, the treaty question: Chief Magistrate Latchman’s deferred ruling on whether a valid Guyana-United States extradition treaty covers this request is the only argument capable of ending the case without any examination of the fraud evidence, which is why the defence led with it and why the magistrate’s prudence in deferring it keeps the entire file in suspense. Second, the sufficiency question: six days of hearings on customs records, banking trails and charging instruments will show whether the American evidence, taken at its highest, discloses extraditable conduct under Guyanese law, with the 50-million-dollar loss figure and the money-laundering counts doing the heavy lifting. Third, the executive question: committal would return the file to the minister for the surrender decision, itself open to further challenge, while discharge would close the American road unless narrow appellate routes revive it. Around those three questions orbit the political facts no court can pretend away: a sanctioned businessman turned opposition leader, a runner-up finish in a 2025 presidential election, and a campaign promise that Washington would lift his sanctions. French law, applied hypothetically to the same file, would most likely end in surrender for the fraud and laundering counts, but only after a nationality check that Guyana never performs, an offence-by-offence double-criminality dissection, a reasoned ruling on the political-purpose claim under article 696-4, a flagrant-denial review under the 2023 Italian precedent, a disproportionality review under the 2025 Korean precedent, and a specialty perimeter policed by the chambre de l’instruction. That is the French lawyer’s lesson of the Mohamed hearings: extradition is never a conveyor belt between the requesting state’s indictment and the surrender plane, and the wider the gap between the two countries’ power, the more each lock matters. Georgetown is testing its locks this week. Paris would test twelve.