US v. David Minkkinen - Intellectual Property Theft and Fraud Case

PUBLISHED UNITED STATES CO URT OF APPEALS FOR THE FOURTH C IRCUIT No. 23-4443 UNITED STATES OF AMERICA, Plaintiff – Appellant, v. DAVID GERALD MI NKKINEN; SIVA RAMAN SAMBAS IVAM, Defendants – Appel l ee s. Appeal from the Unite d States District Court for the Southern Distric t of West Virginia, at Charleston. Irene C. Berger, Distri ct Judge. (2:22-cr-00163-1; 2:22- cr -00163-2) Argued: May 10, 202 4 Decided: February 26, 2026 Before WYNN, RICH ARDSON, and R USHING, Circuit Judge s. Reversed and remande d by published opinion. Judge Rushing wrote t he opinion, in which Judge Wynn and Judge Richardson joined. ARGUED: Jennifer Rada Herrald, OFFICE OF THE UNITED ST ATES ATTORNEY, Charleston, West Virg inia, for Appellant. Stephen S. Stallings, LAW OFFICES OF STEPHEN S. STAL LINGS, ESQ., Pittsb urgh, Pennsylvania; Rabea Jamal Zayed, DORSEY & WHITN EY LLP, Minneapolis, Minnesota, for App ellee. ON BRIEF: William S. Thompson, United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Charlest on, West Virginia, for Appellant. Nic ole Engisch, DO RSEY & WHITNEY LLP, Mi nneapolis, Minne sota; Susan M. Robinson, T HOMAS COMBS & SPAN N, PLLC, Cha rleston, West Virginia, for Appellee Si va raman Sambasivam.

2 Michael Edward Noga y, SELLITTI, NOGA Y & NO GAY, PLLC, W eirton, West Virginia, for Appellee David Ge rald Minkkinen.

3 RUSHING, Circuit Jud ge: Following a whistleblower complaint in 2016, the Government began investigating Defendants David Minkkinen a nd Sivaraman Samba sivam for potential intellect ual property theft and frau d. After a lengthy in vestigation, the Gover nment indicted them around six years later. By then, two relevant w itnesses had died an d three States had purged documents potentially relevant to the charge d offenses. Citing this missing ev idence, Defendants moved to dismi ss the indictme nt against them, alleging unconstitu tional preindictment de lay in violatio n of the Fifth Amendment’s Due Process Clause. The district court granted Defendants’ motion in part, dismissing ten o f the fourteen count s charged in the Govern ment’s superseding ind ictment. T he district court erred in concluding t hat it would viola te the Due Pr ocess Clause to prosecute Defendant s after the Government’s preindictment delay, which was the result of its lengthy invest igation and not tainted by bad faith. Accordingly, we reverse and remand for further proceedings o n all fourteen counts charged in the superseding indictment. I. A. Between 2009 and 2013, Defendants worked for Deloitte, an interna tional company offering “various fin ancial and advisory services, including u nemployment claims management services.” J.A. 76 1. Relevant here, Deloitte developed and marketed “a proprietary web - base d software platform named Unemplo yment Framework for Automated Claim & Tax Services (‘uFA CTS’),” which it used “to process unemplo yment

4 insurance claims for sta te agencies” in Minnes ota, Massachusetts, and New Mex ico. 1 J.A. 761. Minkkinen was Deloitte’s unemployment insurance practice l eader, and Samb asivam was the “lead syste m architect” for uFACT S. J.A. 762. In 2013, Def endants left Deloitte to become s enior partners in the ne wly launched unemployment insuran ce practice of another firm, Sag itec Solutions, LLC. About two years later, Sag itec won a bid to design, develop, a nd implement a feder ally funded unemployment insuran ce claims software system for Maryland and West Virginia. The parties call this project the Maryland– W est Virginia Consortium, or the Consortium. Minkkinen wa s Sagitec’s prim ary point of c ontact for the proje ct, and state emp loyees worked with Sagitec an d other subcontractors t o develop and install th e new software. Starting i n 2016, mu ltiple West Vir ginia state employees bega n “notic[ing] suspicious references to Massachus etts, New Mexico, and Deloitte” in Sagitec’s project materials. J.A. 768. One employee reported t he suspiciou s reference s to West Virginia’s Commission on Speci al Investigations, “alleging possib le wrongdoing by Sagitec, to include the misappropr iation of Deloitte in tellectual property.” J. A. 768. That summer, the United States Attorney ’ s Office, the United States Department of Labor, the Maryland Attorney General’s Office, and Deloitte learned about the whistleblower’s compl aint. T he Maryland Attorney General ’s Of fice sent Defendants letters relaying the whistleblower’s concerns a nd asking specific questions about Sag itec’s ownership of the materials being used for the Consortium. Defenda nts consi stently denied 1 Deloitte’s work with Minnesota stemme d from its predecessor’s contract with t he State.

5 wrongdoing. Investiga tors also me t with Deloitte represen tatives. Those representative s stated that materials fro m previous consulting projects—i.e., the materials Defendants and Sagitec were allegedly using for their Consortium work — were not in the public do main. They also claim ed that the States had “no ownership rig hts” in the software Deloitte develop ed for them; instead, the S tates held only “a license to use the software in perpetuity.” J.A. 447. On February 3, 2017, the United States form ally opened an investigation. Fro m 2017 to late 2019, the Government “conducte d approximately 31 interviews and review ed [a large number of] documents.” U nited States v. Minkkine n, 678 F. Supp. 3d 778, 793 (S.D. W. Va. 202 3); see id. at 788 (acknowl edging Government’s representation tha t it “conduct[ed] a protrac ted and complex investigation involvi ng voluminous discov ery [and] witnesses in mult iple states”). 2 Relevant here, in August 2017, in vestigators identified Bernt Peterso n “as a source of design documents fr om Deloitte’s New Me xico and Massachusetts project s that Sagitec possessed in connectio n with the Consortium project.” Id. at 785. P eterson had worked on the Minnesota p roject as a state employe e and the New Me xico project as a Deloitte employee. He joined Sagitec in 2014. After j oining Sagitec, Peterso n allegedly circulated the Deloitte design doc uments to other Sagitec employees. 2 T he district court noted that the Governme nt “did not supply exhibits or testimony from investigators detailing the investigati ve timeline.” Minkkinen, 678 F. Supp. 3d at 793. The court therefore “ relie[d] on the repres entations in [the Gov ernment’s] brief, in combination with the other evidence subm itted.” Id. On appeal, Defendants do not object to the district court’s re liance on the Government’s represen tations about the investigation.

6 About a month later, i n September 2017, inv estigators learned of Philip Tackett, another former Deloitte employee w ho worked on the New Mexico project. Investigator s became interested in Tackett after discoverin g his name in the metadata of hundreds of Consortium docum ents, which investigators suspected had been take n from Deloitte. At the time investigators i dentified Tackett, he worked at IBM Corp oration. Tackett eventually joined Sagitec in May 20 19, and investigators int erviewed him under a grant of i mmunity in August of that year. 3 Several statements from the inter view are relevant here. First, Tackett stated that when Deloitte’s projects with state entities end ed, “it’s com mon that the software a nd all the delivera bles, meaning the u se cases, the test cases, ... training documents, all those th ings, they become prop erty of the State.. . . So New Mexico o wned all” the m aterials generated by Deloitte for the New M exico project. 4 J.A. 646. But when asked in a follow - up question whether New Mexico owned the materials from its project or whether the State was instead “lic ensed to use” them, Tackett responded: “Th at I don’t know.” J.A. 646. Tackett clarified that he “certainl y had no expectation or unde rstanding” that he — an employee —“owned an y of that.” J.A. 646. Next, when aske d why his name ap peared in some of the documents’ meta data, Tackett explained that several docum ents dealt with subjects he had never worked on and that other docum ents implausibly indica ted he “printed” them during the ti me he was at 3 An excerpted transcript of Tackett’s inte rview is in the record. See J.A. 646–650. 4 A “use case” is a “wr itten document” that “ describes how . . . [a] software system. .. is going to functio n.” J.A. 868. It can have “diagra ms, schematics, [and] screen captures.” J.A. 868. Once use cases are finalized, programme rs use them to write softwar e programs.

7 IBM— not Deloitte or Sagitec. J.A. 6 50. Further, when asked about similarities betwee n some Deloitte and Sa gitec materials, Tackett claimed th at the similarities reflected Minkkinen’s preferences and meth ods for man aging a project. He also remar ked that some of the materials with D eloitte’s name in the m etadata “[c]ould ha ve” come from the N ew Mexico project, but he maintained t hat he “really [couldn’t] say” one way or the other. J.A. 647. A few m onths after Tackett’s interview, i n Dece mber 2019, th e Government informed Sagitec that it was the ta rget of a cri minal investigation. After a proffer from the Government in Februa ry 2020, Sagitec agreed to cond uct its own internal inv estigation. Meanwhile, in March 2020, Tackett “pa ssed away unexpectedly. ” J.A. 652. Sagi tec’s internal investigation concluded in October of that year, and it “prompted additional questions and leads” fo r the Government to pu rsue. J.A. 663. In November 2020 — o ne month after Sagitec finished its internal investigation — investigators interview ed Peterson. 5 Peterson confirmed wor k ing o n the Massachusetts and New Mexico proje cts, as well as writing use cases and supervisi ng others who wrote use cases while worki ng on the Con sortium at Sagitec. He stated t hat he did not re member any “references to othe r states” in the material s he worked on. J.A. 630. But he did a dmit to taking use case doc uments from the New Mexico (and po ssibly the Massachusett s) 5 Peterson’s intervie w was recorded, but the recording w as not submitted to t he district court. Thus, like the district court, we “rel[y] on the Memo randum of Interview and synopsis” in the re cord, Minkkinen, 6 78 F. Supp. 3d at 78 5 n.4, although we note t he synopsis “ does not pur port to be an exact transcription of what was s aid during the course of the interview,” J.A. 626.

8 project when he left Deloitte. He explained that he used the old us e case materials o nly as a template for the Consortium’s wor k. In other words, he claimed tha t he would scrub the documents’ contents w hile retaining the form at, and then tell his su bordinates to use the formatted document to create a use case for the Consortium. A s far as he was aware, no one had access to the substance of the u se case materials he b r o ught from Deloitte. And he had “no idea” if anyone else took any materials from Deloitte. J.A. 634. Peterson stated that use cases are inhere ntly individualistic and thus of no use to compe titors, and he opin ed that the materials he took from Sagitec when he went to work at another firm were “Sagitec’s property” or “Sagitec’s project.” J.A. 632. After the Gov ernment made a proffer to Peterson abou t his potential criminal exposure, Peters on agreed to conduct a follo w - up interview sched uled for December 15, 2020. On December 1 3, 2020, however, Peterson “died unexpect edly.” J.A. 636. That same month, the Government served Mi nkkinen with a target letter. I n March 2021, the Government interviewed Sambasivam, served hi m with a target letter, and made a proffer to Minkkinen. In April, it made a proffer to Sambasiv am. Over the months that followed, the parties “e ngaged in several additional communications,” and the Governme nt continued to review ev idence, including ne wly produced documen ts. Minkkinen, 678 F. Supp. 3d at 787. T he last witness inter view occurred on July 21, 20 22, about on e month before the indictment was returned. B. Defendants were indicted on August 2 3, 2022. Both were charged with misappropriating tra de secrets and conspirin g to misappropriate tra de secrets in violation

9 of 18 U.S.C. § 1832(a) (1), (2), (3), and (5). In addition, Minkkinen was individually charged with wire fraud in violation of 18 U.S.C. § 1343 and se veral counts of false statements in violatio n of 18 U.S.C. § 1001. Sambasivam was also individually charged with two counts of fals e statements. Following the indictm ent, Defendants subpoenaed various th ird parties, including Deloitte, Sagitec, and the States that Deloitte had contracte d with to develo p unemployment softwar e solutions. Notable here, the relev ant Minnesota agency responded with a letter stating Defendants sou ght documents from a project tha t concluded in 2008 and had since been “re peatedly modified.” J. A. 142 – 143. A ccordin g to the agency, its “designated retention period for m ost [documents] is 11 years or shorter, with many [documents ] retain ed f or 3, 6, or 7 years.” J.A. 143. “The retentio n period for contracts and related documents is 6 years.” J.A. 143. The agency explaine d that “in connection with” the end of each retention per iod and “specifically, the [age ncy’s office] mov e completed in May 2022, concerted efforts were made” to destroy any documents n o longer with in the applicable retention period s. J.A. 143. “As such, as of May 2022, d ocuments prior to 2011 and man y others to which shor ter retention periods a pplied, we re securely destroyed.” J.A. 143. T hat said, the Minnesota agency explained that there was “cu rrent staff with knowledge of or involvement in ” the relevant unemploym ent insurance proj ect and produced one hu ndred pages of docu mentation relevant to Defendants’ reque sts, including various contr act components, so me executed and some not. J.A. 143–144. Though the record is less clear about docum ents related to the Ne w Mexico and Massachusetts projects, neither S tat e w as able “ to provide full resp onses” to Defendants’

10 subpoenas “because many docume nts and communications were n o longer available. ” Minkkinen, 678 F. Sup p. 3d at 787–788. After the Government clarified that its tria l evidence would focus on alleged trade secrets related to Massachusetts uFACTS sourc e code and Massa chusetts and New Mexico uFACTS project design documentation, Defendants m oved to dismi ss the indictment on grounds of unconstitutional preindict ment delay. They argued that the Government ’s preindictment delay prejudiced thei r defense because the delay resulted in the unavailability of key testimony from Peterson and Tackett, as well as material documents from state agencies. The Govern ment opposed the moti on. Before the district co urt ruled, the Govern ment filed a supersedi ng indictment, dropping the count for substantive trade se cret theft and adding s everal wire fraud counts. All said, the superse ding indictment c harged both Defenda nts with conspiracy to misappropriate trade se crets in violation of 18 U.S.C. § 1832(a)(5); conspiracy t o commit wire fraud in violation of 18 U.S. C. §§ 1343, 1349; and several counts of aiding and abetting wire fraud in violation of 18 U.S.C. § § 2, 1343. Minkkinen was also charged with seven counts of making false statements in viol ation of 18 U.S.C. § 1001, and Sambasivam was charged with two. In the main, the s uperseding indictment alleged that before, during, and after transitioning from Deloitte t o Sagitec, Defendants illegally “ copied, downloaded, obtained, and transmitted numerous D eloitte files, including highly pr oprietary informatio n such as uFACTS source code, data, and use ca ses,” without Deloitte’ s authorization. J.A. 763. And those “files were subsequently use d by numerous Sagitec employees to desi gn,

11 develop, and market” t he software Sagitec wa s hired to develop fo r the Consortium. J.A. 763. After the superseding indictment, the distri ct court received supple mental briefing from the parties and then granted Defendants’ motion to dismiss in part. The court found that Peterson’s a nd Tackett’s testimo ny would have been “highly significant” and “valuable” to the defen se. Minkkinen, 678 F. Supp. 3d at 791–79 2. It also c oncluded that the documents destroyed by the state agenci es would have been “highly valuab le” and “highly probative.” Id. at 792. In co mbination, t h e missing evidence “could c hange the outcome of a trial.” Id. at 793. T he district court foun d that the Government’s “prolo nged investigation” was the reason for the delay but that the Government offered “litt le explanation for the length of the investigati on.” Id. at 794. Judging the Government’s “justification for the length of the investigatio n” to be “ insufficient in light of the extent of the prejudice caused by the delay,” the court concluded that prosecution o n the affected counts would violate th e Due Process Clause. Id. I t therefore dismis sed ten of the fourteen counts charged in the s uperseding indictment. See id. at 795–797. The Government timely appealed, and we hav e jurisdiction under 18 U.S.C. § 3731. II. “ [S] tatutes of lim itations. .. pr ovide ‘ the primary guara ntee [] agai nst.. . overly stale criminal charges. ’ ” United States v. Lovasco, 431 U.S. 783, 789 (1977) (qu oting United States v. Mar ion, 404 U.S. 307, 322 (1971)). Even so, the Supreme Court has instructed that “ the Due Process Clause has a limited role to play in protecting against oppressive [preindictm ent ] delay. ” Id. It is “ possible for an indict ment within the statute

12 of limitations to raise due process conc erns,” but “a defendant raising such a challenge faces a high burden.” United States v. Pa lmer, 159 F.4th 221, 226 (4th Cir. 2025). This Court evaluates claims of unconstitutio nal preindictment delay under “a two - pronged inquiry.” United States v. Uribe -Ri os, 558 F.3d 347, 358 (4t h Cir. 2009). “First, we ask whether the defendant ha s satisfied his burden of proving ‘actual prejudice’” fr om the delay. Id. (quoting United State s v. Automated Med. Lab’ys, In c., 770 F.2d 399, 403 (4th Cir. 1985)). “‘Th is is a heavy bur den because it requir es not only tha t a defendant show actual prejudice, as opposed to mere speculative prejudic e, but also that he show that any actual prejudice w as substantial — th at he was meaningfully impaired in his ability to defe nd against the state’s charges to such an extent that the dispos ition of the criminal proceeding was likel y affected.’” U nited States v. Shealey, 6 41 F.3d 627, 633 – 634 (4th Cir. 2011) (quoting Jon es v. Angelone, 94 F.3d 900, 907 (4th Cir.1996)). “ Second, if that thresho ld requirement is met, we consider the govern ment’s reasons for the delay, ‘balanc ing the prejudice to the defendant with the Government’s justification for delay.’” Uribe- Rios, 558 F.3d at 3 58 (quoting Automated Med. Lab ’ ys, 7 70 F.2d at 404). “‘The basic inquiry’” un der this pron g is “‘whether the government’s action in prosecuting after substantial delay violat es fundamental conceptions of justice or the community’s sense of fair play and decency.’” Jones, 94 F.3d at 9 04 (quotin g Howell v. Barker, 904 F.2d 889, 895 (4th Cir. 1990)); se e also id. at 910. We need only address the second prong of th e analysis to resolve thi s appeal. On the first prong, Defendants claim prejudice from the loss of two kinds of evidence: (1) Peterson’s and Tac kett’s testimony, and (2) the state agency documents allege dly

13 relevant to past Deloitte projects. We can accept for the sake of a rgument the di strict court’s fi nding that the loss of this evidence actual ly and sub stantially prejudiced Defendants’ defense. 6 For even if that’s the case, the district court err ed in concluding that the Government’s action in prosecuting Defe ndants violated the Due Process Claus e. See, e.g., Jones, 94 F.3 d at 910 (concluding defendant failed to satisf y the second prong of the due process analysis “[e]ven if [he] had establ ished that he was actually and substantially prejudiced”); Automated Med. Lab’y s, 770 F.2d at 403 (“[W]h ile proof of actual prejudice makes a due process claim concrete an d ripe for adjudicatio n, it does not automatically make the claim valid.”). 6 Though we assume without deciding that Defendants have shown actual, substantial prejudice, we pause to note t hat the district cour t’s reaso ning on this score was deficient. To begin, when the “‘ claimed prejudice is th e unavailability of [a] witness[],’ the defendant must ... ‘sh ow that the informatio n the witness would ha ve provided was not available from other sources.’” Palmer, 15 9 F.4th at 227 (quoting Jones, 94 F.3d at 908). Logically, the same rule applies to lo st documents. Yet the district c ourt did not require this showing and made no findings abo ut whether other sources could provide the information conveyed by Peterson’s and Tack ett’s witness interviews and the unspecified state agency documents. In addition, the di strict court found substantial prej udice based on th e lost documents even though Defendants never expl ained what any of the d ocuments (save o ne) would have said. Proof of prejudice m ust be actual and defini te, not “ speculative.” Jo nes, 94 F.3d at 907. Yet Defendants merely cl aim ed, without any s pecificity, that some unidentified documents were “ potentially excu lpatory.” J.A. 420 (em phasis added). The one document D efendants did iden tify was a contract Deloitte employ ees signed before be ginning work on the Ma ssachusetts project, which stat ed that project deliverables belonged t o Massachusetts. Alt hough Defendants obtained an unsigned copy of the contract, t he distr ict court found prejudic e from the loss of signe d copies. Generally, loss of “the best means of proving [D efendants ’] contentions” is not actual prejudice when Defendants possess other means to present th eir defense. United Stat es v. Cederquist, 641 F.2d 1347, 1351 – 1352 (9th Cir. 1981) (holding that district court erred in findin g actual prejudice from missi ng documents when witnesses coul d testify concerning the information).

14 A. Before evaluating the merits of the district court’s due process an alysis, we pause to identify the s tandard of review. As we h ave often repeat ed when considering clai ms of preindictment delay, “[w] e review the district court’s factua l findings for clear error a nd its conclusions of law de novo.” Palmer, 15 9 F.4th at 226; see Uni ted States v. Villa, 70 F.4th 704, 715 (4t h Cir. 2023). Both p rongs o f the preindictment delay inquir y, however, may involve re view of factual findings and legal conclu sions, so dis cerning the nature of the precise question be fore us is important. We conclude that t he question we address here— whether a p rosecution “violates fundamental conceptions of justice or the com munity’s sense of fair play and decency” — presents a mixed question of law and fa ct subject to de novo revie w. Automated Med. Lab’ys, 770 F.2d at 404 (internal qu otation marks omitted). A mixed qu estion asks “whether the historical facts found sa tisfy the legal test chosen.” U.S. Bank Nat’l Ass’n e x rel. CWCap. Asset Mgmt. LLC v. Vill. at L akeridge, LLC, 583 U.S. 387, 39 4 (2018). Th e second prong of the preindictmen t delay inquiry does ju st that. It requires a court to determine whether, c onsidering the court’s findings about the j ustification for the government’s delay and the prejudice to the defendant, the prosecuti on comports with the relevant legal sta ndard, that is, “fun damental conceptions of justice or the comm unity’s sense of fair play and d ecency. ” Jones, 94 F. 3d at 910 (internal qu otation marks omitted). T he standard of review for a mixed questi on depends on “ whethe r answering it entails primarily legal or factual work.” U.S. Bank, 583 U.S. at 396. Generally, where the mixed question “requ ire[s] courts to expound on the law, pa rticularly by amplif ying or

15 elaborating on a broad legal standard,” d e novo review is appropria te. Id. (citing Salve Regina Coll. v. Russell, 499 U.S. 225, 231 – 23 3 (1991)). W here the question “immerse[s] courts in case - specific factual is sues,” more d eferential review applies. Id. (citing Pierce v. Underwood, 487 U. S. 552, 561 – 562 (1988)). “In the constitutional realm,” howeve r, “the calculus changes. ” Id. at 396 n.4. “There, [the Supreme Court has] often held that the role of appellate courts ‘in marking out the limits of [a] sta ndard through the process o f case -by- case adjudication’ favors de novo review even when answering a mixed question primarily involves plunging into a factual reco rd.” Id. (quoting Bose Corp. v. Consu mers Union of U.S., Inc., 46 6 U.S. 485, 503 (1984), and collecting cases). The mixed question at the second prong of the preindictment delay in quiry requires courts to expound on a broad constitutional standard; therefore, de novo review is most appropriate. It’ s hard to imagi ne a broader s tandard than one focu se d on “fundame ntal conceptions of justice o r the community’s sens e of fair play and decency.” Jones, 94 F. 3d at 910 (internal quota tion marks omit ted). And this broad standard is also a c onstitution al one, making de novo r eview even more fittin g. See Bose Cor p., 466 U.S. at 503 (“ When th e standard governing the decision of a particular case is pr ovided by the Constitution, [appellate courts’] role in marking out the limits of the standard th rough the process of case -by- case adjudicati on is of special imp ortance.”); Lovasco, 431 U.S. at 790 (drawing the due process standar d for preindictme nt delay from Rochin v. Cali fornia, 342 U.S. 165,

16 173 (1952), and Mooney v. Holohan, 294 U. S. 103, 112 (1935)). 7 Accordingly, we will review de novo the district court’s de termination that, balancing the justification for th e delay and the prejudice caused by it, t rial on the dismissed co unts would violate Defendants’ d ue process rights. A ccord She aley, 641 F.3d at 634 (“This court revie ws legal issues, including claims of due pro cess violations, de novo.” (internal quotation mark s omitted)); United States v. Harvel, 11 5 F.4th 714, 7 27 (6th Cir. 2 024) (reviewing the district court’s “finding s about historical facts under the deferential clear - error stan dard but giv[ing] fresh (de novo) review to its ultimate holding that the gove rnment did not violate due process”). B. Having se ttled the stan dard of review, we no w turn to the merits. A t th e se cond prong of the pre indictment delay analysis, w e consider whether, ba lancing Defendants ’ prejudice against the G overnment’s justif ication for the delay, the Government’s action in prosecuting after suc h delay “ violate[s] fundamental concept ions of justice or th e community’s sense of fair play and decency. ” Jones, 94 F.3d at 910 (internal quotati on marks omitted). The district court found, and Defendants do not dispute, that the Government’s delay was the product of a “ prolonged inv estigation ” untainted by any “improper motive.” Minkkinen, 678 F. Su pp. 3d at 794. 7 The Supreme Court h as also “asked whet her there is ‘a long histo ry of appellate practice’ indicating the appropriate standard” of review. Monas ky v. Taglieri, 140 S. Ct. 719, 730 (2020) (quot ing Pierce, 487 U.S. at 558). Here, caselaw does not revea l a “ uniform, reasoned pra ctice” concerning the st andard of review for th e second prong of the preindictment delay in quiry. Id.

17 The Supreme Court has square ly “h[e]ld that to prosecute a defendant following investigative delay d oes not depriv e him of due process, e ven if his defense might h ave been somewhat prej udiced by the lapse of tim e.” Lovasco, 43 1 U.S. 796; see id. (fin ding no due process violati on from preindictment delay despite the death of two witnesses, where “the only reaso n the Government p ostponed ac tion was to await the results of additional investigation”). We have adhered to the Supreme Court’s holding in our ow n decisions, reiterating t hat “investigative delay, as opposed to i ntentional delay under taken to gain tactical advantage, would not v iolate due process.” Howell, 904 F.2d at 894 –895; s ee also, e.g., Uribe-Rios, 558 F.3d at 358 (“If delay results from a protracted investigation that was nevertheless conducted in good f aith, the Supreme Court has held that ‘to prosecute a defenda nt following investigative delay does not depriv e him of due process. . . .’” (quoting Lovas co, 431 U.S. at 796)). By contrast, the Due Process Clause “re quires the dismissal of an indi ctment ... if the defendant can pr ove that the Govern ment’s [prejudicial] d elay in bringin g the indictment was a deli berate device to gain an advantage over him.” United Sta tes v. Gouveia, 467 U.S. 180, 192 (1984). “[A] due process vio lation [might also] be made ou t ‘upon a showing of prosecutorial dela y incurred in reckless disregard of circumstances, known to the prosecuti on, suggesting th at there existed an appreci able risk that delay w ould impair the ability to m ount an effective de fense.’” Howell, 90 4 F.2d at 895 n.9 (qu oting Lovasco, 431 U.S. at 7 95 n.17). And this Court has found a due process violation when the prosecution admitte d it “was ‘negligent’ in not prosecuti ng the defendant earlier,” was

18 not “engaged in preindi ctment investigation,” and caused actual preju dice to the defendant for “mere convenience. ” Id. at 895. 8 As the Supreme C ourt has explained, “i nvestigative delay is fundam entally unlike delay undertaken” for t actical advantage or ot her reasons, “because i nvestigative delay is not so one -sided.” Lovasco, 431 U.S. at 795. P rosecutors are not “ constitutionall y obligated to file charge s against a suspec t as soon as they have p robable cause ” or “ as soon as they marshal en ough evidence to prove guilt beyond a reasonable d oubt but before the ir investigations are com plete.” Gouveia, 467 U.S. at 192 n.7. To h old otherwise would “prevent[] society from bringing lawbrea kers to justice,” would “pres sure prosecutors into resolving doubtful case s in favor of early — an d possibl y unwarranted— prosecuti ons,” and would “preclude the Government fro m giving full consideration to the desirability of not prosecuting in partic ular cases.” Lovasco, 43 1 U.S. at 792 –794. A prosecutor who “refuses to seek indictments un til he is complete ly satisfied that he should p rosecute and will be able promptly to establ ish guilt beyond a reas onable doubt” abides b y the “standards of ‘fair play and decency’” rather than deviating f rom them. Id. at 795. Accordingl y, the Due Process Clause is not offended when pro secutors “defer action f or [investigatory] reasons.” Id. The district court er red in concluding t hat the Govern ment’s investigative preindictment de lay violated t he Due Proc ess Clause. In the c ourt’s view, (1) the 8 Unlike most circuits, this Court does not require defenda nts alleging preindictment delay to show “ th at ‘the government intentionally dela yed the indict ment to gain an unfair tactical advantage or for other bad fait h motives.’” Jones, 94 F.3d at 905 (quoting United States v. Crooks, 766 F.2d 7, 11 (1st Cir. 1985) (Breyer, J.), and colle cting cases).

19 Government’s expl anation of its investigatory activities did not justif y “ the length of the investigation ” and (2) t he Government’s justifi cation was “ insufficient in light of the e xtent of the prejudice caused by the delay.” Minkkinen, 678 F. Sup p. 3d at 794. Nei ther reason supports dismissal of t he charges for unconsti tutional delay. First, regarding the length of the investigation, the district court’s own reconstruction of the investigati ve timeline shows that the Government’s good faith investigation accounts for the entire period of preindictm ent delay. The Government became aware of the allegations against Sagit ec in t he summer of 2 016, and it opened a formal investigation only months later in Fe bruary 2017. Id. at 793. From then until December 2019, the Government conducted a pproximately 31 int erviews and rev iewed a large number of doc uments. Id. at 793 – 794. In Decem ber 2019, t he Government inf ormed Sagitec that it wa s the target of a criminal investigation, an d in February 2020 t he Government made a pr offer to the compan y. See id. at 787, 794. Sagitec then undertook an internal investi gation that lasted appr oximately one year, and in the meanti me the Governm ent con tinued its work a nd interviewed Sagi tec employees. I d. at 787, 794 & n.10. Sagitec’s i nvestigation “prompted additional questions and leads for investigation,” whic h the Government pur sued. J.A. 663. Investigator s interviewed Peterson in November 2020 and made a proffer to him one mo nth later. Th e Government infor med Minkkinen that h e was a target that sam e month (December 202 0) an d informed Sambasi vam that he was a targe t in March 2021. Throughout 2021, the Government “engag[ed ] in pre - indictment disc ussions” with Defendants’ counsel. Minkkinen, 678 F. Su pp. 3d at 794. These included pro ffers of

20 evidence in March and April, a video conference with a defense e xpert in June, phone conversations with def ense counsel, and a meeting with defense co unsel in December 2021. Investigators als o “continued reviewing emails turned over by Sagitec” duri ng this time and conducted more interviews. Id. The last witness interview occurred on July 2 1, 2022, about one month before the original indi ctment. Id. As the distric t court’s timeline shows, no port ion of the preindict ment delay is left unaccounted for by th e Government’s good faith investigative efforts. Of course, the Government has not e xplained what occurre d on ever y single day i n the relevant period. But such a detai led exp lanation is not r equired. Indeed, the Supreme Court has rejected a standard that would require prosecu tors “ to trace the day -by- day progress of each investigation” for distri ct courts, as though the “courts were re quired to deci de. . . when the prosecution should have commenced.” Lovasco, 431 U. S. at 793 n.14. T o the contrary, “t he Due Process Cla use does not per mit courts to abort cri minal prosecutions si mply because they disagree with a prosecutor’s j udgment as to when to seek an i ndictment.” Id. at 790. The preindictment d elay here, though leng thy, is entirely attrib utable to the Government’s good fai th investigation. Prose cuting Defendants afte r this “investigati ve delay does not deprive [them] of due pr ocess.” Id. at 796; see id. (f inding no due process violation where “the only reason the Govern ment postponed action w as to await the results of additional investigat ion”); United States v. Lopez, 860 F.3d 201, 213 (4th Cir. 2017) (finding no due proc ess violation where six - year delay was “a result of co ntinued reasonable investigatio n”); United States v. McKoy, 129 Fed. App. 815, 819 (4th Cir. 2005)

21 (finding no due process violation when addit ional investigation was a “legitimate need for the delay”). T hough “[t]he wheels of governm ent bureaucracy may, at times, seem to turn at a frighteningly slow pace,” “careful investigation and con sideration prior to the brin ging of criminal charges” accord with fair play and decency. Automated Med. Lab’ys, 770 F.2d at 404. Second, the district court reasoned that the Government’ s investigat ion could no t justif y the delay becau se the “ ‘ actual prejudice ... is substantial. ’” Minkkinen, 67 8 F. Supp. 3d at 794 (quotin g Automated Med. Lab’ys, 770 F.2d at 404). T hat conclusion stand s in tension with our “tw o- pronged inquiry ” for evaluating preindictme nt delay. Uribe-Rios, 558 F.3d at 358. The first prong requires a defendant to prove “actual substantial prejudice.” Jones, 94 F.3d at 907; see Sheale y, 641 F.3d at 633–634; Mario n, 404 U.S. a t 324. But actual substantial prejudice is n ot sufficient; the court must consider the reasons for the delay as well to determine wheth er prosecution after such delay viola tes “fundamental concepti ons of justice or the co mmunity’s sense of fair play and decency.” Jones, 94 F.3d at 9 10 (internal quotatio n marks omitte d); see Lovas co, 431 U.S. at 790. And “a protracted investigation ... conducted in good faith” can justify even a delay that results in actual substantial prejudice. Uribe-Rios, 558 F.3d at 358; s ee Lovasco, 431 U.S. at 796; Howell, 904 F.2 d at 894–895. In reaching its conclusion, the di strict court r el ied on this Court’ s decision in Automated Medical Laborato ries. There, the Government att ributed the delayed indictment to “the leng thy administrative revi ew process at the FDA,” “the time req uired for Government attorne ys to become familiar with a complex case,” “manpower pro blems”

22 in the prosec u tor’s offi ce, and, in part, “addit ional investigative acti vities.” 770 F.2d at 404. This Court stated: “While such reasons may not suffice where the actual prejudice to the defendant from the delay is substantial, we find that there is no due process violation when these reasons ar e considered in light of the slight, possibly nonexistent p rejudice suffered by [the defen dant].” Id. From this statement, the district court reasoned that, because the prejudice here was subst antial, the Govern ment’s investigation could not suffice to justify the de lay. Automated Medical Laboratories does not sta nd for that proposition. For one, the Court did not hold that the government’s reason s for the delay in that ca se would not suffice if the prejudice wer e substantial but simply allowed t hat those reasons “ may not suffice” in such circumst ances, which were not presented there. Id. (emphasis added); see als o Payne v. Taslimi, 99 8 F.3d 648, 65 4 – 655 (4th Cir. 2021) (defining dicta). Moreover, the Government’s reasons for the delay in Aut omated Medical Laborato ries —which included “manpower proble ms” and a noncrimin al regulatory process, 770 F.2d at 404 — do not match the reason for the delay here, a sustained cri minal investigation i nvolving “voluminous discovery ” and numerous “witnesses in multi ple states,” Minkkinen, 678 F. Supp. 3d at 78 8. For these reasons, the district court ’s reliance on Automated Med ical Laboratories was misp laced. At bottom, the district court’s analy sis would compel the Governme nt to seek an indictment before its in vestigation was comple te. Due process does n ot require that result. See Lovasco, 431 U.S. at 790; Gouveia, 467 U.S. at 192 n.7. There has been “no sh owing that the Government intentionally delaye d to gain some tactical adva ntage over

23 [Defendants] or to ha rass them,” Ma rion, 404 U.S. at 3 25, or that t he Government acted recklessly or for “mere convenience” in not concluding its investigation soone r than it did, Howell, 904 F.2d at 89 5 & n.9. Rather, the district court’s own factual findings establish a timeline showing that the Government ’s go od faith investigative efforts continue d until at least one month bef ore the indictment was returned. Even assum ing Defendants have proven actual substantial prejudice to their de fense, requiring them to stand trial after t he Government delay ed indictment to investiga te further does not vi olate “fundamental conceptions of justice o r the community’s sens e of fair play and decency.” Jones, 94 F.3d at 904 (internal quotati on marks omitted). III. For the foregoing reasons, we reverse the district court’s order dismissing ten of the fourteen counts charge d in the supers eding indictment, and we remand the ca se for further proceedings on all four teen counts. REVERSED AND REM ANDED