Palantir Usg, Inc. v. United States

United States Court of Appeals for the Federal Circuit ______________________ PALANTIR USG, INC., Plaintiff-Appellee v. UNITED STATES, Defendant-Appellant ______________________ 2017-1465 ______________________ Appeal from the United States Court of Federal Claims in No. 1:16-cv-00784-MBH, Judge Marian Blank Horn. ______________________ Decided: September 7, 2018 SEALED OPINION ISSUED: September 7, 2018 PUBLIC OPINION ISSUED: September 13, 2018 * ______________________ THEODORE OLSON, Gibson, Dunn & Crutcher LLP, Washington, DC, argued for plaintiff-appellee. Also represented by KAREN LOUISE MANOS, AMIR C. TAYRANI; JOSH KREVITT, New York, NY; HAMISH HUME, STACEY K. GRIGSBY, JON KNIGHT, JOSHUA RILEY, Boies, Schiller & * This opinion was originally filed under seal and has been unsealed in full. 2 PALANTIR USG, INC. v. UNITED STATES Flexner, LLP, Washington, DC; DAVID BOIES, Armonk, NY. DOMENIQUE GRACE KIRCHNER, Commercial Litigation Branch, Civil Division, United States Department of Justice, Washington, DC, argued for defendant-appellant. Also represented by CHAD A. READLER, ROBERT EDWARD KIRSCHMAN, JR., DOUGLAS K. MICKLE. GIDEON A. SCHOR, Wilson, Sonsini, Goodrich & Rosati, PC, New York, NY, for amicus curiae Technology Net- work. Also represented by ADAM WILLIAM BURROWBRIDGE, Washington, DC. ______________________ Before NEWMAN, MAYER, and STOLL, Circuit Judges. STOLL, Circuit Judge. The government appeals from a permanent injunction on its solicitation of bids for Distributed Common Ground System – Army Increment 2 (“DCGS-A2”), the Army’s primary system for processing and disseminating multi- sensor intelligence and weather information. The United States Court of Federal Claims granted the injunction after concluding that the Army failed to comply with the requirements of 10 U.S.C. § 2377. We affirm. PALANTIR USG, INC. v. UNITED STATES 3 BACKGROUND 1 Palantir USG, Inc. (“Palantir”) filed a pre-award bid protest in the Court of Federal Claims, challenging the Army’s solicitation 2 for DCGS-A2. The solicitation seeks a single contractor to be the system data architect, devel- oper, and integrator of DCGS-A2. Palantir’s complaint alleges that the Army violated § 2377(c) by, among other things, failing to determine whether its needs could be met by commercial items before issuing the contested solicitation. See § 2377(c)(2). To provide background, we introduce the applicable statute and regulations, the DCGS-A2 system, the relevant facts regarding pre- and post-solicitation activity, and the procedural history of this case. I. The Statutory and Regulatory Preference for the Acquisition of Commercial Items This appeal is centered on the Federal Acquisition Streamlining Act (“FASA”), which requires that federal agencies, to the maximum extent practicable, procure commercially available technology to meet their needs. Federal Acquisition Streamlining Act of 1994, Pub. L. No. 103-355, § 8104, 108 Stat. 3243 (1994) (codified as amended at 10 U.S.C. § 2377). As the government 1 The facts and procedural history of this case are extensive. A more exhaustive recitation of the facts underlying this appeal may be found in the Court of Federal Claims’ opinion. See Palantir USG, Inc. v. United States, 129 Fed. Cl. 218, 221–43 (2016) (“CFC Op.”). We provide a summary of the relevant facts and procedural history necessary to resolve this appeal. 2 United States Department of the Army, Army Contracting Command, Aberdeen Proving Group’s Re- quest for Proposals (“RFP”) No. W56KGY-16-R-0001. 4 PALANTIR USG, INC. v. UNITED STATES acknowledges, the legislative history reflects Congress’s understanding that “[t]he purchase of proven products such as commercial and nondevelopmental items can eliminate the need for research and development, mini- mize acquisition leadtime, and reduce the need for de- tailed design specifications or expensive product testing.” S. Rep. No. 103-258, at 5 (1994), reprinted in 1994 U.S.C.C.A.N. 2561, 2566. FASA provides that the “Federal Acquisition Regula- tion [FAR] shall provide regulations to implement” FASA. 41 U.S.C. § 3307(e)(1). Under FASA: (a) Preference.—The head of an agency shall en- sure that, to the maximum extent practicable— (1) requirements of the agency with respect to a procurement of supplies or services are stat- ed in terms of— (A) functions to be performed; (B) performance required; or (C) essential physical characteristics; (2) such requirements are defined so that commercial items or, to the extent that com- mercial items suitable to meet the agency’s needs are not available, nondevelopmental items other than commercial items, may be procured to fulfill such requirements; and (3) offerors of commercial items and nondevel- opmental items other than commercial items are provided an opportunity to compete in any procurement to fill such requirements. 10 U.S.C. § 2377(a); see FAR 11.002(a)(2). FASA further requires that agencies shall, “to the maximum extent practicable . . . acquire commercial items or nondevelop- PALANTIR USG, INC. v. UNITED STATES 5 mental items [NDIs] other than commercial items to meet the needs of the agency.” § 2377(b)(1); see FAR 12.101(b). FASA achieves its preference for commercial items in part through preliminary market research. Before solicit- ing bids or proposals, agency officials must conduct mar- ket research 3 concerning the availability of commercial items pursuant to § 2377(c)(1), which states: (1) The head of an agency shall conduct market research appropriate to the circumstances— (A) before developing new specifications for a procurement by that agency; (B) before soliciting bids or proposals for a contract in excess of the simplified acquisition threshold; and (C) before awarding a task order or delivery order in excess of the simplified acquisition threshold. § 2377(c)(1); FAR 10.001(a)(2). Next, agency officials must use that market research to determine whether commercial items can meet the agency’s requirements, with or without modification of either the commercial items or the agency’s requirements, pursuant to § 2377(c)(2): 3 The implementing regulations define market re- search as “collecting and analyzing information about capabilities within the market to satisfy agency needs.” FAR 2.101(b). Conducting market research “involves obtaining information specific to the item being acquired” and the regulation explains that the “extent of market research will vary, depending on such factors as urgency, estimated dollar value, complexity, and past experience.” FAR 10.002(b)(1). 6 PALANTIR USG, INC. v. UNITED STATES (2) The head of an agency shall use the results of market research to determine whether there are commercial items or, to the extent that commer- cial items suitable to meet the agency’s needs are not available, nondevelopmental items other than commercial items available that— (A) meet the agency’s requirements; (B) could be modified to meet the agency’s re- quirements; or (C) could meet the agency’s requirements if those requirements were modified to a rea- sonable extent. § 2377(c)(2) (emphasis added); FAR 10.001(a)(3). Palantir argues—as it did below—that the Army violated both of these FASA mandates, § 2377(c)(1) and (2). II. The Distributed Common Ground System – Army Increment 2 (DCGS-A2) We briefly introduce the purpose and evolution of the Army system at issue here. The Distributed Common Ground System (“DCGS”) is made up of Army, Air Force, Navy, and Marine Corps ground processing systems that can share information across the Joint Force. The over- arching purpose of the Army’s Distributed Common Ground System (“DCGS-A”) is to combine all of the Ar- my’s intelligence software/hardware capabilities into one program with the ability to access and be accessed by Army intelligence and command components, as well as other military and intelligence systems. DCGS-A in- cludes many software products—commercial, government, and open source—as well as software integration that allows all the different products and components to com- municate and operate seamlessly. CFC Op., 129 Fed. Cl. at 223. PALANTIR USG, INC. v. UNITED STATES 7 The original DCGS-A Increment 1 (“DCGS-A1”) is op- erational and deployed worldwide, but its “data architec- ture is over 10 years old and is based upon technology that is nearing obsolescence, with no growth margin.” Id. at 233. Therefore, in 2014, the Army began investigating the best way to approach DCGS-A2, which would “intro- duce a new and modernized data management architec- ture (DMA) using a modular system approach to perform Army intelligence analysis capabilities.” Id. at 223. The performance work statement (“PWS”) for this so- licitation stated that the requirements of DCGS-A2 in- cluded the ‘‘development of new data architecture, standards based enhanced visualization and analytical tools, cloud computing and ‘big data’ analytic capabilities; cyber analytics and data integration, visualization capa- bilities, Cyber Operations, Interoperability, Counter Intelligence/HUMINT, Weather, GEOINT, Geospatial Engineering and Sensor Management,’’ and explained that ‘‘[t]hese efforts include Software Development, Capa- bility Enhancements, Integration, Limited Fielding and Training support, Maintenance, and Support for logistics development, for a period of performance of six years from contract award.’’ Id. The draft version of the perfor- mance work statement for the DCGS-A2 solicitation stated that ‘‘[t]he DCGS-A Increment approach utilizes spiral deliveries to maintain interoperability with Army and Joint ISR [Intelligence, Surveillance and Reconnais- sance] architectures and to address capability insertion and enhancements. This system must remain interoper- able and compatible with the Joint command system infrastructure and mission applications.’’ Id. As indicat- ed by the contracting officer who issued the solicitation, the data management architecture “will serve as the architecture foundation and the heart with which the rest of the capabilities will depend on to function. The [data management architecture] development is therefore the 8 PALANTIR USG, INC. v. UNITED STATES focus of the first task order executed under the DCGS-A Increment 2 contract.’’ Id. III. The Palantir Gotham Platform As noted above, Palantir argues that the Army violat- ed § 2377(c) by failing to determine whether the Army’s needs could be met by commercial items, and that, had the Army done so, it would have issued one or more solicitations to procure commercial or nondevelopmental items to satisfy the DCGS-A2 requirements. J.A. 194–95. Specifically, Palantir submits that its flagship software product, the Gotham Platform, could satisfy the Army’s requirements. Palantir’s Gotham software product is a data- management platform that Palantir began to market to private sector and government customers in 2009. The software enables agencies to integrate, visualize, and analyze large amounts of data from different sources that reside in different databases in different formats. The parties stipulate that the government has previously procured the Palantir Gotham Platform on a commercial item basis, and that the GSA Schedule lists both term licenses and perpetual licenses for Palantir’s platform. Parties’ Joint Stipulations of Fact ¶¶ 3–4, Palantir USG, Inc. v. United States, No. 1:16-cv-00784-MBH (Fed. Cl. Apr. 24, 2017), ECF No. 125 (“Joint Stip.”). The parties also stipulate that Palantir Gotham is a commercially available data management platform. IV. Pre-Solicitation Activity The parties do not dispute the following Court of Fed- eral Claims’ fact findings. In 2014, the Army decided to curtail the third release of DCGS-A1 and redirect its efforts to acquiring and launching DCGS-A2. The Army PALANTIR USG, INC. v. UNITED STATES 9 chartered an independent Data Integration, Visualization and Analytics (“DIVA”) Market Study. 4 This market study, dated July 2014, was completed by the MITRE Corporation, a not-for-profit research and development organization. According to the parties, the DIVA Market Study was intended to “provide situational awareness and market trends to the Army leadership of the ‘state-of-the- practice’ within the commercial DIVA software platform landscape.” Joint Stip. ¶ 11. The DIVA Market Study report summarized MITRE’s recommendations for the DCGS-A2 acquisition effort. According to the report, the DIVA Market Study assessed three acquisition approach- es: a. Cloud Infrastructure Platform Provider: Pro- vide highly-scalable and reliable computing infra- structure services (e.g., data bases [sic]; analytic engines; computing and storage; identity man- agement); b. Turn-Key: Procure a commercial product as ba- sis of [DCGS-A2] infrastructure. Integrate addi- tional applications onto this infrastructure[;] c. Hybrid approach: both an Enterprise Cloud Platform and a Turn-Key Platform, including in- tegration of additional applications . . . . Id. ¶ 11 (citations omitted). Having considered each of these three acquisition approaches, the DIVA Market Study recommended the hybrid approach. The study’s “Key Observation” was that the hybrid approach blends the benefits of the other two approaches, provides “the global scale of the cloud infrastructure with the ‘out-of- 4 A DIVA platform is sometimes referred to as a “data management platform.” Compare J.A. 18400, with J.A. 12226–27. 10 PALANTIR USG, INC. v. UNITED STATES the-box’ capabilities of the DIVA ‘Turn Key’ platform,” and provides “better tactical edge support.” J.A. 12234. The DIVA study outlined how to apply a hybrid approach to the DCGS-A2 capabilities. In particular, the hybrid approach would start with procurement of two Commer- cial-off-the-Shelf (“COTS”) foundation components: (a) cloud infrastructure services and (b) a DIVA “Turn Key” infrastructure platform. Though the DIVA Market Study did not analyze any potential vendors, it assessed overall market trends and served as an early indication that commercial items should be considered for the DCGS-A2 infrastructure platform. Following completion of the DIVA Market Study, the Army issued three requests for information (“RFIs”). It issued its RFI #1 in August 2014, just one month after release of the DIVA Market Study. The goal of RFI #1 was to assess “the level of relevant competition and capabilities in the market place and elicit industry feed- back to assist the Program Office in developing the Acqui- sition Plan” for the potential DCGS-A2 procurement. J.A. 11802. It requested “respondents’ corporate overview information and basic qualifications in managing software development projects that are similar in scope and process to the DCGS-A program.” J.A. 11876 (emphasis added); see also J.A. 11876–81. The Army issued RFI #2 in December 2014. Palantir responded, expressing concern that RFI #2 was focused on collecting information on the respondents’ ability to conduct a large-scale development effort, instead of as- sessing existing software capabilities that would be applicable to DCGS-A1 capability gaps. The Army issued RFI #3 in May 2015, which was meant to “[i]nform the small business role for Increment 2 [and] [d]etermine if [a Small Business Set-Aside] is ap- propriate.” J.A. 11803. In response, Palantir again highlighted its concerns with the Army’s acquisition PALANTIR USG, INC. v. UNITED STATES 11 approach, asserting that “[t]he successful delivery of Increment 2 depends on the answer to a central question: will the Army acquire a data platform from the commer- cial market or will it attempt to build one itself?” J.A. 11918. In July 2015, the Army Materiel Systems Analysis Activity issued a Trade Space Analysis, which identified and evaluated technical functionality, cost, usability, schedule risk, and technical risk for DCGS-A2. The report indicated that the Trade Space Analysis would inform the economic analysis and RFP for DCGS-A2 and analyzed the following options: COTS, Government-off- the-Shelf (“GOTS”), and hybrid. The report concluded that a hybrid COTS-development approach was the best of the three alternatives, noting that such an approach was currently functioning in the Department of Defense Intelligence Community and would only require minor development to fill capability gaps. On July 13, 2015, however, the Army issued a Market Research Report that concluded the opposite—that “the [DCGS-A2] development effort cannot be procured as a commercial product.” J.A. 11840. This Market Research Report indicated that three features were not available as commercial products: Data Fusion, Intelligence Support to Cyber, and DCGS Integrated Backbone Upgrade. Id. It further addressed Palantir directly, finding Palantir’s response to earlier RFPs non-responsive because Palantir “did not provide any examples of past experience relevant to the development of Increment 2.” J.A. 11835–36. The Market Research Report further stated that, “[b]ased on the Market Research to date, the recommended approach” for DCGS-A2 “is a five (5) year Engineering and Manufac- turing Development (EMD) effort consisting of two releas- es.” J.A. 11841 (emphasis added). It also stated, without any explanation, analysis, or support, that “[s]ignificant portions of the anticipated Increment 2 scope of work” are “not available as a commercial product.” J.A. 11840. 12 PALANTIR USG, INC. v. UNITED STATES Two days later, on July 15, 2015, the Army issued its draft performance work statement, defining the efforts required to acquire services for the development and integration of DCGS-A2. J.A. 10410–594. The perfor- mance work statement defined the requirements for DCGS-A2 to include “development of new data architec- ture” and completion of the “design, development, integra- tion and test.” J.A. 10418–19. Palantir filed a response to the draft performance work statement in October 2015, asserting that the “Army does not need to build that [data management] platform, as it can buy it today.” J.A. 10693. In the same month, on October 21, 2015, Ms. Heidi Shyu, as the Senior Procurement Executive, signed a Determination & Findings for “Award of a Single Source Indefinite-Delivery Indefinite-Quantity (IDIQ) Single Award Contract Exceeding $103M” for DCGS-A2. J.A. 12298–304. The Determination & Findings noted that DCGS-A2 “is heavily focused on design and develop- ment of a new data management architecture by a con- tractor as the systems integrator,” and “[d]evelopment of the data integration layer is pivotal and complicated by multiple interfaces and interoperability requirements with external intelligence systems.” J.A. 12299 ¶ 4. The Determination & Findings concluded that: [I]ssuing a single award IDIQ contract will miti- gate many of the risks identified herein and is in the best interest of the Government. Due to the complex developmental efforts this work entails, further competition at the task order level would interrupt development, ultimately increase price, and cause schedule slippages. .... [A] single-source task or delivery order contract estimated to exceed $103 million for [DCGS-A2] Engineering Manufacturing and Development contract is authorized because the task or delivery orders expected under the contract are so integral- PALANTIR USG, INC. v. UNITED STATES 13 ly related that only a single source can reasonably perform the work. J.A. 12302, 12304. V. The Solicitation On December 23, 2015, the Army issued the solicita- tion that is the subject of this appeal. The solicitation contemplated the award of a single indefinite-delivery, indefinite-quantity contract for DCGS-A2, with the simul- taneous issuance of a cost-reimbursement type task order. It sought a single contractor to be the system data archi- tect, developer, and integrator of DCGS-A2. The solicita- tion also required a software capability demonstration, which the Army contemplated “could include a Govern- ment Furnished Information (GFI), Commercial Off-[t]he- Shelf (COTS), Government Off-the-Shelf (GOTS), or Open Source product(s).” J.A. 10960. The performance work statement accompanying the solicitation explained that the successful offeror would be responsible for, among other things, the development of new data architecture; cloud computing and big data analytic capabilities; data integration; and interoperability with counter intelli- gence/human intelligence. The performance work state- ment also stated that the software design release/development should include “maximization of reuse of GOTS/COTS products.” J.A. 11101 ¶ 3.4.1. VI. Post-Solicitation Activity Shortly after the Army issued the solicitation, Palan- tir filed a pre-award bid protest, which the Government Accountability Office (“GAO”) denied in May 2016. See generally Palantir USG, Inc., No. B-412746, 2016 WL 3035029 (Comp. Gen. May 18, 2016) (“GAO Op.”). Then, on June 30, 2016, Palantir filed the current pre-award bid protest in the Court of Federal Claims. Count one of Palantir’s complaint alleged that the Army violated § 2377 and 48 C.F.R. §§ 10.002 and 11.002 by refusing to 14 PALANTIR USG, INC. v. UNITED STATES solicit the data management platform as a commercial item. Count two alleged that the Army violated § 2377 and 48 C.F.R. §§ 10.002 and 11.002 by refusing to solicit a commercial item for the entirety of DCGS-A2. Count three alleged that the Army violated § 2377(c) by failing to determine whether its needs could be met by commer- cial items. On July 1, 2016, one day after Palantir filed its complaint in the Court of Federal Claims, the govern- ment issued a Determination of Non-Commercial Item, laying out the government’s justification for its determi- nation regarding commercial items following market research. In the Court of Federal Claims, the parties filed cross- motions for judgment on the administrative record. The Court of Federal Claims granted judgment in Palantir’s favor, concluding that the Army failed to determine whether commercial items meet or could be modified to meet the agency’s needs and that, by failing to do so, the Army acted in an arbitrary and capricious manner in violation of 10 U.S.C. § 2377. CFC Op., 129 Fed. Cl. at 282, 290. The court explained: Not only did the agency fail to explain or indicate what commercial items possibly were available or had been considered, the Market Research Report is devoid of any information regarding the possi- ble commercial items that could be modified to meet the Army’s requirements. . . . [T]here is no evidence that the agency made [a determination regarding the suitability of Palantir’s data man- agement platform] after the market research was complete or prior to issuing the solicitation. The total absence of any discussion regarding commer- cial items, or possible modifications to commercial items, reinforces the court’s understanding that the Army was focused on a developmental ap- proach to the DCGS-A Increment 2 at an early PALANTIR USG, INC. v. UNITED STATES 15 point in the procurement process, to the exclusion of commercially available alternatives. Id. at 276. Further finding that the Army’s actions caused Palantir to suffer a ‘‘non-trivial competitive injury which can be addressed by judicial relief,” the Court of Federal Claims permanently enjoined the Army from issuing a contract award under the protested solicitation. Id. at 289–95 (quoting Weeks Marine, Inc. v. United States, 575 F.3d 1352, 1362 (Fed. Cir. 2009)). According to the injunction, the Army would have to properly and sincerely comply with FASA § 2377 before awarding a contract to meet its DCGS-A2 requirements. Id. at 295. The United States appeals. We have jurisdiction under 28 U.S.C. § 1295(a)(3). DISCUSSION I. Standard of Review We review the Court of Federal Claims’ ruling on the parties’ cross-motions for judgment on the administrative record de novo, applying the same standard of review as the trial court. Glenn Def. Marine (Asia), PTE Ltd. v. United States, 720 F.3d 901, 907 (Fed. Cir. 2013). Cross- motions for judgment on the administrative record are governed by Rule 52.1(c) of the Rules of the United States Court of Federal Claims (“RCFC”). See RCFC 52.1(c). In deciding these motions, the court considers “whether, given all the disputed and undisputed facts, a party has met its burden of proof based on the evidence in the record.” A & D Fire Prot., Inc. v. United States, 72 Fed. Cl. 126, 131 (2006) (citing Bannum, Inc. v. United States, 404 F.3d 1346, 1356 (Fed. Cir. 2005)). The Army’s procurement decision must be set aside only if it is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law,” or “without observance of procedure required by law.” 5 U.S.C. 16 PALANTIR USG, INC. v. UNITED STATES § 706(2)(A), (D). As the United States Supreme Court has explained about § 706(2)(A) review: [T]he court must consider whether the decision was based on a consideration of the relevant fac- tors and whether there has been a clear error of judgment. Although this inquiry into the facts is to be searching and careful, the ultimate standard of review is a narrow one. The court is not em- powered to substitute its judgment for that of the agency. Citizens to Pres. Overton Park, Inc. v. Volpe, 401 U.S. 402, 416 (1971), abrogated on other grounds by Califano v. Sanders, 430 U.S. 99, 105 (1977) (internal citations omit- ted). “Effective contracting demands broad discretion. Accordingly, agencies ‘are entrusted with a good deal of discretion in determining which bid is the most advanta- geous to the Government.’” Lockheed Missiles & Space Co. v. Bentsen, 4 F.3d 955, 958–59 (Fed. Cir. 1993) (quot- ing Tidewater Mgmt. Servs., Inc. v. United States, 573 F.2d 65, 73 (Ct. Cl. 1978) (internal citations omitted)). The reviewing “court’s task is to determine whether ‘(1) the procurement official’s decision lacked a rational basis; or (2) the procurement procedure involved a viola- tion of regulation or procedure.’” Savantage Fin. Servs., Inc. v. United States, 595 F.3d 1282, 1285–86 (Fed. Cir. 2010) (quoting Weeks Marine, 575 F.3d at 1358). The government raises two issues on appeal: (1) whether the trial court went beyond the statutory and regulatory language of FASA and its implementing regu- lations and imposed heightened obligations; and (2) whether the trial court wrongly discarded the pre- sumption of regularity and substituted its judgment in determining that the Army acted arbitrarily and capri- ciously and in violation of 10 U.S.C. § 2377. We address each issue in turn. PALANTIR USG, INC. v. UNITED STATES 17 II. The Trial Court Properly Concluded that the Army’s Actions Violated FASA § 2377(c)(2) The government first argues that the trial court erro- neously added requirements to § 2377, including that the Army was required to “fully investigate,” “fully explore,” “examine,” and “evaluate” whether all or part of its re- quirements could be satisfied by commercially available items, such as Palantir’s product. Appellant Br. 35–36; see also CFC Op., 129 Fed. Cl. at 282. We are not per- suaded that the Court of Federal Claims imposed addi- tional requirements beyond those required by the statute. FASA requires an agency to use the results of market research to “determine” whether there are commercial items that “meet the agency’s requirements; could be modified to meet the agency’s requirements; or could meet the agency’s requirements if those requirements were modified to a reasonable extent.” § 2377(c)(2). While the trial court’s thorough opinion sometimes uses words other than “determine,” we conclude that, read in context, those words were intended to be synonymous with “determine.” In any event, we need not devote significant discussion to this argument, as we “sit to review judgments, not opin- ions,” Stratoflex, Inc. v. Aeroquip Corp., 713 F.2d 1530, 1540 (Fed. Cir. 1983), and our de novo review leads us to the same conclusion as the one reached by the Court of Federal Claims. As discussed above, we give deference to the Army’s procurement decisions. Advanced Data Concepts, Inc. v. United States, 216 F.3d 1054, 1058 (Fed. Cir. 2000). Even with that deference, however, we conclude that the Ar- my’s procurement actions in this case were arbitrary and capricious and in violation of § 2377. First, as explained in detail below, the administrative record demonstrates that the Army, while conducting its market research, was on notice of the desirability of hybrid options that used commercial solutions and that Palantir claimed to have a commercial item that could meet or be modified to meet 18 PALANTIR USG, INC. v. UNITED STATES the Army’s needs. Furthermore, the record shows that the Army did not use the results of that market research to determine whether there were commercial items that could meet its requirements, could be modified to meet its requirements, or could meet its requirements if those requirements were modified to a reasonable extent. See § 2377(c)(2). The administrative record reflects that the Army was on notice of the possibility that commercial items 5 could satisfy its needs for portions of DCGS-A2. Indeed, the Army was on notice as early as July 2014, when it re- ceived the DIVA Market Study it had commissioned. As discussed above, the DIVA Market Study recommended a Phased Acquisition and Integration Approach as a poten- tial strategy. In this phased strategy, the Army would 5 The term “commercial item” is defined at 41 U.S.C. § 103 and FAR 2.101, which apply government- wide. 10 U.S.C. § 2302. Regarding products, a “commer- cial item” is defined as an “item . . . that is of a type customarily used by the general public or by non- governmental entities for purposes other than govern- mental purposes; and—(i) [h]as been sold, leased or licensed to the general public; or (ii) [h]as been offered for sale, lease, or license to the general public.” FAR 2.101. The definition is broad enough to include products that would meet the above provisions, but for “[m]odifications of a type customarily available in the commercial market- place,” or “[m]inor modifications of a type not customarily available in the commercial marketplace made to meet Federal Government requirements.” Id. “Minor modifica- tions means modifications that do not significantly alter the nongovernmental function or essential physical char- acteristics of an item or component, or change the purpose of a process.” Id. PALANTIR USG, INC. v. UNITED STATES 19 first acquire the two foundation components: a COTS cloud infrastructure service and a COTS DIVA “Turn Key” platform. Next, integrating these two components with each other and the DCGS-A Enterprise data man- agement architecture would establish a baseline DCGS- A2—a core suite of applications and analytics functions; a new data management architecture. Indeed, the DIVA Market Study explained that “[a] key advantage of lever- aging COTS cloud infrastructure services and a COTS DIVA platform is that doing so provides a significant amount of technical infrastructure and end-user capabili- ties.” J.A. 12251. Thus, the Army was aware of a possible commercial approach for at least portions of the DCGS-A2 procurement. Similarly, just six months before the solicitation, the Army Materiel Systems Analysis Activity’s July 2015 Trade Space Analysis indicated that a hybrid approach using commercial items was the best of three alternatives considered, including COTS, GOTS, and hybrid. The Trade Space Analysis described this hybrid approach as a “compilation of commercially available software packages augmented with integrated tools/widgets written by a third-party using requirements/specifications generated by the Government (i.e., combination of COTS and GOTS).” J.A. 11954. Notably, the analysis acknowledged that “[h]ybrid software option alternatives are currently functioning in the [Department of Defense Intelligence Community]” and would “only require minor development to fill capability gaps.” J.A. 11976. Palantir also put the Army on notice of its capabilities to provide a commercial item that could be modi- fied/integrated to meet the Army’s needs for DCGS-A2. Palantir responded to the Army’s RFI #1 by explaining that the Army should consider existing commercial solu- tions: 20 PALANTIR USG, INC. v. UNITED STATES The acquisition cycle should fully leverage exist- ing commercial solutions. Prioritizing the rapid procurement of commercial capabilities minimizes the anticipated scope of development needed to deliver Increment 2 capabilities. Narrowing the development scope requires expanding the use of commercially available COTS capabilities—it does not require narrowing the overall scope of the DCGS-A program. The Government does not need to build Increment 2 functionality; the Govern- ment can buy the core functionality from the commercial market and integrate any number of additional applications. J.A. 11885 (footnote omitted). Palantir explained that “we recommend the Government pursue a different acqui- sition strategy than the strategy behind the Increment 1 challenges.” Id. Palantir further informed the Army that it had successful contracts with the U.S. Marine Corps, U.S. Immigration and Customs Enforcement, and the Defense Intelligence Agency with its COTS solution. Suggesting that the COTS approach would also work for DCGS-A2, Palantir proposed using a firm-fixed-price (“FFP”) model with an “outcomes-based Performance Work Statement based on a proven product and incorpo- rating support services.” J.A. 11889. The Army’s December 2014 RFI Response Analysis includes a summary of Palantir’s response: Palantir has developed an intelligence fusion sys- tem that has been used by various entities within the Department of Defense. Palantir was found capable to provide Data management and Work- flow Management upgrades, and partially capable of providing Data Fusion and Cyber capabilities to Increment 2. J.A. 11868. This confirms that the Army was aware of Palantir’s commercially available intelligence fusion PALANTIR USG, INC. v. UNITED STATES 21 system, which was already in use within the Department of Defense, and considered Palantir capable of delivering some of the required functionality of DCGS-A2. Palantir again tried to explain the value of a commer- cial—rather than developmental—approach in response to the Army’s RFI #2: We continue to believe that the success of Incre- ment 2 requires a proven commercial solution to ensure the delivery of a working capability on time and within budget. We are concerned that the present RFI . . . is focused on collecting infor- mation on each respondent’s ability to conduct a services-based, large-scale, and custom software engineering effort . . . rather than to assess exist- ing software capabilities applicable to Increment 1 capability gaps. J.A. 11910. Likewise, Palantir’s response to RFI #3 explained that “[i]n cooperation with the government, Palantir fields and manages 25 Palantir deployments at every major Marine Corps command, representing over 15,000 accounts at peak usage across the Marine Intelli- gence community.” J.A. 11922. Palantir further stated that “Increment 2 should use a fielded commercial solu- tion” and that “[d]elivering Increment 2 on a commercial platform ensures the data layer advances at the same pace as commercial technology.” J.A. 11918. Additional- ly, Palantir flagged that it thought the Army’s “initial decision to embark on a significant software development effort, rather than acquiring a COTS solution,” would cause challenges like the ones that faced DCGS-A1. Id. In addition, the Administrative Record includes three Operational Needs Statements from other Department of Defense personnel requesting Palantir’s data manage- ment platform. One such statement, dated February 2015, explained that “[t]he Palantir Command platform is a proven capability that is currently in use to provide 22 PALANTIR USG, INC. v. UNITED STATES COP, data integration, and staff integration capabilities across multiple commercial and government organiza- tions.” CFC Op., 129 Fed. Cl. at 224. It further stated that Palantir “offers a solution that meets all of our requirements.” Id. Based on this record, we agree with the trial court that the Army was, or should have been, aware of Palan- tir’s data management platform. Despite repeated notice that commercial products might well be available and could be modified to meet the Army’s needs, the Army concluded that DCGS-A2 could not be procured as a commercial product with scant explanation. Indeed, the Army’s July 2015 Market Research Report simply stated that “[s]ignificant portions of the anticipated Increment 2 scope of work such as Data Fusion, Intelligence Support to Cyber, and [DCGS Integrated Backbone] upgrade are not available as a commercial product. As such, the [DCGS-A2] development effort cannot be procured as a commercial product.” J.A. 11840 ¶ 8.3.5; CFC Op., 129 Fed. Cl. at 231. It concluded that “[b]ased on the Market Research to date, the recommended approach is a five (5) year Engineering and Manufacturing Develop- ment (EMD) effort consisting of two releases.” J.A. 11841. There was no discussion in the Market Research Report to support the Army’s conclusory assessment that these three requirements—data fusion, intelligence support, and DCGS integrated backbone—were not commercially available. Nor was there any discussion of whether any commercial items could have been modified to meet the Army’s needs or the Army’s requirements could have been modified so that commercial items could be used. Further, on the first requirement—data fusion— record evidence shows that Palantir Gotham may provide “data fusion” capability. The Army itself described Palan- tir Gotham as being used as an “integrated fusion and analysis platform.” J.A. 18183–84. In addition, Palantir explained in its response to the Army’s draft performance PALANTIR USG, INC. v. UNITED STATES 23 work statement that the commercial market offers nu- merous existing tools with this capability. Palantir took a step further and explained that this requirement, as written, envisioned building the capability from scratch instead of evaluating whether such functionality was commercially available. The record evidence likewise demonstrates that the second requirement—intelligence support to cyber—may have been commercially available. Indeed, the July 2013 MITRE Palantir Platform Information Brief, which is in the administrative record, noted that “Palantir has NETOPS capability to audit/log potential cyber events and has Cyber Analysis Tools to detect/analyze suspicious Cyber events.” J.A. 17851. Furthermore, Palantir ex- plained in its response to the draft performance work statement that having a separate requirement for cyber intelligence functionality is unnecessary because the Army could acquire such intelligence support to cyber capabilities by simply acquiring the Palantir Gotham Data Management Platform. Finally, the record demonstrates that Palantir Go- tham could be interoperable with the existing DCGS integrated backbone. For example, Palantir contracted with a U.S. military command to provide Gotham as an information bridging solution, including to satisfy the requirement that the data structure would support evolv- ing DCGS integrated backbone standards and upgrades to new versions. Furthermore, Palantir explained in its response to the draft performance work statement that the proposed requirement to integrate the DCGS inte- grated backbone is unnecessary because it should be treated as an interoperability standard, not a software platform that should be integrated. On this record, we agree with the trial court that the Army failed in its obligation under § 2377 to determine whether a commercial item could meet or be modified to 24 PALANTIR USG, INC. v. UNITED STATES meet the Army’s procurement requirements. We acknowledge that there is no statutory or regulatory requirement for agencies to document their determina- tions pertaining to § 2377 and FAR Part 10. See Ad- vanced Am. Constr., Inc. v. United States, 111 Fed. Cl. 205, 227 (2013) (“[T]he language of section 10.002(e) is precatory in nature and does not establish any mandatory documentation requirement. That section states that agencies ‘should’ document the results of their market research; it does not state that those agencies ‘shall’ do so.”). Nevertheless, the record must be sufficient to permit meaningful judicial review consistent with the Administrative Procedure Act, 5 U.S.C. § 706. See Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983) (“[T]he agency must examine the relevant data and articulate a satisfactory explanation for its action including a ‘rational connection between the facts found and the choice made.’” (quoting Burlington Truck Lines v. United States, 371 U.S. 156 (1962))); see also Sierra Club v. Gorsuch, 715 F.2d 653, 660–61 (D.C. Cir. 1983) (“If there is reasoned deci- sionmaking lurking behind such agency behavior, it is yet to be articulated. For agency action to be upheld, it must not only be explainable; it must also be explained.”); Bagdonas v. Dep’t of Treasury, 93 F.3d 422, 426 (7th Cir. 1996) (“The statement of reasons need not include de- tailed findings of fact but must inform the court and the petitioner of the grounds of decision and the essential facts upon which the administrative decision was based.” (citing Kitchens v. Dep’t of Treasury, 535 F.2d 1197, 1199–1200 (9th Cir. 1976))); Impresa Construzioni Geom. Domenico Garufi v. United States, 238 F.3d 1324, 1338 (Fed. Cir. 2001) (“Garufi”) (citing Supreme Court deci- sions establishing that, even if the agency is not obligated to provide reasons, a court may nonetheless order the agency to provide explanation if such an explanation is required for meaningful judicial review). Here, the ad- ministrative record plainly shows that the Army was on PALANTIR USG, INC. v. UNITED STATES 25 notice that Palantir’s product might be a commercial item that would satisfy its requirements, whether as-is or with modifications. Despite that notice, the Army’s ultimate determination regarding its market research excluded commercial items from consideration in a conclusory fashion. On this record, we conclude that the Army did not rationally use its market research results to deter- mine whether there are available commercial items that: “(A) meet the agency’s requirements; (B) could be modi- fied to meet the agency’s requirements; or (C) could meet the agency’s requirements if those requirements were modified to a reasonable extent.” § 2377(c)(2); FAR 10.001(a)(3)(ii). The government argues that the Trade Space Analy- sis demonstrates that it satisfied its obligations under § 2377. In particular, the government asserts that this document shows that a hybrid approach—using “commer- cially available software” and software developed by the government (GOTS)—was superior to a commercial item procurement. Appellant Reply Br. at 9. But the record undermines the government’s position, showing that the Army’s procurement efforts were focused on a develop- mental approach without determining the viability of a commercial or even a hybrid approach. Indeed, in its Determination & Findings for Award of a Single Source IDIQ Single Award Contract, the government emphasized that DCGS-A2 “is heavily focused on design and develop- ment of a new data management architecture by a con- tractor as the systems integrator.” J.A. 12299 ¶ 4 (emphasis added). The government further emphasized that “[d]evelopment of the data integration layer is pivot- al.” Id. Nowhere in that document does the government address implementation of the Trade Space Analysis’s recommended hybrid approach or Palantir’s proposed COTS approach. Moreover, although the Trade Space Analysis recommended a hybrid approach using commer- cial items, the government did not use that information to 26 PALANTIR USG, INC. v. UNITED STATES determine whether a commercial item could be modified to meet the agency’s requirements. See § 2377(c)(2). As such, the Trade Space Analysis does not alter our conclu- sion that the Army did not comply with § 2377. III. The Trial Court Properly Accounted for the Presumption of Regularity We now turn to the government’s second ground for challenging the trial court’s judgment. The government alleges that the trial court wrongly discarded the pre- sumption of regularity in determining that the Army’s action was arbitrary and capricious and did not comply with § 2377(c)(2). We do not agree. Under the Administrative Procedure Act, even where an explanation or reason is not required for an agency’s determination, a reviewing court has the power to require an explanation. Garufi, 238 F.3d at 1338. “[I]n determin- ing whether to require an explanation, the agency deci- sion is entitled to a presumption of regularity.” Id. (citing Bowen v. Am. Hosp. Ass’n, 476 U.S. 610, 626–27 (1986)). “Because of that presumption of regularity, the agency should not be required to provide an explanation unless that presumption has been rebutted by record evidence suggesting that the agency decision is arbitrary and capricious.” Id. Here, the court extensively cited record evidence showing that the Army’s decision was arbitrary and capricious and in violation of 10 U.S.C. § 2377. In partic- ular, the court performed a searching review and analysis of the DIVA Study, Trade Space Analysis, RFIs and RFI responses, July 2015 Market Research Report, Octo- ber 21, 2015 Determination & Findings for Award of a Single Source IDIQ Single Award Contract, and July 1, 2016 Determination of Non-Commercial Item. Based on this review, it concluded that the Army neglected to determine whether possible commercially available alter- natives meet or could be modified to meet the require- PALANTIR USG, INC. v. UNITED STATES 27 ments of the Army’s acquisition. See CFC Op., 129 Fed. Cl. at 275–82. Accordingly, the court properly determined that the record evidence rebutted the pre- sumption of regularity. CONCLUSION We do not reach the Court of Federal Claims’ finding of prejudice because the government does not contest it. Therefore, we need not reach its argument that the Court of Federal Claims erred in admitting the expert testimony of Mr. Bryant Choung, which the Court of Federal Claims relied on solely for its prejudice analysis. We have considered the government’s remaining ar- guments and find them unpersuasive. We affirm the judgment of the Court of Federal Claims that the Army must satisfy the requirements of 10 U.S.C. § 2377, which, thus far, the Army has failed to do. Only after the Army has complied with 10 U.S.C. § 2377 should it proceed to award a contract to meet its DCGS-A2 requirements. To be clear, we are not suggesting that the Army must choose Palantir as the awardee. We simply affirm that the Army must satisfy the requirements of 10 U.S.C. § 2377. AFFIRMED COSTS Costs to Appellee.