MOORE, Circuit Judge.
Amphastar Pharmaceuticals, Inc., International Medication Systems, Ltd., Watson Pharmaceuticals, Inc., and Watson Pharma., Inc. (collectively, Amphastar) appeal the district court’s order denying the Emergency Motion to Dissolve or Stay the preliminary injunction entered in this case. Because the district court applied an unduly narrow interpretation of the HatchWaxman safe harbor, 35 U.S.C. § 271(e)(1), we vacate the grant of a preliminary injunction and remand for further proceedings consistent with this opinion.
Background
This case is a patent litigation involving a generic version of Lovenox (enoxaparin), a drug that prevents blood clots. Enoxaparin is a low molecular weight version of heparin, a naturally occurring molecule. Heparin is a polymer, known as a polysaccharide, made up of long chains of sugar molecules. Heparin is not a single defined molecule. Instead, heparin molecules have considerable diversity in (1) the length of the polysaccharide chain and (2) in the component disaccharide units and the corresponding distribution of disaccharide unit sequences in the polysaccharide chains. FDA Letter to Aventis Pharmaceuticals, Inc., July 23, 2010, FDA Docket No. FDA-2003-P-0273 (FDA Letter), J.A. [1350]*1350291. For example, the molecular weight of heparin molecules varies between 5,000 and 40,000 daltons. Id. Likewise, the disaccharide units can vary between two different uronic acid components, and each of four positions on the disaccharide unit can be modified. Id., J.A. 291-92. The natural diversity inherent to heparin stems from the biosynthetic pathway used to produce the molecule. Id., J.A. 292.
Enoxaparin is produced by breaking the heparin polysaccharide into smaller pieces, called oligosaccharides. Because the heparin starting material is a diverse set of molecules, enoxaparin is also made up of different chain lengths and disaccharide units corresponding to the diversity in the original mix of heparin molecules. Id. Additional diversity is introduced based on the way in which the heparin molecule is broken down into the low molecular weight heparin product. Id., JA 292-93. Thus, unlike a typical small molecule drug like penicillin, enoxaparin is made up of a range of different molecules.
This molecular diversity raises a potential problem in light of the Food and Drug Administration’s (FDA’s) abbreviated new drug application (ANDA) approval process. AND As are typically used by generic companies to obtain approval to market a generic version of an existing drug. Unlike a new drug application (NDA), an ANDA applicant is not required to submit the same extensive clinical studies typically needed to prove the drug’s safety and efficacy. Instead, the ANDA applicant must submit studies to establish that its drug is bio-equivalent to the reference drug. The ANDA must also include sufficient information to establish that the generic drug has the same active ingredients as the reference drug.
The obvious complication with using an ANDA application to gain approval for enoxaparin is that it is a mixture of a number of different low molecular weight heparin molecules. In fact, Aventis, which marketed Lovenox, asked the FDA to deny approval for a generic version of enoxaparin via an ANDA unless the applicant either (1) completely characterized enoxaparin by isolating, purifying, and sequencing each of its unique polysaccharide chains, which Aventis claimed was impossible; (2) used Aventis’s manufacturing process; or (3) conducted clinical trials to prove safety and efficacy (the very type of duplicative studies the ANDA approval process was designed to avoid). FDA Letter, J.A. 286. The FDA rejected Aventis’s arguments, and instead explained that the ANDA “statutory provisions do not describe the type or amount of information that an ANDA applicant must submit to demonstrate that the active ingredient in the generic drug product is the same as the active ingredient in the [reference drug].” Id., J.A. 294. As a result, the FDA concluded that Congress recognized that the FDA has “broad discretion with respect to the information [it] may consider in making a finding on the ‘sameness’ of an active ingredient.” Id.
Consistent with this discretion, the FDA identified five criteria, or “standards for identity,” that “together provide sufficient information to conclude that generic enoxaparin has the ‘same’ active ingredient as Lovenox.” Id., J.A. 295. These criteria included, inter alia, “[equivalence in disaccharide building blocks, fragment mapping, and sequence of oligosaccharide species.” Id. The FDA explained that such equivalence is proven by “exhaustive digestion of enoxaparin with purified heparin digesting enzymes (heparinases I, II, III) and nitrous acid, among other means, to yield the constituent disaccharide building blocks comprising enoxaparin.” Id., J.A. [1351]*1351300. These disaccharides can then potentially be “separated and quantified” by a number of techniques, including capillary electrophoresis (CE), reverse phase high performance liquid chromatography (RPHPLC), and strong anion exchange high performance liquid chromatography (SAXHPLC). Id.
The FDA also suggested the identity of the disaccharides could be determined via standard techniques, including mass spectroscopy, NMR spectroscopy, modifying reagents, or modifying enzymes. These techniques identify the nature of the constituent sugars and their substitution patterns, including the sulfation and acetylation patterns, as well as “whether the disaccharide possesses, among other structures, a ... 1, 6 anhydro ring” structure. Id., J.A. 300-01. Detecting the presence of a 1, 6 anhydro ring structure is particularly important for proving equivalence because “[equivalence in disaccharide building blocks together with equivalence in molecular weight distribution shows that generic enoxaparin contains the 1, 6 anhydro ring structure at the reducing ends of between 15 percent and 25 percent of its poly(oligo)saccharide chains.” Id. n. 68, J.A. 301.
Amphastar was the first company to file an ANDA for a generic version of enoxaparin. It submitted its ANDA to the FDA in March 2003, and subsequently engaged in a lengthy patent litigation with Sanofb-Aventis. Amphastar received FDA approval to market its generic enoxaparin on September 19, 2011. Despite the fact that Amphastar was the first company to file an ANDA, Momenta Pharmaceuticals, Inc. and Sandoz, Inc. (collectively Momenta), who collaborated to develop a generic enoxaparin product, were the first to bring generic enoxaparin to the market-place. Momenta received FDA approval to market enoxaparin in July 2010, more than a year before Amphastar’s approval. Being the only generic version of enoxaparin has it benefits: its sales generated revenues of $260 million per quarter. J.A. 189. The approval of Amphastar’s version of enoxaparin, and the resultant ruinous competition of another generic version of the drug, threatened this unique market position. Understandably unwilling to give up a billion dollars in yearly revenue, Momenta initiated the present litigation two days after Amphastar received final FDA approval to market its generic enoxaparin.
Momenta is the assignee of United States Patent No. 7,575,886 ('886 patent). The '886 patent generally relates “to methods for analyzing heterogeneous populations of sulfated polysaccharides, e.g. heparin [and] ... LMWH [e.g., enoxaparin.]” '886 patent col.4 11.53-55. Claim 6 is typical. It is a method for analyzing an enoxaparin sample “for the presence or amount of a non naturally occurring sugar ... that results from a method of making enoxaparin that included (3-eliminative cleavage with a benzyl ester and depolymerization.” Id. col.64 11.35-39. Momenta also asserted independent claims 15, which assesses the level of non-naturally occurring sugar, and 53, which allows selection of an appropriate batch. These claims are similar to claim 6. The asserted claims generally require digestion of an enoxaparin sample with a heparin degrading enzyme, followed by the use of a separation method to detect the presence of the non-naturally occurring sugar resulting from the B-eliminative cleavage. The signal corresponding to the non-naturally occurring sugar can then be used to analyze the test sample based on a comparison with a reference standard. Id. col.6411.40-57.
[1352]*1352Momenta alleged that Amphastar infringed the '886 patent by “manufacturing generic enoxaparin for commercial sale” using the claimed methods. J.A. 58. Momenta asserted that Amphastar “included in their process for manufacturing batches of enoxaparin sodium ... a method for determining that a defined percentage of the oligosaccharide chains that make up enoxaparin include ... a non-naturally occurring sugar that includes a 1, 6-anhydro ring structure, which method infringes the '886 patent.” J.A. 57. Momenta also alleged that this infringing testing was necessary because the “FDA requires a generic manufacture to include in its manufacturing process the analysis of each batch of its enoxaparin drug substance to confirm that ... [it] includes a 1, 6-anhydro ring structure.” J.A. 56. Momenta moved for and received a temporary restraining order to prevent the irreparable harm of additional generic entry from Amphastar. J.A. 4. The district court subsequently granted Momenta a preliminary injunction based on its belief that Amphastar’s quality control batch testing infringed the '886 patent. J.A. 30. Amphastar later filed two emergency motions for relief from the preliminary injunction, which the district court denied.
Amphastar sequentially appealed the preliminary injunction and the two denials for relief from the preliminary injunction. These three appeals were consolidated. We have jurisdiction to hear these appeals pursuant to 28 U.S.C. § 1292. After hearing oral argument in this case, we stayed the preliminary injunction. This stay, however, was not a final decision on the merits of Amphastar’s appeal. We now explain why the district court incorrectly concluded that Momenta was likely to succeed on the merits of its infringement claim, and conclude that the preliminary injunction must be vacated.
Analysis
“The issuance of a preliminary injunction ... is a matter of discretion for a district court. That discretion, however, is not absolute and must be reviewed in light of the equitable standards governing the issuance of injunctions.” Intel Corp. v. ULSI Sys. Tech., Inc., 995 F.2d 1566, 1568 (Fed.Cir.1993). To determine whether a preliminary injunction is appropriate, the district court weighs factors including “(1) whether the movant has sufficiently established a reasonable likelihood of success on the merits; (2) whether the movant would suffer irreparable harm if the injunction were not granted; (3) whether the balance of hardships tips in the movant’s favor; and (4) the impact, if any, of the injunction on the public interest.” Id. The grant of a preliminary injunction can be overturned “by showing that the court made a clear error of judgment in weighing relevant factors or exercised its discretion based upon an error of law or clearly erroneous factual findings.” Genentech, Inc. v. Novo Nordisk A/S, 108 F.3d 1361, 1364 (Fed.Cir. 1997). As the party seeking the injunction, the burden is on Momenta to establish it is entitled to this extraordinary relief. Id. In order to prove a likelihood of success on the merits, Momenta must prove that Amphastar likely infringes its patent. Id. Conversely, if Amphastar establishes that Momenta is unlikely to succeed on its claim of infringement, a preliminary injunction is likely not appropriate. Id.
In its opposition to the preliminary injunction, Amphastar argued, among other things, that its testing falls within the scope of the Hatch-Waxman safe harbor, 35 U.S.C. § 271(e)(1). Section 271(e)(1) indicates that “[i]t shall not be an act of infringement to ... use ... a patented [1353]*1353invention ... solely for uses reasonably related to the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs.... ” The district court found that “the alleged infringing activity involves the use of plaintiffs’ patented quality control testing methods on each commercial batch of enoxaparin that will be sold after FDA approval.” J.A. 31; see also J.A. 56 (Momenta’s complaint alleging that the “FDA requires” the testing). While acknowledging that Amphastar’s use of the patented method was for the purpose of developing information to submit to the FDA, the district court nevertheless concluded that the safe harbor does not apply to Amphastar’s testing: “although the safe harbor provision permits otherwise infringing activity that is conducted to obtain regulatory approval of a product, it does not permit a generic manufacturer to continue in that otherwise infringing activity after obtaining such approval.” J.A. 23. In reaching this conclusion, the district court focused primarily on the legislative history of the safe harbor, as quoted in one of our prior cases, Classen Immunotherapies, Inc. v. Biogen IDEC, 659 F.3d 1057 (Fed. Cir.2011). J.A.23.
On appeal, Amphastar argues that the district court took an unduly restrictive view of the safe harbor, and that its activities fall within the plain language of 35 U.S.C. § 271(e)(1). Momenta counters that the district court correctly held that the safe harbor does not apply to Amphastar’s testing for two reasons. First, Momenta argues that the safe harbor does not apply to post-approval activity: “In Classen, this court squarely held that ‘[t]he [safe harbor] does not apply to information that may be routinely reported to the FDA long after marketing approval has been obtained.’ ” Appellee’s Br. at 43 (quoting Classen, 659 F.3d at 1070, alterations made by Momenta). Because Amphastar’s batch testing is carried out as a condition for the post-FDA approval sale of enoxaparin, Momenta argues it falls outside the scope of the safe harbor.
Second, despite its allegations and concessions, Momenta asserts the safe harbor does not apply because “the FDA does not require the use of the particular procedure that is claimed in the '886 patent.” Id. at 41. Instead, Momenta claims that the FDA’s interpretation of its statutory mandate in its letter response to Aventis’s petition, J.A. 300-01, allows a variety of testing methods to be used to establish equivalence, both for the submission of an ANDA and for the undisputedly required batch testing. Appellee’s Br. at 41. Momenta argues that the availability of other acceptable testing methods means that Amphastar’s alleged use of the patented method is not required by the FDA, and is therefore outside of the safe harbor provision.
The parties thus present us with conflicting views about the scope of the safe harbor. If Amphastar is correct that its post-approval activities actually fall within the scope of 35 U.S.C. § 271(e)(1), Momenta is unlikely to succeed on its claim of infringement and the preliminary injunction is likely inappropriate. Genentech, 108 F.3d at 1364. In order to determine whether the preliminary injunction was appropriate in this case, we must first ascertain the scope of the Hatch-Waxman safe harbor provision, 35 U.S.C. § 271(e)(1).
I.
“[A]ll statutory construction cases ... begin with the language of the statute.” Barnhart v. Sigmon Coal Co., 534 U.S. [1354]*1354438, 450, 122 S.Ct. 941, 151 L.Ed.2d 908 (2002). The “first step in interpreting a statute is to determine whether the language at issue has a plain and unambiguous meaning with regard to the particular dispute in the case.” Robinson v. Shell Oil Co., 519 U.S. 337, 340, 117 S.Ct. 843, 136 L.Ed.2d 808 (1997). If the language of the statute is unambiguous, there is no second step: “Our inquiry must cease if the statutory language is unambiguous and ‘the statutory scheme is coherent and consistent.’ ” Id. (quoting United States v. Ron Pair Enters., Inc., 489 U.S. 235, 240, 109 S.Ct. 1026, 103 L.Ed.2d 290 (1989)). Whether the text of a statute is plain or ambiguous “is determined by reference to the language itself, the specific context in which the language is used, and the broader context of the statute as a whole.” Id. at 341, 117 S.Ct. 843.
The Drug Price Competition and Patent Term Restoration Act (Hatch-Waxman Act), Public Law No. 98-417 (1984) (codified in relevant part at 35 U.S.C. § 271(e)) set up a statutory system to “balance the need to stimulate innovation against the goal of furthering the public interest.” H.R. Rep. 98-857, pt. 2, at 2714 (Aug. 1, 1984), 1984 U.S.C.C.A.N. 2686, 2714. This balance is embodied, in part, in the “safe harbor” provision of 35 U.S.C. § 271(e)(1), which provides (with emphasis added) that:
It shall not be an act of infringement to make, use, offer to sell, or sell within the United States or import into the United States a patented invention (other than a new animal drug or veterinary biological product (as those terms are used in the Federal Food, Drug, and Cosmetic Act and the Act of March 4, 1913) which is primarily manufactured using recombinant DNA, recombinant RNA, hybridoma technology, or other processes involving site specific genetic manipulation techniques) solely for uses reasonably related to the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs or veterinary biological products.
Congress could not have been clearer in its choice of words: as long as the use of the patented invention is solely for uses “reasonably related” to developing and submitting information pursuant to “a Federal law” regulating the manufacture, use, or sale of drugs, it is not “an act of infringement.”
Although the Hatch-Waxman safe harbor provision was enacted in the context of the then-novel ANDA approval process, 35 U.S.C. § 271(e)(1) does not reference the portion of the Federal Food, Drug, and Cosmetic Act describing the ANDA requirements, e.g., 21 U.S.C. § 355(j). Instead, Congress used more flexible and expansive language to define the scope of § 271(e)(1), referring generally to “the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs.” This broad language unambiguously applies to submissions under any federal law, providing that the law “regulates the manufacture, use, or sale of drugs.” Limiting the scope of 35 U.S.C. § 271(e)(1) to just the submission of information pursuant to the Federal Food, Drug, and Cosmetic Act generally, or to the ANDA provision of the Federal Food, Drug, and Cosmetic Act in specific, would read words into the statute in violation of the express language chosen by Congress.
This interpretation is also consistent with the rest of the statutory scheme. When Congress wanted to impose a limita[1355]*1355tion based on the Federal Food, Drug, and Cosmetic Act, it expressly referenced the Act. For example, in the safe harbor provision, Congress excluded “a new animal drug or veterinary biological product (as those terms are used in the Federal Food, Drug, and Cosmetic Act and the Act of March 4, 1913)” made using certain genetic techniques. 35 U.S.C. § 271(e)(1) (emphasis added). Likewise, when Congress wanted to limit the statute to just a certain kind of submission, for example the submission of an ANDA application under 21 U.S.C. § 355(j), it specifically referenced the statutory section governing those submissions. For example, in the subsection immediately following the safe harbor, Congress defined as an act of infringement the submission of “an application under section 505(j) of the Federal Food, Drug, and Cosmetic Act [codified at 21 U.S.C. § 355(j) ] or described in section 505(b)(2) of such Act for a drug claimed in a patent or the use of which is claimed in a patent.” 35 U.S.C. § 271(e)(2)(A).
Unlike the closely related infringement provision, 35 U.S.C. § 271(e)(2), Congress did not link the safe harbor to the submission of an application for approval under the Federal Food, Drug, and Cosmetic Act. Compare 35 U.S.C. § 271(e)(1) (not an act of infringement when used for “the development and submission of information under a Federal law”) with 35 U.S.C. § 271(e)(2)(A) (it is an act of infringement to submit “an application under section 505(j) of the Federal Food, Drug, and Cosmetic Act or described in section 505(b)(2) of such Act”). We cannot change the statutory language. We will not import the limitation of § 271(e)(2) into § 271(e)(1). “[Ojur obligation is to take statutes as we find them.” Diamond v. Chakrabarty, 447 U.S. 303, 315, 100 S.Ct. 2204, 65 L.Ed.2d 144 (1980); see also, e.g., Reiter v. Sonotone Corp., 442 U.S. 330, 344, 99 S.Ct. 2326, 60 L.Ed.2d 931 (1979) (“We must take the statute as we find it.”). The statute here applies to any use of a patented invention as long as the use is “reasonably related to the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs----” 35 U.S.C. § 271(e)(1).
In light of these provisions, the only coherent and consistent interpretation of “a Federal law which regulates the manufacture, use, or sale of drugs” is that it must be broad enough to encompass submissions made pursuant to the Federal Food, Drug, and Cosmetic Act. Since there is no ambiguity in the language used by Congress in 35 U.S.C. § 271(e)(1), our inquiry into the scope of the safe harbor is complete. Robinson, 519 U.S. at 340, 117 S.Ct. 843. When the intent of Congress is expressed so clearly and consistently throughout the statute, there is neither the need nor the occasion to refer to the legislative history. Id. The scope of the Hatch-Waxman safe harbor does not stop at activities reasonably related to development of information submitted in an ANDA. Instead, the safe harbor applies “to the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs or veterinary biological products.” As long as the allegedly infringing use is “for uses reasonably related” to the development and submission of that information it is not an act of infringement, regardless of where that requirement resides in the law.
This analysis is not groundbreaking: the Supreme Court came to essentially the same conclusion in 1990. In Eli Lilly & Co. v. Medtronic, Inc., the Court explained that “the phrase ‘a Federal law which regulates the manufacture, use, or sale of [1356]*1356drugs’ more naturally summons up the image of an entire statutory scheme of regulation,” and not just a particular provision of the law. 496 U.S. 661, 666, 110 S.Ct. 2683, 110 L.Ed.2d 605 (1990) (emphasis added). Although the legislative history of the safe harbor only mentioned drugs, id. at 669 n. 2, 110 S.Ct. 2683, the Court nevertheless concluded that the safe harbor also extended to medical devices, which were also part of “a Federal law which regulates the manufacture, use or sale of drugs,” namely the Federal Food, Drug, and Cosmetic Act, id. at 674, 110 S.Ct. 2683.
The Court later reaffirmed this expansive view, explaining: “we think it apparent from the statutory text that § 271(e)(l)’s exemption from infringement extends to all uses of patented inventions that are reasonably related to the development and submission of any information under the FDCA [ (Food, Drug, and Cosmetic Act) ].” Merck KGaA v. Integra Lifesciences I, Ltd., 545 U.S. 193, 202, 125 S.Ct. 2372, 162 L.Ed.2d 160 (2005) (citing Eli Lilly, 496 U.S. at 665-69, 110 S.Ct. 2683). Merck KGaA expressly rejected the notion that the safe harbor only applies to information developed during a clinical trial. 545 U.S. at 202 n. 6, 125 S.Ct. 2372. Instead, “the statutory text makes clear that it provides a wide berth for the use of patented drugs in activities related to the federal regulatory process.” Id. at 202, 125 S.Ct. 2372 (emphasis added). In light of the unqualified exemption for uses reasonably related to the development and submission of information, “[tjhere is simply no room in the statute for excluding certain information from the exemption on the basis of the phase of research in which it is developed or the particular submission in which it could be included.” Id. (emphasis added). The use of the word “under” in the statute is expansive. Caraco Pharm. Labs., Ltd. v. Novo Nordisk A/S, 566 U.S.-, 132 S.Ct. 1670, 1683-84, 182 L.Ed.2d 678 (2012). “Under a federal law” extends beyond just the “most barebones information” required by the FDA, and instead encompasses all “materials the FDA demands in the regulatory process.” Id.
While it is clear that the safe harbor applies to a broad set of “activities related to the federal regulatory process,” Merck KGaA, 545 U.S. at 202, 125 S.Ct. 2372, there is an important limitation: the use must be “for uses reasonably related to the development and submission of information,” 35 U.S.C. § 271(e)(1). “Reasonably related,” however, does not mean that the use of the patented invention must necessarily result in submission of information to the FDA: “Congress did not limit § 271(e)(l)’s safe harbor to the development of information for inclusion in a submission to the FDA; nor did it create an exemption applicable only to the research relevant to filing an ANDA for approval of a generic drug.” Merck KGaA 545 U.S. at 206, 125 S.Ct. 2372. Instead, the Court explained that the safe harbor “exempted from infringement all uses of patented compounds ‘reasonably related’ to the process of developing information for submission under any federal law regulating the manufacture, use, or distribution of drugs.” Id. (emphasis in original). Thus, the Court explicitly rejected the notion that § 271(e)(1) was limited “to the activities necessary to seek approval of a generic drug.” Id. As long as the accused infringer “has a reasonable basis for believing” that use of the patented invention might yield information that “would be appropriate to include in a submission to the FDA, that use is ‘reasonably related’ to the ‘development and submis[1357]*1357sion of information under ... Federal law.’ ” Id. at 207, 125 S.Ct. 2372.
II.
At the outset we are met with the contention that the information in question was not “submitted” to the FDA, see 35 U.S.C. § 271(e)(1) (“... solely for uses reasonably related to the development and submission of information ... ”), but rather was retained by the ANDA holder. We do not agree. Amphastar, as a generic drug manufacturer under an ANDA, cannot sell a batch of enoxaparin unless it has established that its strength and quality is consistent with the standards set forth in the relevant official compendium. See 21 U.S.C. §§ 331(a), 351(b). FDA regulations require that all records associated with a produced batch of drugs, including these batch records, “be retained for at least 1 year after the expiration date of the batch.” 21 C.F.R. § 211.180(a). These records “shall be readily available for authorized inspection” by the FDA at any time. 21 C.F.R. § 211.180(c). We think that the requirement to maintain records for FDA inspection satisfies the requirement that the uses be reasonably related to the development and submission of information to the FDA. It is not disputed by the parties that these records are produced in order to develop and submit to the FDA proof that the Amphastar products comply with a Federal law. The fact that the FDA does not in most cases actually inspect the records does not change the fact that they are for the “development and submission of information under a Federal law.” 35 U.S.C. § 271(e)(1); cf. Merck KGaA, 545 U.S. at 207, 125 S.Ct. 2372 (holding that uses which are not ultimately included in a submission to the FDA are nonetheless exempted by the safe harbor). Thus, we consider this information “submitted” for purposes of the statute. We turn then to the question of whether these submissions are within the safe harbor.
In Merck KGaA v. Integra Lifesciences I, Ltd., 545 U.S. 193, 125 S.Ct. 2372, 162 L.Ed.2d 160 (2005), the Supreme Court held that uses of patented inventions in preclinical research, the results of which are not ultimately included in a submission to the FDA, are nevertheless exempted from infringement by the safe harbor provision. Id. at 208, 125 S.Ct. 2372. The Court explained that
Congress did not limit § 271(e)(l)’s safe harbor to the development of information for inclusion in a submission to the FDA; nor did it create an exemption applicable only to the research relevant to filing an ANDA for approval of a generic drug. Rather, it exempted from infringement all uses of patented compounds “reasonably related” to the process of developing information for submission under any federal law regulating the manufacture, use, or distribution of drugs.
Id. at 206, 125 S.Ct. 2372. Thus, it was not an act of infringement to use patented compounds in preclinical studies which were not ultimately submitted to the FDA where “there [was] a reasonable basis for believing that the experiments [would] produce the types of information that are relevant to an IND or NDA.” Id. at 208, 125 S.Ct. 2372.
However, in Classen Immunotherapies, Inc. v. Biogen IDEC, 659 F.3d 1057, 1070 (Fed.Cir.2011), we held that § 271(e)(1) “does not apply to information that may be routinely reported to the FDA, long after marketing approval has been ob[1358]*1358tained.” At issue in Classen were studies to evaluate the association between the timing of childhood vaccinations and the risk of developing certain immune-mediated disorders. The studies themselves were not mandated by the FDA, but any vaccine license holder was required to report to the FDA “adverse experience information,” such as adverse side effects, it acquired as a result of vaccine studies. See 21 C.F.R. § 600.80. We found that the studies conducted by the vaccine license holder according to patented methods were not insulated by the safe harbor because the studies did not facilitate marketing a generic drug by “expediting] development of information for regulatory approval.” Classen, 659 F.3d at 1070. We, of course, are bound by the Classen decision unless it is overruled en banc or by the Supreme Court. Accordingly, the scope of the safe harbor provision does not extend to “information that may be routinely reported to the FDA, long after marketing approval has been obtained.”
This case, however, fits well within Classen because the information submitted is necessary both to the continued approval of the ANDA and to the ability to market the generic drug. Here, the submissions are not “routine submissions” to the FDA, but instead are submissions that are required to maintain FDA approval. Amphastar is required to conduct a laboratory determination of identity and strength of the active ingredient for each batch of enoxaparin. See 21 C.F.R. § 211.165(a). This test must be done according to the patented methods described in an official compendium, in this case the United States Pharmacopeia (USP). See 21 U.S.C. § 351(b) (Any “determination as to strength, quality, or purity shall be made in accordance with the tests or methods of assay set forth in such compendium.”). Moreover, as described above, FDA regulations require that all such batch records “be retained for at least 1 year after the expiration date of the batch,” 21 C.F.R. § 211.180(a), and that such records “shall be readily available for authorized inspection” by the FDA at any time, 21 C.F.R. § 211.180(c); see also 21 C.F.R. §§ 211.186, 211.188, 211.194 (requiring “master production and control records,” “batch production and control records,” and “laboratory records”). Failure to comply with these requirements could result in suspension or revocation of Amphastar’s ANDA approval to market the drug. See 21 U.S.C. §§ 335a(g), 355(e). Furthermore, such testing is “a condition for [the drug’s] approval and release” into commerce, 21 C.F.R. § 211.165(d), thus acting as a predicate to the ability to market the ANDA-approved drug to the public.
The submissions to the FDA in this case are anything but “routine” — they implicate Amphastar’s very ability to continue its FDA approval for its ANDA and to continue manufacturing and marketing enoxaparin under its ANDA. We also note that, unlike in Classen where the patented studies performed were not mandated by the FDA, the information here is not generated voluntarily by the manufacturer but is generated by FDA requirements the manufacturer is obligated under penalty of law to follow. Under such circumstances, the information can be said to have been gathered solely for submission to the FDA and not, as in Classen, primarily for non-FDA purposes. While Momenta urges us to adopt the pre-/post-approval distinction used by the district court, we cannot: Classen did not turn on this artificial distinction, and the plain language of the statute is not restricted to pre-approval [1359]*1359activities.1 We therefore hold that post-approval studies that are “reasonably related to the development and submission of information under a Federal law which regulates the manufacture, use, or sale of drugs” fall within the scope of the § 271(e)(1) safe harbor.
In this case, Momenta concedes that Amphastar’s tests “are conducted in order to satisfy the FDA’s requirements that each batch of enoxaparin that is sold commercially after FDA approval is actually the same as the brand name drug.” Appellee’s Br. at 40-41 (emphasis added); see also J.A. 56 (allegation that the “FDA requires” the accused testing). Under a proper construction of 35 U.S.C. § 271(e)(1), the fact that Amphastar’s testing is carried out to “satisfy the FDA’s requirements” means it falls within the scope of the safe harbor, even though the activity is carried out after approval. Unlike Classen, where the allegedly infringing activity “may” have eventually led to an FDA submission, there is no dispute in this ease that Amphastar’s allegedly infringing activities are carried out to “satisfy the FDA’s requirements.” The district court’s interpretation of § 271(e)(1) was erroneous. Under the correct construction, Momenta cannot establish a likelihood of success on infringement and the preliminary injunction must be vacated. Genentech, Inc., 108 F.3d at 1364.
Momenta also argues that even if 35 U.S.C. § 271(e)(1) extends to post-approval activities, Amphastar’s testing is not protected because there are FDA endorsed non-infringing alternatives available. The safe harbor, however, does not mandate the use of a noninfringing alternative when one exists. The only limitation in the safe harbor is that the use must be “reasonably related to the development and submission of information” pursuant to a federal law regulating the “manufacture, use, or sale of drugs or veterinary biological products.” 35 U.S.C. § 271(e)(1). The safe harbor’s protection is not limited to the dire situation where the patented invention is the only way to develop and submit the information. Instead, the safe harbor expressly allows the submitter the freedom to use an otherwise patented means to develop the necessary information demanded by the “Federal law.” This makes good sense because it eliminates liability for infringement when that act of infringement is, in effect, required by the federal government as part of the continuing safety and efficacy monitoring of an approved drug. It also avoids the situation here, where a drug has received approval, but is nevertheless kept from the market based on an FDA mandated testing requirement.
Momenta’s interpretation is predicated upon the incorrect assumption that “solely” in the context of 35 U.S.C. § 271(e)(1) means that the patented invention must be the “sole” means of providing the information for the safe harbor to apply. This is not the language of the statute: under 35 U.S.C. § 271(e)(1), as long as the use is “reasonably related to the development and submission of information” under a relevant statute, it is not an act of infringement. “Solely” modifies “uses reasonably [1360]*1360related to the development and submission of information,” but does not place any other restriction on when the patented invention may be used without infringing. As long as the use of the patented invention is done to generate information that will be submitted pursuant to a relevant federal law, that use falls within the safe harbor. Merck KGaA, 545 U.S. at 205-206, 125 S.Ct. 2372. Momenta is therefore incorrect that the possibility that the FDA would accept the use of other, non-patented, testing methods for the development and submission of information precludes Amphastar from relying on the safe harbor in this case.2
Even if Momenta’s strained reading of the statute was supportable, Amphastar’s allegedly infringing activities are clearly carried out according to the dictates of the Federal Food, Drug, and Cosmetic Act. Under the Act, Amphastar is prohibited from selling a drug if it is adulterated. 21 U.S.C. § 331(a). A drug is adulterated if it purports to be a drug listed in an official compendium, for example the USP, but in actuality differs in composition. 21 U.S.C. § 351(b); see also 21 U.S.C. § 321© (defining “official compendium”). In order to demonstrate that a drug is not adulterated, testing must be carried out pursuant to the methods articulated in the compendium, in this case the USP. See 21 U.S.C. § 351(b) (Any “determination as to strength, quality, or purity shall be made in accordance with the tests or methods of assay set forth in such compendium.”). “For each batch of drug product, there shall be appropriate laboratory determination of ... the identity and strength of each active ingredient....” 21 C.F.R. § 211.165(a). FDA regulations characterize this testing as “a condition for [the drug’s] approval and release” into commerce. 21 C.F.R. § 211.165(d). The FDA also mandates maintenance of appropriate records related to this type of testing. See 21 C.F.R. § 211.180(a) (production, control, and distribution records associated with a batch of drug must be retained for at least one year after the expiration date of the batch); see also 21 C.F.R. §§ 211.186, 211.188, 211.194 (requiring “master production and control records,” “batch production and control records,” and “laboratory records”).
The USP entry for enoxaparin, the drug at issue in this litigation, states: “About 20 percent of the materials contain a 1, 6-anhydro derivative on the reducing end of the chain, the range being between 15 and 25 percent.” J.A. 365 (USP Revision Bulletin, Official December 1, 2008). Thus, in order to be “enoxaparin” as defined in the [1361]*1361USP entry, the marketed drag product must contain between 15 and 25 percent of the 1, 6-anhydro derivative. Id,.; see also 21 U.S.C. § 351(b) (drug adulterated if purports to be a drug in an official compendium but its strength, quality, or purity differs from the standard set forth in the compendium). The USP also includes a specific test for the 1, 6-anhydro derivative, which “involves HPLC analysis of a depolymerized enoxaparin sodium solution by a mixture of heparinases.” J.A. 369 (USP Method <207>). As the district court explained: “Claims 6, 16, and 53 of the '886 patent describe how to analyze a sample of enoxaparin to ensure its conformity to the USP Monograph standard.” J.A. 8. Amphastar is required by the FDA to use this test in order to ensure its enoxaparin is not adulterated. 21 U.S.C. § 351(b). This testing, which generates information for submission pursuant to the Food, Drug, and Cosmetic Act, therefore falls squarely within the scope of the safe harbor.
Finally, the dissent suggests that we must reject any disequilibrium between sections 201 and 202 of the Hatch-Wax-man Act, that is, the safe harbor should not be available unless a patent term extension is also available. Dissenting Op. at 1370-71. This is not correct. The Supreme Court in Eli Lilly noted that equilibrium was not always achieved. See Eli Lilly, 496 U.S. at 671-72, 110 S.Ct. 2683. We too have rejected this strict interpretation of the safe harbor, explaining that “statutory symmetry is preferable but not required.” Abtox, 122 F.3d at 1029 (holding that Class II medical devices, which are not subject to a “rigorous premarket approval process” and thus cannot receive patent term extensions, are nonetheless covered by the safe harbor).
III.
Under the correct interpretation of 35 U.S.C. § 271(e)(1), Momenta’s admission that Amphastar’s testing is carried out to “satisfy the FDA’s requirements,” Appellee’s Brief at 40-41, makes it unlikely that Momenta will succeed on the merits of its infringement claim. The district court’s findings with respect to the irreparable harm, balance of the hardships, and public interest factors were all, to some extent, predicated on its erroneous conclusion that Momenta’s' patent was likely infringed by Amphastar’s product. See J.A. 24 (applying a presumption of irreparable harm in view of Momenta’s “showing of infringement and validity”); J.A. 29 (explaining that in light of the “showing of likelihood of success on the merits, the balance of hardship tips in [Momenta’s] favor”); J.A. 30 (public interest favors protection of patent rights secured by valid patents). Because Momenta has not established a likelihood of success on its claim of infringement, the preliminary injunction must be vacated.
On remand, the district court may want to consider whether Momenta’s admission that Amphastar’s use of the patented invention is to “satisfy the FDA’s requirements” makes this case amenable to summary judgment of non-infringement in favor of Amphastar. Because the safe harbor issue is dispositive, we need not reach the other arguments on appeal.
VACATED AND REMANDED.
Costs
Costs to Appellants.