New York’s usury laws prohibit unlicensed lenders from lending money at an interest rate above 16 percent per year, and criminalize loans with interest rates higher than 25 percent per year. N.Y. Gen. Oblig. Law § 5-501(1), N.Y. Banking Law § 14-a(l), N.Y. Penal Law §§ 190.40-42. The plaintiffs are two Native American tribes, tribal regulatory agencies, and companies owned by the tribes that provide short-term loans over the internet, all of which have triple-digit interest rates that far exceed the ceiling set by New York law. When the New York State Department of Financial Services (“DFS”) tried to bar out-of-state lenders, including the plaintiffs, from extending loans to New York residents, plaintiffs sought a preliminary order enjoining DFS from interfering with the tribes’ consumer lending business.
Plaintiffs contended that New York had projected its regulations over the internet and onto reservations in violation of Native Americans’ tribal sovereignty, which is protected by the Indian Commerce Clause of the Constitution. U.S. CONST. art. 1, § 8, cl. 3. But the United States District Court for the Southern District of New York (Richard J. Sullivan, Judge) held that plaintiffs had not offered sufficient proof that the loans fell outside New York’s regulatory domain. After examining the evidence marshaled by plaintiffs in support of their motion, the District Court concluded that plaintiffs had failed to establish that the challenged loan transactions occurred on Native American soil, a fact necessary to weaken New York State’s regulatory authority over them. Because this conclusion was a reasonable one, we AFFIRM the District Court’s denial of plaintiffs’ motion for a preliminary injunction.
BACKGROUND
This case arises from a conflict between two sovereigns’ attempts to combat poverty within their borders. Native American tribes have long suffered from a dearth of economic opportunities. Plaintiffs in this case, the Otoe-Missouria Tribe of Indians, the Lac Vieux Desert Band of Lake Supe*108rior Chippewa Indians, and wholly owned corporations of those tribes (collectively, “the lenders”), established internet-based lending companies in the hopes of reaching consumers who had difficulty obtaining credit at favorable rates but who would never venture to a remote reservation. The loans were made at high interest rates, and the loans permitted the lenders to make automatic deductions from the borrowers’ bank accounts to recover interest and principle. New York has long outlawed usurious loans. DFS aggressively enforced those laws in order to “protect desperately poor people from the consequences of their own desperation.” Schneider v. Phelps, 41 N.Y.2d 238, 243, 391 N.Y.S.2d 568, 359 N.E.2d 1361 (1977). Thus, the tribes’ and New York’s interests collided.
It is unclear, however, where they collided-in New York or on a Native American reservation. The lenders assert that the challenged transactions occurred on reservations. The “loan application process” took place via “website[s] owned and controlled by the Tribe[s].” Loans were “reviewed and assessed by ... Tribal loan underwriting system[s].” Loans complied with rules developed, adopted, and administered by tribal regulatory authorities. The loans were funded out of “Tribally owned bank accounts.” And each loan application notified borrowers that the contract was “governed only by the laws of [the Tribe] and such federal law as is applicable under the Indian Commerce Clause of the United States Constitution ... [and] [a]s such, neither we nor this Agreement are subject to any other federal or state law or regulation.” In sum, as the Chairman of the Lac Vieux Desert Tribe explained in an affidavit, “[t]hrough technological aids and underwriting software, loans are approved through processes that occur on the Reservation in various forms.”1
But loans approved on Native American reservations and other out-of-state locations flowed across borders to consumers in New York. New York borrowers never traveled to tribal lands or other jurisdictions; they signed loan contracts remotely by keying in an electronic signature. Borrowers listed their New York addresses on applications, and provided lenders with routing information for their personal bank accounts in New York. Moreover, the lenders did more than simply transfer loan proceeds into New York bank accounts. Under the terms of the loans, the lenders reached into New York to collect payments: the lenders placed a hold on borrowers’ accounts that resulted in an automatic debit every two weeks over the course of many months.2 The harm inflicted by these high-interest loans fell upon customers in New York: DFS received complaints from residents faltering under the weight of interest rates as high as 912.49 percent; as one complaint explained, “I am attempting to get out of a hole, not dig a deeper one.”
Thus, both the tribes .and New York believed that the high-interest loans fell within their domain, both geographic and regulatory, and acted accordingly. The tribes re-invested profits into their eom-*109munities, and New York authorities began an investigation into online payday lending. In the summer of 2013, those initiatives clashed.
In August, DFS launched what the tribal lenders describe as a “market-based campaign explicitly designed to destroy Tribal enterprises,” and what New York defends as a “comprehensive effort to determine how best to protect New Yorkers from the harmful effects of usurious online payday loans.” At issue are two related mailings.
First, DFS sent cease-and-desist letters to thirty-five online payday lenders that it had identified as having made loans to New York residents. Its efforts were directed generally at such lenders, including not only tribal lenders, but also foreign lenders and lenders headquartered in states that do not cap interest rates on short-term loans. The letters accused lenders of “using the Internet to offer and originate illegal payday loans to New York consumers,” in violation of “New York’s civil and criminal usury laws.” The letters instructed lenders to “confirm in writing” within fourteen days “that [they were] no longer soliciting] or mak[ing] usurious • loans in New York.”
Second, DFS wrote to the lenders’ partners in the financial services industry. The lenders relied on outside hanks to hold money and transfer it to customers. Those banks, in turn, depended upon an electronic wire service called the Automated Clearing House (“ACH”) to move money from their coffers into borrowers’ accounts, and to extract repayment from those accounts. DFS’s letters solicited banks and ACH for their “cooperative effort[s]” to “stamp out these pernicious, illegal payday loans.” In the letters sent to banks, DFS warned that “it [was] in ... [the] bank’s long-term interest to take appropriate action to help ensure that it is not serving as a pipeline for illegal conduct.” It, urged the banks to “work with” the agency “to create a new set of model safeguards and procedures to choke-off ACH access” to the 35 payday lenders that had lent money to New York customers. “Doing so,” the letter counseled, was “in the best interest of your member banks and their customers.” The letters ended with a request that the companies meet with New York officials to discuss a cooperative “undertaking.”
According to plaintiffs, DFS’s outreach had immediate and , devastating effects on tribal lenders. Banks and ACH abruptly ended their relationships with the lenders, stymieing their transactions not just with New York borrowers, but with consumers in every other state in the union. Without revenue from lending, the tribes faced large gaps in their budgets. According to the Chairman of the Otoe-Missouria tribe, proceeds from lending account for almost half of the tribe’s non-federal income. Profits from lending have fueled expansion of tribal early childhood education programs, employment training, healthcare coverage, and child and family protection services. The Chairman of the Lac Vieux Desert tribe attested to similar fiscal reliance, noting that lending revenue supports tribal housing initiatives, youth programs, health and wellness services, and law enforcement.
Faced with crumbling businesses and collapsing budgets, plaintiffs filed suit, claiming that New York’s efforts to curb the lenders’ online business violated the Indian Commerce Clause of the Federal Constitution by infringing on tribes’ fundamental right to self-government. Plaintiffs moved for a preliminary injunction barring DFS from further interfering with the lenders’ transactions with consumers in New York and elsewhere. The District Court denied the motion. The court found *110that the lenders had “built a wobbly foundation for their contention that the State is regulating activity that occurs on the Tribes’ lands,” and concluded that New York’s “action [was] directed at activity that [took] place entirely off tribal land, involving New York residents who never leave New York State.” Otoe-Missouria Tribe of Indians v. N.Y. State Dep’t of Fin. Servs., 974 F.Supp.2d 353, 360 (S.D.N.Y.2013). Thus, the court held that New York acted within its rights to regulate business activity within the state.
This appeal followed.
DISCUSSION
I. Preliminary Injunctions: Standard for Granting, Standard of Review
A district court’s denial of a motion for a preliminary injunction is reviewed for abuse of discretion. WPIX, Inc. v. ivi, Inc., 691 F.3d 275, 278 (2d Cir.2012). In general, district courts may grant a preliminary injunction where a plaintiff demonstrates “irreparable harm” and meets one of two related standards: “either (a) a likelihood of success on the merits, or (b) sufficiently serious questions going to the merits of its claims to make them fair ground for litigation, plus a balance of the hardships tipping decidedly in favor of the moving party.” Lynch v. City of N.Y., 589 F.3d 94, 98 (2d Cir.2009) (internal quotation marks omitted). This two-track rule, however, is subject to an exception: A plaintiff cannot rely on the “fair-ground-for-litigation” alternative to challenge “governmental action taken in the public interest pursuant to a statutory or regulatory scheme.” Plaza Health Labs., Inc. v. Perales, 878 F.2d 577, 580 (2d Cir.1989) (relying on Union Carbide Agric. Prods. Co. v. Costle, 632 F.2d 1014, 1018 (2d Cir.1980) and Med. Soc’y of N.Y. v. Toia, 560 F.2d 535, 538 (2d Cir.1977)). As we have explained, “[t]his exception reflects the idea that governmental policies implemented through legislation or regulations developed through presumptively reasoned democratic processes are entitled to a higher degree of deference and should not be enjoined lightly.” Able v. United States, 44 F.3d 128, 131 (2d Cir.1995).
DFS’s attempt to curb online payday lending in New York was a paradigmatic example of “governmental action taken in the public interest,” Plaza Health Labs., 878 F.2d at 580, one that vindicated proven “policies implemented through legislation or regulations.” Able, 44 F.3d at 131. New York’s usury prohibitions date back to the late 18th century. New York enacted the current cap — 16 percent interest on short-term loans made by non-bank, unlicensed lenders — decades ago. See N.Y. Banking Law § 14-a (McKinney 2014) (noting original enactment date of Dec. 31, 1979). New York courts have consistently upheld and enforced such laws; as the New York Court of Appeals wrote in 1977, usury laws protect “impoverished debtors from improvident transactions drawn by lenders and brought on by dire personal financial stress.” Schneider, 41 N.Y.2d at 243, 391 N.Y.S.2d 568, 359 N.E.2d 1361. New York regulatory authorities, both at the behest of successive Attorneys General and now the Superintendent of Financial Services,3 have pursued businesses that *111lent money at interest rates above the legal limit. See e.g., Press Release, New York State Office of the Attorney General, Spitzer Not Preempted in Suit to Stop Illegal Payday Lending Scheme (May 28, 2004), available at http://vmm.ag.ny.gov/ pressrelease/spitzer-notpreempted-suit-stop-illegal-payday-lending-scheme (describing lawsuit brought by former Attorney General Eliot Spitzer). Although plaintiffs argue that New York lacks the authority to enforce its laws against tribal lenders (and they may be right in the end), there is no question as to what those laws require.
For this reason, plaintiffs must establish a likelihood of success on the merits to win injunctive relief at this early stage. Our decision in Haitian Centers Council, Inc. v. McNary, 969 F.2d 1326 (2d Cir.1992), is not to the contrary. There, we upheld an order enjoining the Immigration and Nationalization Service (“INS”) from limiting Haitian asylum applicants’ contact with counsel while they were detained at Guantanamo Bay. Id. at 1347. We did so even though the plaintiffs demonstrated only a fair ground for litigation rather than a likelihood of success on the merits. Id. at 1339. The government could not identify any specific statute or regulation that allowed it to deny counsel to applicants at their screening interviews — a top official had announced the policy in a memo in response to a flood of applicants following a coup. The agency sought to moor its policy choice in the “broad grant of authority in the [Immigration and Nationality Act]” to screen emigrants. Id. We deemed that too general an authority to trigger the higher standard for a preliminary injunction. Id. “We believe that in litigation such as is presented herein,” we explained, “no party has an exclusive claim on the public interest.” Id. The “likelihood of success” prong, we held, “need not always be followed merely because a mov-ant seeks to enjoin government action.” Id.
This ease is distinguishable from Haitian Centers Council in two respects. First, DFS acted to enforce a rule embodied in a specific statute. In contrast, the INS enforced a much more informal policy, hastily adopted without the benefit of either specific statutory instructions or regulations issued after a public notice- and-comment process. Second, New York’s view of the “public interest” has been defined and reaffirmed by all three branches of government for many years. Unlike the novel issue presented by Haitian detainees seeking counsel while they awaited transfer to the continental United States, New York long ago confronted and answered the policy question posed in this case — whether businesses should be allowed to make triple-digit, short-term loans to those with an acute liquidity problem but no credit with which to solve it. Thus, “the full play of the democratic process involving both the legislative and executive branches has produced a policy in the name of the public interest embodied in a statute and implementing regulations.” Able, 44 F.3d at 131. That policy is entitled to “a higher degree of deference” than a private party’s position would merit, and we must be sure that, in all likelihood, New York has acted unlawfully before we substitute our judgment for that of the political branches. Id.
We recognize that the plaintiffs’ argument that there are “public interests on both sides” in this case, is not without force. The tribes are independent nations, and New York’s regulatory efforts may hinder the tribes’ ability to provide for *112their members and manage their own internal affairs. But as we explained in Oneida Nation of N.Y. v. Cuomo, 645 F.3d 154 (2d Cir.2011), “[a] party seeking to enjoin governmental action taken in the public interest pursuant to a statutory or regulatory scheme cannot rely on the fair ground for litigation alternative even if that party seeks to vindicate a sovereign or public interest.” Id. at 164 (holding that Oneida Nation must prove a likelihood of success on the merits to merit a preliminary injunction enjoining New York from enforcing tax scheme on the tribe’s cigarette sales). Despite the possibly serious intrusion on tribal interests posed by this case, the plaintiffs must still meet the higher standard.4
II. Likelihood of Success on the Merits
Plaintiffs claim that DFS infringed upon tribal sovereignty in two ways. They argue that New York had no authority to order tribes to stop issuing loans originated on Native American reservations, and that New York regulated activity far outside its borders when it launched a “market-based campaign” to shut down tribal lending in every state in the Union. But to prove either of these claims, plaintiffs had to demonstrate that the challenged transactions occurred somewhere other than New York, and, if they occurred on reservations, that the tribes had a substantial interest in the lending businesses. As described below, the district court reasonably concluded that plaintiffs failed to do so.
A. The “Who,” “Where,” and “What” of the Indian Commerce Clause
Indian Commerce Clause jurisprudence balances two conflicting principles. On the one hand, Native Americans retain the right to “make their own laws and be ruled by them.” Williams v. Lee, 358 U.S. 217, 220, 79 S.Ct. 269, 3 L.Ed.2d 251 (1959). On the other, tribes are only “semi-independent”; their sovereign authority is “an anomalous one and of a complex character,” McClanahan v. State Tax Comm’n of Az., 411 U.S. 164, 173, 93 S.Ct. 1257, 36 L.Ed.2d 129 (1973), because tribes remain “ultimately dependent on and subject to the broad power of Congress,” White Mountain Apache Tribe v. Bracker, 448 U.S. 136, 143, 100 S.Ct. 2578, 65 L.Ed.2d 665 (1980). With these two principles in mind, the Supreme Court has held that states may regulate tribal activities, but only in a limited manner, one constrained by tribes’ fundamental right to self-government, and Congress’s robust power to manage tribal affairs.5 Id. at 142-43, 100 *113S.Ct. 2578. That delicate balance results in an idiosyncratic doctrinal regime, one that, as the Ninth Circuit has described, requires “careful attention to the factual setting” of state regulation of tribal activity. Barona Band of Mission Indians v. Yee, 528 F.3d 1184, 1190 (9th Cir.2008).
The breadth of a state’s regulatory power depends upon two criteria — the location of the targeted conduct and the citizenship of the participants in that activity. Native Americans “going beyond the reservation boundaries” must comply with state laws as long as those laws are “nondiscriminatory [and] ... otherwise applicable to all citizens of [that] State.” Mescalero Apache Tribe v. Jones, 411 U.S. 145, 148-49, 93 S.Ct. 1267, 36 L.Ed.2d 114 (1973) (“Mescalero /”). For example, in Mescalero I, the Supreme Court held that New Mexico could collect sales and use taxes from a ski resort owned by a Native American tribe that was located outside a reservation’s borders. Id. at 149, 93 S.Ct. 1267. Every business in the state had to pay the tax, and the Indian Commerce Clause did not create an exception to that rule.
But once a state reaches across a reservation’s borders its power diminishes and courts must weigh the interests of each sovereign — the tribes, the federal government, and the state — in the conduct targeted by the state’s regulation. The scales will tip according to the citizenship of the participants in the conduct. As the Supreme Court explained in Bracker, “[w]hen on-reservation conduct involving only Indians is at issue, state law is generally inapplicable, for the State’s regulatory interest is likely to be minimal and the federal interest in encouraging tribal self-government is at its strongest.” 448 U.S. at 144, 100 S.Ct. 2578. A state’s interest waxes, however, if “the conduct of non-Indians” is in question. Id. A court conducts a more “particularized inquiry into the nature of the state, federal, and tribal interests at stake.” Id. at 144-45, 100 S.Ct. 2578. In Bracker, the Supreme Court engaged in that “particularized inquiry” and held that Arizona could not impose fuel and use taxes on a non-Indian hauler moving timber across a reservation. Although Arizona wished to raise revenue, the federal government and the tribe’s shared commitment to the continued growth and productivity of tribal logging enterprises outweighed Arizona’s interest.
Thus, “the “who’ and the ‘where’ of the challenged [regulation] have significant consequences,” ones that are often “dispos-itive.” Wagnon v. Prairie Band Potawatomi Nation, 546 U.S. 95, 101, 126 S.Ct. 676, 163 L.Ed.2d 429 (2005). And even when the “who” and “where” are clear, a court must still understand “what” a regulation targets to weigh interests appropriately. A tribe’s interest peaks when a regulation threatens a venture in which the tribe has invested significant resources. In New Mexico v. Mescalero Apache Tribe, 462 U.S. 324, 103 S.Ct. 2378, 76 L.Ed.2d 611 (1983) (‘Mescalero II”), the Supreme Court held that a state could not enforce its hunting laws against non-Indian sportsmen who hunted and fished on a reservation. Id. at 341, 103 S.Ct. 2378. The tribe had “engaged in a concerted and sustained undertaking to develop and manage the reservation’s wildlife and land resources,” and state regulations threatened to unsettle and supplant those investments. Id.
*114Four years later, the Court echoed that conclusion in California v. Cabazon Band of Mission Indians, 480 U.S. 202, 107 S.Ct. 1083, 94 L.Ed.2d 244 (1987). There, the Court permitted Native American tribes to continue operating on-reservation bingo games without complying with California’s gambling restrictions, even though the tribes catered their games to nonNative American customers. The tribes had “built modern[,] ... comfortable, clean, and attractive facilities,” and developed rules and procedures to ensure “well-run games.” Those sunk costs were a “substantial interest” that outweighed California’s interest in curbing organized crime’s “infiltration of the tribal games.” Id. at 219-21, 107 S.Ct. 1083.
In contrast, a tribe has no legitimate interest in selling an opportunity to evade state law. In Washington v. Confederated Tribes of the Colville Indian Reservation, 447 U.S. 134, 100 S.Ct. 2069, 65 L.Ed.2d 10 (1980), the Supreme Court held that tribal stores had to collect a state tax on cigarettes sold to non-Native American customers. Id. at 161, 100 S.Ct. 2069. All the “smokeshops offer[ed to non-member] customers, [that was] not available elsewhere, [was] solely an exemption from state taxation.” Id. at 155, 100 S.Ct. 2069. “[W]hether stated in terms of pre-emption, tribal self-government, or otherwise,” tribes did not have any legitimate interest in “marketing] an exemption from state taxation to persons who would normally do their business elsewhere.” Id.
Factual questions, then, pervade every step of the analysis required by the Indian Commerce Clause. A court must know who a regulation targets and where the targeted activity takes place. Only then can it either test for discriminatory laws, as in Mescalero I, or balance competing interests, as in Bracker. And even if a court knows enough to trigger a weighing of competing interests, a court must still know what the nature of those interests are. Only then can it assess whether a regulation threatens a significant investment, as in Mescalero II and Cabazon, or whether a tribe has merely masked a legal loophole in the cloak of tribal sovereignty, as in Colville. Given the fact-dependent nature of these inquiries, it is no surprise that, as detailed below, plaintiffs have failed to prove a likelihood of success on the merits at this early stage of the litigation.
B. The Ambiguity of Internet Loans and Cooperative Campaigns
Loans brokered over the internet seem to exist in two places at once. Lenders extend credit from reservations; borrowers apply for and receive loans without leaving New York State. Neither our court nor the Supreme Court has confronted a hybrid transaction like the loans at issue here, e-commerce that straddles borders and connects parties separated by hundreds of miles. We need not resolve that novel question today — the answer will depend on facts brought to light over the course of litigation. On the record now before us, plaintiffs have not offered sufficient proof of the “who,” “where,” and “what” of the challenged loans. Without knowing more facts, we cannot say that the District Court unreasonably concluded that New York regulated transactions brokered “entirely off tribal land,” or that District Court erred when, relying on that conclusion, it held that New York’s evenhanded treatment of payday lenders did not violate the Indian Commerce Clause. Otoe-Missouria Tribe of Indians, 974 F.Supp.2d at 360.
First, plaintiffs claim that New York had no authority to demand that the lenders “cease and desist” from extending loans to New York residents. At the outset, we *115note that even if these letters, which were sent to tribal lenders (among other payday lenders), constitute attempted regulation of on-reservation activities, plaintiffs do not allege that the letters caused them harm; the damage to their business derived not from the cease-and-desist letter, which plaintiffs appear to have ignored, but from actions discussed below that allegedly caused the tribal lenders’ non-tribal off-reservation banking partners to cease doing business with them.
In any event, plaintiffs provided insufficient evidence to establish that they are likely to succeed in showing that the internet loans should be treated as on-reservation activity. As the district court noted, plaintiffs “built a wobbly foundation for their contention that [New York] ... regulated] activity that occur[ed] on the Tribes’ lands.” Id. The lenders’ affidavits boldly (but conclusorily) assert that “loans are approved through processes that occur on ... Reservations],” but nowhere do they state what specific portion of a lending transaction took place at any facility physically located on a reservation. Plaintiffs averred that loans were processed through “website[s] owned and controlled by the Tribes,” but never identified the citizenship of the personnel who manage the websites, where they worked, or where the servers hosting the websites were located. Loans were approved by a “Tribal loan underwriting system,” a vague description that could refer to the efforts of Native American actuaries working on a reservation, but could also refer to myriad other “systems” — software developed and administered by an off-site company, paid consultants located anywhere in the world, or any number of other arrangements. Loans were funded out of “Tribally owned bank accounts,” but those accounts were apparently held in, and perhaps funded by, non-tribal banks; the necessary involvement of non-tribal financial institutions is the very basis of plaintiffs’ claim that their business collapsed when banks pulled out of the payday lending business after receiving New York’s letter. Thus, even if we agreed that New York customers traveled elsewhere when they opened an internet browser, the lenders failed to establish where those customers metaphorically went, and who exactly approved their loans.
The complexities introduced by modern electronic commercial transactions also weaken plaintiffs’ arguments. Much of the commercial activity at issue takes place in New York. That is where the borrower is located; the borrower seeks the loan without ever leaving the state, and certainly without traveling to the reservation. Even if we concluded that a loan is made where it is approved, the transaction New York seeks to regulate involves the collection as well as the extension of credit, and that collection clearly takes place in New York. The loan agreements permit the lenders to reach into the borrowers’ accounts, most or all of them presumably located in New York, to effect regular, automatic wire transfers from those accounts to make periodic payments on the loans.
A court might ultimately conclude that, despite these circumstances, the transaction being regulated by New York could be regarded as on-reservation, based on the extent to which one side of the transaction is firmly rooted on the reservation. Because significant aspects of the transaction and its attendant regulation are distinctly not located on-reservation, however, ambiguities in the record about those portions of the transaction that purportedly are loom all the larger.6
*116Given this decidedly ambiguous and insufficient record as to the details of the purportedly on-reservation portions of the loan transactions, plaintiffs insist that the courts’ traditional “on-or-off reservation” analysis is an “overly-simplistic” approach to the “modern world of e-commerce.” It is enough, plaintiffs argue, that tribes bear the “legal burden of the regulation,” and, with that in mind, they contend that the court should proceed directly to the interest balancing prescribed in Bracker.
As discussed above, Supreme Court precedent that we are not free to disregard directs us to make the initial inquiry into the location of the regulated activity. Even assuming that the electronic nature of the transaction at issue here would permit us to distinguish those cases and proceed to interest balancing, plaintiffs “have not provided sufficient evidence of what we would weigh were we to adopt that test. At first blush, the tribal lenders’ payday loans resemble the Colville tribes’ tax-free cigarettes: Tribes profit from leveraging an artificial comparative advantage, one which allows them to sell consumers a way to evade state law.7 In theory, the tribes may have built the electronic equivalent of “modern[,] ... comfortable, clean, attractive facilities” like the ones in Cabazon, and they may have “engaged in a concerted and sustained undertaking to develop and manage” limited capital resources as the tribe did in Mescalero II. But the record does not reflect any such “substantial interest.” Cabazon, 480 U.S. at 220, 107 S.Ct. 1083. As noted above, it is not entirely clear just what the lenders have virtually “built,” and in any event the record contains no information about the extent of investment that was required.8
Second, plaintiffs claim that DFS infringed upon tribal sovereignty by launch*117ing a “national campaign” with the “express purpose of destroying out-of-state tribal businesses.” That claim rests on equally tenuous ground: Read in their strongest form, DFS’s letters requested that ACH and banks stop processing payday loans made to New York customers, But, again assuming that New York’s letters requesting that banks and ACH cooperate with DFS constitute regulation, that effort was directed to those aspects of online lending that are remote from the reservation. The direct force of DFS’s request fell upon parties located far from a reservation, on financial institutions that plaintiffs themselves claim are indispensable outside partners.
For DFS’s “campaign” to have run afoul of the Indian Commerce Clause, the lenders must demonstrate that DFS treated financial intermediaries as a proxy for Native American tribes. To do so, plaintiffs would have to show that DFS acted with the intent of regulating tribes, or that its outreach had that effect. New York’s alleged efforts to influence the banks and ACH can hardly be considered discriminatory, or specifically aimed at tribal lenders, as the state asked that the banks and ACH stem loans made by any online lender. The letters targeted a diverse group of lenders, the majority of whom had no affiliation with Native American tribes. If DFS cast a broad net with the ulterior motive of ensnaring just the tribes, that intent was certainly well-hidden.
It is not clear, moreover, that the DFS letters required the banks and ACH to take any particular action. To be sure, the letters contained a few ominous turns of phrases; they requested that financial institutions “choke-off ACH access” and “stamp out ... pernicious,' illegal payday loans.” But the letters also concluded with soft requests, asking for a simple meeting to explore “cooperation.” It is impossible to know what this ambiguous tone, at once bombastic and conciliatory, implies about DFS’s intent to take regulatory action to coerce the banks and ACH to act.
Nor is it clear that New York’s actions would have had any different effects if the tribal lenders had not been explicitly identified by DFS. New York’s usury laws apply to all lenders, not just tribal lenders, and DFS’s letters to the banks and ACH made clear that New York regulators disapproved of the facilitation by banks of high-interest payday lending from outside the state. The Indian Commerce Clause has no bearing on New York’s efforts to discourage banks from cooperating with non-Indian payday lenders.9 Because it is not clear why the banks and ACH reacted as they did to DFS’s letters, it is uncertain that they would have continued to do business with tribal lenders if DFS had cited *118only the general problem of payday lending.
Thus, it is not clear what to infer, if anything, from the decisions made by ACH and other banks. Although it is possible that the companies believed that they had to comply with DFS’s agenda, it is equally possible that they simply made an independent calculation that the benefits of avoiding potential violations of New York law outweighed the benefits of doing business with payday lenders in general or with tribal lenders in particular. It is far from clear that the banks and ACH would have continued to do business with plaintiffs if DFS had simply requested that they drop their business relationships with payday lenders in general.
In sum, the record presented to the district provided ambiguous answers to what are fundamentally factual questions. With the benefit of discovery, plaintiffs may amass and present evidence that paints a clearer picture of the “who,” “where,” and “what” of online lending, and may ultimately prevail in this litigation. But at this stage, the record is still murky, and thus, the District Court reasonably held that plaintiffs had not proven that they would likely succeed on the merits.
CONCLUSION
For the foregoing reasons, we AFFIRM the District Court’s denial of plaintiffs’ motion for a preliminary injunction.