dissenting.
Because (1) the North Dakota Supreme Court has approved tort recovery for damage to “other property” and directed its state courts away from the foreseeability approach we erroneously predicted twenty years ago in Dakota Gasification Co. v. Pascoe Building Systems, 91 F.3d 1094 (8th Cir.1996), and (2) even if Dakota Ga*298sification states the relevant test, it would not control the facts of this case, I dissent.3
I. North Dakota Law
“The economic loss doctrine recognizes the distinction between the bargain expectation interests protected by contract law under the Uniform Commercial Code and the safety interests protected by tort law.” Clarys v. Ford Motor Co., 592 N.W.2d 573, 578 (N.D.1999). “‘[I]t prevents tort law from altering the allocation of costs and risks negotiated by the parties’ ” by “ ‘den[ying] a remedy in tort to a party whose complaint is rooted in disappointed contractual or commercial expectations.’ ” Dannix Painting, LLC v. Sherwin-Williams Co., 732 F.3d 902, 906 (8th Cir. 2013) (quoting Marvin Lumber & Cedar Co. v. PPG Indus., Inc., 223 F.3d 873, 882 (8th Cir.2000), and Mut. Serv. Cas. Ins. Co. v. Elizabeth State Bank, 265 F.3d 601, 615 (7th Cir.2001)).
In 1992, the North Dakota Supreme Court held, “[A] manufacturer of a machine sold in a commercial transaction may not be held liable in negligence or strict liability for economic loss caused by a failure of a component part of the machine which causes damage to the machine only.” Coop. Power Ass’n v. Westinghouse Elec. Corp., 493 N.W.2d 661, 667 (N.D.1992). In doing so, the North Dakota Supreme Court adopted the economic loss doctrine as announced in East River Steamship Corp. v. Transamerica Delaval, Inc., 476 U.S. 858, 106 S.Ct. 2295, 90 L.Ed.2d 865 (1986), agreeing with East River that the economic loss doctrine is necessary to preserve the distinction between tort liability — which flows from the responsibility and capability of a manufacturer to ensure the safety of its products — and contract liability — which flows from “losses [that] essentially relate to the benefit of the bargain between business entities.” Coop. Power, 493 N.W.2d at 664-65, 667. The court in Cooperative Power recognized that because “the rules of negligence and strict liability for damage to property other than the product itself and for personal injury adequately protect [society’s] interest [in safe products] and provide manufacturers with incentive to produce safe products,” economic loss to the product itself is recoverable only under contract and warranty principles. Id. at 665.
Four years later, we decided Dakota Gasification. “Although there [wa]s no North Dakota case directly on point” deciding whether the economic loss doctrine applied to property other than the product itself, we thought North Dakota would follow “the modern trend in many jurisdictions ... that tort remedies are unavailable for property damage experienced by the owner where the.damage was a foreseeable result of a defect at the time the parties contractually determined their respective exposure to risk, regardless whether the damage was to the ‘goods’ themselves or to ‘other property.’ ” Dakota Gasification, 91 F.3d at 1099. We thus extended North Dakota’s economic loss doctrine and predicted the state’s supreme court would apply the doctrine to all property damage that could foreseeably result from a defect rather than rely upon a distinction between the purchased product and other property. See id.
The majority here perpetuates this flawed prediction because the North Dakota Supreme Court has not directly contradicted Dakota Gasification. “While we *299are loath to reject the considered judgment of a prior panel decision of our court, our task is to apply state law.” Marvin Lumber, 223 F.3d at 883. Because it is not federal law we apply, we must utilize all “ ‘relevant state precedent,' analogous decisions, considered dicta, and any other reliable data’ ” to anticipate how the state’s highest court would decide this case. Ashley Cnty., Ark. v. Pfizer, Inc., 552 F.3d 659, 665 (8th Cir.2009) (quoting In re W. Iowa Limestone, Inc., 538 F.3d 858, 866 (8th Cir.2008)). We cannot blithely look past the North Dakota Supreme Court’s recent statements by adhering to a prediction based on a decades-old legal landscape. See, e.g., Holden Farms, Inc. v. Hog Slat, Inc., 347 F.3d 1055, 1066-67 (8th Cir.2003) (giving a non-binding intermediate state court decision controlling weight over our earlier panel decision given the absence of “data suggesting the [state] Supreme Court would disagree with the” intermediate state court’s decision).
A. “Modern Trend” of Foreseeability
The unstable foundation upon which we built our position in Dakota Gasification has eroded over time. In Dakota Gasification, we relied primarily on a “modern trend in many jurisdictions” which we found consistent with the general rationale underlying Cooperative Power’s adoption of the economic loss doctrine. See Dakota Gasification, 91 F.3d at 1099. Shortly after our decision, that trend lost its momentum. In 1997, the United States Supreme Court explained that East River-the primary authority for the North Dakota Supreme Court in Cooperative Power— “held that an admiralty tort plaintiff cannot recover for the physical damage the defective product causes to the ‘product itself; but the plaintiff can recover for physical damage the product causes to ‘other property.’ ” Saratoga Fishing Co. v. J.M. Martinac & Co., 520 U.S. 875, 877, 117 S.Ct. 1783, 138 L.Ed.2d 76 (1997) (emphasis added). The Supreme Court rejected the idea that “tort recovery for damage to ... other property” could be precluded by “the mere possibility” that a product’s manufacturer and initial purchaser could have negotiated a contract term apportioning potential loss for other property. Id. at 882. Soon after, the American Law Institute published the Restatement (Third) of Torts (Restatement), which employs an economic loss rule permitting recovery in tort for “harm to ... the plaintiffs property other than the defective product itself.” Restatement (Third) of Torts: Product Liability § 21(c) (1998); see also id. cmt. e. Dakota Gasification’s “modern trend” stalled.
The majority acknowledges the Restatement and Saratoga Fishing but dismisses “Saratoga Fishing [a]s inapposite in the instant analysis primarily because it does not construe North Dakota law ... but rather was an admiralty case and thus does not address the heart of the matter we now face.” Ante at 295. This distinction ignores (1) the Restatement (which obviously applies outside admiralty), and (2) the fact that North Dakota derived its economic loss doctrine from East River, an admiralty law case. See Coop. Power, 493 N.W.2d at 664-67. The majority accepts Saratoga Fishing as “[a]pplying and building upon ... East River,” ante at 294, the very case North Dakota expressly followed. See Coop. Power, 493 N.W.2d at 667.
The foreseeability approach’s earlier momentum, upon which we relied in Dakota Gasification, has since dissipated against the contrary views of the economic loss doctrine. See, e.g., 2-J Corp. v. Tice, 126 F.3d 539, 544 n. 4 (3d Cir.1997) (noting the disagreement as to whether “the economic loss doctrine bars tort recovery where the ‘other property1 damaged was always like*300ly to have been injured upon the failure of ‘the product’ itself’ and finding “more persuasive” the cases rejecting “this expansion of the economic loss doctrine”); Lesiak v. Cent. Valley Ag Coop., Inc., 283 Neb. 103, 808 N.W.2d 67, 84-85 (2012) (per curiam) (rejecting the foreseeability approach because it “precludes tort recovery based on the mere possibility that the parties could have included a contract term dealing with the occurrence of the damage at issue”); Grams v. Milk Prods., Inc., 283 Wis.2d 511, 699 N.W.2d 167, 180 (2005) (Abrahamson, C.J., dissenting) (reasoning that with a foreseeability threshold, Wisconsin’s economic loss doctrine has become “a swelling globule on the [state’s] legal landscape” akin to “the ever-expanding, all-consuming alien life form portrayed in the 1958 B-movie classic The Blob”).
B. Recent Statements by the North Dakota Supreme Court
Since Dakota Gasification, the North Dakota Supreme Court unquestionably has shown it adheres to the dichotomy between the product itself and other property. In 1999, the North Dakota Supreme Court “consider[ed] for the first time whether the economic loss doctrine should be applied to consumers” in addition to commercial purchasers. Clarys, 592 N.W.2d at 575. The Clarys court found the economic loss doctrine applied to consumer transactions, reasoning, “When a defective product causes damage to itself, only the consumer’s product expectations are infringed, and those interests are protected under contract principles found within the remedies provided by the U.C.C.” Id. at 576 (emphasis added). Throughout its discussion, the Clarys court reaffirmed its agreement with East River and reiterated that contract liability is for damage to “the product itself’ while indicating tort liability is available for damage to “other property.” Id. at 574-75, 577, 578-79. The court explained:
When a defective product causes damage to persons or other property, the interest at stake is health and safety.... Those safety interests are protected wider tort law, which allows recovery by injured plaintiffs against a seller or manufacturer of an unreasonably dangerous defective product. When, however, a product is defective and damages only itself, the interest at stake is the purchaser’s expectation of receiving the bargained-for product. That interest is protected by the remedies provided under Article 2 of the Uniform Commercial Code.
Id. at 578-79 (emphasis added) (citation omitted); see also id. at 577-78 (interpreting North Dakota’s Tort Reform Act of 1993 to permit “tort actions ... when a defective product causes damage to persons or other property,” but not “when only the defective product has been damaged”).
The Clarys court expressly agreed with the drafters of the Restatement “that the Uniform Commercial Code, not product liability tort law, governs actions of persons seeking redress for damages when the injury is confined to the defective product itself, an$ neither persons nor other property are damaged.” Id. at 578. In announcing its agreement with the Restatement, the court emphasized that its approach permits recovery in a product liability suit for harm to “ ‘property other than the defective product itself’ ” but not for harm “to the defective product itself.” Id. (emphasis in original) (quoting Restatement (Third) of Torts: Product Liability § 21(c)).
Since Clarys, the North Dakota Supreme Court has only twice addressed the economic loss doctrine, and in both instances, it reiterated the distinction be*301tween the product itself and other damaged property. See Leno v. K & L Homes, Inc., 803 N.W.2d 543, 550 (N.D. 2011) (“[U]nder the economic loss doctrine in North Dakota, ‘economic loss resulting from damage to a defective product, as distinguished from damage to other property or persons, may be recovered in a cause of action for breach of warranty or contract, but not in a tort action.’ ” (quoting Steiner v. Ford Motor Co., 606 N.W.2d 881, 884 (N.D.2000))); Steiner, 606 N.W.2d at 884 (“ ‘[D]amage to persons or other property ... [is] protected under tort law.’” (quoting Clarys, 592 N.W.2d at 578)).
The majority dismisses these indications of North Dakota law, disregarding the discussion in Clarys as “dicta.” Ante at 295-96. However, the signals in Clarys are more than just fleeting comments or unassuming word choices. The North Dakota Supreme Court’s statements about “other property” represent an important part of the careful weighing and molding that went into the court’s understanding and description of how best to balance contract law with tort. “[C]onsidered dicta” such as this provides crucial insight into the North Dakota Supreme Court’s thinking, Ashley Cnty., 552 F.3d at 665, particularly when that same court cites and repeats the point in later cases, see, e.g., Leno, 803 N.W.2d at 550; Steiner, 606 N.W.2d at 884.
The majority notes the North Dakota Supreme Court’s later cases do not expressly “point[ ] to a rejection of the foreseeability approach” advanced in Dakota Gasification. Ante at 295. While the majority may fit Dakota Gasification’s foreseeability approach within portions of Clary s’s general discussion of the role and underpinnings of the economic loss doctrine, the majority does not justify Dakota Gasification’s outright disregard for the product/other property analysis in light of the North Dakota Supreme Court’s recent decisions favoring such a distinction. Compare Dakota Gasification, 91 F.3d at 1099, 1101 (applying its foreseeability rale to preclude tort recovery “regardless whether the damage was to the ‘goods’ themselves or to ‘other property’ ” (emphasis added)), with Clarys, 592 N.W.2d at 578 (explaining tort law protects “damage to persons or other property”).
I share the majority’s respect for our prior panel’s opinion. But time proved our prediction wrong, and the prediction should now give way to our obligation to apply North Dakota law in the way described by the highest court of that state. I believe North Dakota would not ask whether the harm was foreseeable but whether the harm occurred to “other property.”
II. Foreseeability Under Dakota Gasifícation
Even accepting Dakota Gasification’s test as controlling, that case does not preclude Arena Holdings’s tort recovery under the summary judgment standard. See Fed.R.Civ.P. 56(a) (requiring summary judgment only where “the movant shows that there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law”).
As “ ‘the fundamental boundary between contract law ... and tort law,’ ” Clarys, 592 N.W.2d at 576 (quoting Casa Clara Condo. Ass’n v. Charley Toppino & Sons, Inc.; 620 So.2d 1244, 1246 (Fla.1993)), the economic loss doctrine retains the delicate task of balancing the competing policy interests of both areas of the law — not just resisting “the usurpation of contract law by tort law,” ante at 297 (citing Clarys, 592 N.W.2d at 577). The majority’s foreseeability analysis does not account for the two primary circumstances that led the *302panel in Dakota Gasification to find the harm in that case foreseeable. Both circumstances are noticeably absent here.
First, in Dakota Gasification, the foreseeability inquiry was whether damage to a plant and its contents was a foreseeable result of defects in “structural components such as steel rafters, columns, and purlins,” which were used in building the plant. Dakota Gasification, 91 F.3d at 1096-97, 1100. Such harm is nearly per se foreseeable when the potential defect lies in the plant’s structural components because the components’ ability to hold up the building and protect its contents is the primary motivation for purchasing such materials. Our discussion of foreseeability in Dakota Gasification made special note of this fact on more than one occasion. See id. at 1100 (reasoning, in discussing our adoption of the foreseeability approach, that “it is difficult to imagine a scenario in which the natural consequence of an installed structural component’s failure would be damage only to the structural component itself without any damage to the surrounding property”); id. (assessing foreseeability and analogizing the case to a Sixth Circuit decision because in both cases, “the defective structural components caused great damage to surrounding property”). In this case, it is not clear whether fire safety was a reason for purchasing Harman’s amplifiers, and unlike the dangers of collapsing structural components, the bargained-for function of an amplifier normally would not entail an inherent danger of widespread fire damage and put these risks “within the scope of bargaining” as a matter of law. Grams, 699 N.W.2d at 175.
Second, in Dakota Gasification we looked to one other fact in assessing foreseeability: A limitation of liability provision in the purchase contract for the steel components stating, “SELLER SHALL NOT BE RESPONSIBLE FOR ... DAMAGES TO THE CONTENTS OR FURNISHINGS IN ANY BUILDING.” Dakota Gasification, 91 F.3d at 1100. That contract specifically allocated the risk of defects in the product harming the building’s furnishings and contents. See id. (reasoning that “[t]his language demonstrates that the damage which occurred here was well within the contemplation of the parties” (emphasis added)).
But Harman relies on language declaring,
YOU ARE NOT ENTITLED TO RECOVER FROM U.S. ... ANY DAMAGE TO ANOTHER PRODUCT OR PRODUCTS RESULTING FROM ... A DEFECT [IN THE PURCHASED PRODUCT],
This disclaimer acknowledges that an amplifier defect might damage “another product,” but this statement is far too vague to give any sense that the fire and smoke damage here was one of the allocated risks considered during the bargaining.
Foreseeability of harm generally “is a question of fact for the jury, unless the facts are such that reasonable minds could not differ.” Barsness v. Gen. Diesel & Equip. Co., 383 N.W.2d 840, 843 (N.D. 1986); accord Haugénoe v. Workforce Safety & Ins., 748 N.W.2d 378, 387 (N.D. 2008); see also Foremost Farms USA Coop. v. Performance Process, Inc., 297 Wis.2d 724, 726 N.W.2d 289, 297 (Ct.App. 2006) (considering foreseeability, as used in the economic loss doctrine, a “factual question”). In Dakota Gasification, the structural nature of the defect made the resulting damage a practical inevitability, a fact reflected by a limitation of liability clause specifically addressing damage to the building’s furnishings and contents.
Before granting summary judgment to Harman, I would require a similarly undeniable indication of foreseeability. Har*303man’s evidence here — (1) the user manual shipped with the amplifier, which referenced at one point the possibility of speakers overheating and warned at another point against the increased risk of fire should the amplifier be exposed to rain or moisture; (2) knowledge by the initial purchasing subcontractor that amplifiers could (but rarely did) cause fires; and (3) Arena Holdings’s implementation of technical fire-safety specifications for contractors working on the arena project — is not so clear. On this evidence, the fire-risk of an amplifier defect might be foreseeable enough to bring the risk within the purview of the bargaining process, but that question is not beyond reasonable dispute and thus should be left to a jury.
III. Conclusion
Because I disagree that Dakota Gasification governs, I would return this case to the district court to address the economic loss doctrine under a dichotomy between the product itself and other property. Even applying Dakota Gasification, I would still remand because there are material factual disputes precluding summary judgment.
. Under the principles of Erie R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938), it is axiomatic that "[a]s a federal court, our role in diversity cases is to interpret state law, not to fashion it” or “expand [it] in ways not foreshadowed by state precedent.” Kingman v. Dillard’s, Inc., 643 F.3d 607, 615 (8th Cir.2011) (quotations omitted).