I concur with the majority in affirming the trial court’s judgment awarding compensatory damages. However, I dissent in regard to Plaintiffs’ cross-appeal on punitive damages. In my view, the record does not contain sufficient clear and convincing evidence from which a reasonable juror could conclude that, at the time of sale in 1996, Doncasters had actual knowledge of the specific defect on which the compensatory award is based. I would affirm the trial court’s judgment notwithstanding the verdict (JNOV) as to punitive damages.
“No Missouri - case has permitted submission of a punitive damage claim in a strict products liability case on the theory that the defendant should have known of a dangerous defect in its product.” School *48Dist. of City of Independence, Mo., No. 30 v. U.S. Gypsum Co., 750 S.W.2d 442, 446 (Mo.App.1988). To make a submissible case, a plaintiff must prove that the defendant had actual knowledge of the defect. Id. And because punitive damages are extraordinary and harsh, the evidence must meet the “clear and convincing” standard of proof. Rodriguez v. Suzuki Motor Corp., 936 S.W.2d 104, 111 (Mo. banc 1996). Neither Plaintiffs nor the majority opinion supply authority upholding an inference of actual knowledge based on expert speculation of what a defendant should or must have known, which describes Plaintiffs’ evidence here. Moreover, even accepting Plaintiffs’ charge that Doncasters must have known something, there is no evidence whatsoever that Don-casters was aware of the defect on which the compensatory verdict is based. Ample precedent suggests that “clear and convincing” requires more than the record contains here, even in a threshold submis-sibility determination.1
In U.S. Gypsum, the school district sought punitive damages from the manufacturer of a ceiling plaster containing asbestos, sold to the district from 1957 to 1969. The district’s expert testified that he showed his asbestos research to USG’s safety manager in 1955 and later wrote to USG “protesting the ‘folly’ of adding asbestos to gypsum.” Id. at 447. The district also cited “USG’s longstanding knowledge of the hazards of asbestos when used at its own plants, the burgeoning medical knowledge of the hazards of asbestos exposure in non-occupational settings, and continued reports of water damage to [plastered] ceilings” as facts sufficient to place USG on notice of the danger of its product. The trial court granted JNOV and the Western District affirmed, reasoning that, even if USG was informed of the expert’s experiments and opinions concerning the potential danger of incorporating asbestos into gypsum products, and even accepting the district’s characterization of USG’s general knowledge of asbestos dangers and water damage to the district’s ceilings, the evidence was insufficient to demonstrate USG’s actual knowledge of, specifically, its plaster’s propensity to release asbestos. Id. For the proposition that actual knowledge can be inferred from a defendant’s general knowledge of product hazards, the district cited Hale v. Firestone Tire & Rubber Co., 756 F.2d 1322 (8th Cir.1985). The Western District deemed Hale inapposite because, there, the plaintiff adduced an internal memorandum in which Firestone sought a list of complaints regarding not only the product (a tire rim) but also, specifically, the danger associated with that product (explosive separations). Such evidence “strongly supported the finding of actual knowledge,” whereas the school district produced no evidence comparable to the Firestone memo. Id.
Providing an even starker contrast from the present case than U.S. Gypsum is Angotti v. Celotex Corp., 812 S.W.2d 742 (Mo.App.1991). There, the defendant’s worker’s compensation manager hired a lung disease expert to study the company’s *49plant. Both the expert and the manager who hired him submitted multiple and increasingly emphatic communications cautioning the defendant of the hazards of asbestos to employees and consumers and of the inevitability of resulting liability. Yet the majority of the court (J. Kennedy dissenting) still found the evidence insufficient to prove the defendant’s actual knowledge of a specific health hazard to its insulators. Id. at 748-750. See also Hogan v. Armstrong World Industries, 840 S.W.2d 230 (Mo.App.1992) (following An-gotti to conclude that evidence failed to establish defendant’s actual knowledge of the dangers of asbestos at the time of plaintiffs exposure).
Likewise, in Peters v. General Motors Corp., 200 S.W.3d 1 (Mo.App.2006), two GM engineers testified that they reported concerns to GM about the subject cruise control system, and 74 reports involving cruise control incidents were admitted into evidence. Yet the majority of the court (J. Lowenstein dissenting) still found the evidence insufficient to prove GM’s actual knowledge of the specific design defect. “GM must have had actual knowledge of the cruise control’s dangerous proclivity to permit a transient charge to enter the device and cause the vehicle on which it was installed to accelerate without driver input.” Id. at 27. In Sparks v. Consolidated Aluminum Co., 679 S.W.2d 348 (Mo.App.1984), involving slipping ladder feet, the defendant’s engineer testified that there had been other claims of injury and that “there might have been a better substance” than the product in question. But the court deemed such evidence insufficient, noting that the evidence of other claims was too vague to establish knowledge of the particular danger at issue, and that the product was tested and certified and met applicable standards. Id. at 354. In Lewis v. Envirotech Corp., 674 S.W.2d 105 (Mo.App.1984), the plaintiff asserted a “failure to test” theory, but the court held that evidence that the defendant relied on field tests and service reports rather than simulated tests was insufficient to submit the issue of punitive damages to the jury. Id. at 114.
Applying the above precedent to the record here, it becomes apparent that the evidence of Doncasters’ actual knowledge of the blades’ specific design defect is insufficient to make a submissible case. Even viewed in the most favorable light, Plaintiffs’ strongest evidence is not clear and convincing evidence of Doncasters’ actual knowledge of the specific defect supporting the compensatory award, namely the alloy and coating materials. As noted by the majority, Plaintiffs’ experts testified about two unsuccessful endurance tests. But neither test identified the materials as the cause of failure. The first was interrupted because both the PWC and Don-casters products failed, so nothing about that test could have indicated an inferiority of Doncasters’ product. The second test was interrupted due to “excessive stretching” caused by engine heat. The report stated that “the nearly burnt thru compressor turbine vane may have been a contributing factor to the excessive blade stretching .... All of the other characteristics on the blades which ran 113.5 hours were exemplary.” Thus none of the testing documentation suggested that Don-casters’ coating and alloy materials were prone to cracking, corrosion and oxidation. As they were quoted in the majority opinion, the experts merely asserted that Don-casters knew of the unsuccessful tests and therefore knew that the blades were defective. Although the experts were prompted by counsel to utter the words “actual knowledge,” logic dictates that, in reality, the experts were either speculating on Doncasters’ state of mind, which is improper and inadmissible (but presented *50here without objection), or opining as to what Doncasters should have known based on the tests, which didn’t identify the specific defect. To the extent the testimony relates at all to the specific defect supporting the compensatory award, it is speculation deemed insufficient under Missouri law. From Dr. Hoeppner:
Q: Did you determine whether or not Doncasters knew that the base material in the coating of the CT blade could not meet the operating parameters for the life limit of the 20 engine?
A: Absolutely. They set a life limit— well, the time of the first overhaul, to my recollection, of 5000 hours.... and 3000 hours thereafter. They had no basis for doing that, and they had to know that....
Ironically, Plaintiffs’ counsel further illustrated this point at oral argument. Noting that the only difference between PWC’s blades and Doncasters’ blades was in the coating and alloy materials, he insisted that the defect “had to have been the inferior coating and/or in combination with the inferior substituted base metal.... There was no other explanation ...” While this deduction may be sufficient as a basis for compensatory damages, it does not constitute clear and convincing evidence to prove Doncasters’ actual knowledge of that specific defect for purposes of imposing punitive damages. (See footnote 1 supra.) Rather, the experts’ testimony is mere speculation and conjecture as to what Doncasters should have known.
When punitive damages are at stake, Missouri precedent universally requires clear and convincing evidence of actual knowledge of the specific defect at issue. Measured against those cases, Plaintiffs’ best evidence here — that the blades never passed a test in the subject engine — hardly constitutes sufficient evidence from which a reasonable juror could conclude, without leaping, that by 1996 Doncasters actually knew specifically that the alloy and coating materials rendered the blades defective.
A submissible case requires substantial evidence to support each element. U.S. Gypsum, 750 S.W.2d at 446. When the record is insufficient on the element of actual knowledge, it is unnecessary to address the sufficiency of the evidence of the defendant’s alleged conscious disregard for others’ safety. Id. at 448. Indeed, logic dictates that one must have knowledge of a danger in order to disregard it consciously. As such, this second prerequisite for punitive damages necessarily fails as well. Nevertheless, to exhaust any doubt, even examining the “conscious disregard” prong independent of its predicate, again Plaintiffs’ strongest evidence fails to establish that Doncasters acted with “evil motive or reckless indifference.” See Peters, 200 S.W.3d at 25. The record on this element, too, is insufficient to make a submissible case.
Because the remedy is so extreme, punitive damages should be applied sparingly and only when the evidence is “clear and convincing.” Peters, 200 S.W.3d at 25. Plaintiffs failed to satisfy this standard. I would affirm the trial court’s JNOV in favor of Doncasters as to punitive damages.
. There is an obvious tension between the standards of review applicable to this case. The majority correctly notes that the general standard of review for submissibility is whether the record contains any evidence to support the jury’s verdict. However, when the precise issue before the court is whether the evidence supports a finding of actual knowledge warranting the imposition of punitive damages, the cases discussed infra instruct that the more specific standard applicable to that issue is "clear and convincing” evidence. There is a vast difference between the quantum and quality of evidence comprising "a complete absence of probative facts” (i.e., none) versus "clear and convincing evidence.” In my view, the latter applies here and supports JNOV.