AMENDED MEMORANDUM OPINION AND ORDER
JAMES B. ZAGEL, District Judge.I. INTRODUCTION
The issue of willful infringement is, by agreement, to be tried before me in one or two phases. First, I must find whether there is willful infringement. Second, if and only if, I find willful infringement, I must go on to consider what damages should be awarded to Plaintiff. My decision as to liability for willful infringement follows.
II. FINDINGS OF FACT AND OBSERVATIONS UPON THE FACTS1
What I must decide here is whether Ford Motor Company’s infringement of the patent of Jacob Krippelz, Sr. was willful. The current law of willful infringement is found in the en banc case of In re Seagate Tech., LLC, 497 F.3d 1360 (Fed.Cir.2007) (overruling precedent that established a principle of willful conduct which was closely akin to a negligence rule, since it imposed upon an infringer a duty of due care to avoid infringement, and defining a new standard):
[T]o establish willful infringement, a patentee must show by clear and convincing evidence that the infringer acted despite an objectively high likelihood that its actions constituted infringement of a valid patent. The state of mind of the accused infringer is not relevant to this objective inquiry. If this threshold objective standard is satisfied, the patentee must also demonstrate that this objectively-defined risk (determined by the record developed in the infringement proceeding) was either known or so obvious that it should have been known to the accused infringer.
497 F.3d at 1371 (citation omitted). The Federal Circuit left it to future cases to develop the ways in which the standard should be applied. Some opinions refer to the question to be determined after the objective risk threshold is crossed as the subjective aspect of the test. I tend to think of it as the state of mind issue since, in law, state of mind includes both actual states of mind (knowing) and inferred states of mind (should have known means known). Under the Seagate standard, both “knowing” and “should have known” are legally culpable states of mind.
Claims of willful infringement are governed by time and circumstance at each stage of the defendant’s infringement. *809The first stage customarily begins when the inventor sends the alleged infringer some form of notice of his, her or its invention, either as part of a cease-and-desist demand or an offer to sell one or all of the owner’s patent rights. The second stage usually begins at a point soon after the lawsuit is filed. Other stages may commence after a decision on re-examination, particular rulings by the court or the return of a verdict of infringement. In this case, I use the stages suggested by Plaintiff; they seem reasonable as dividing lines in this case.
A. Period 1
Period 1 includes the time before the suit was filed in April 1998 and ends in late March 1999 when parties secured an agreed order concerning discovery, and a reexamination of the patent was sought. The significant facts I find with respect to this period (and all following periods) are these:
1. Krippelz sent a copy of his '903 patent to Ford soon after it issued in 1991.2
2. Ford truthfully told Krippelz, at one time, that it was not interested in the invention.3
3. Ford’s puddle lamp supplier, Donnelly, had possession of the '903 patent in mid-1993 before it sold a puddle lamp mirror housing to Ford years later.
4. The Donnelly '659 patent, which Ford had and examined, refers to the Krippelz patent.
5. Ford did not (a) investigate the '903 patent to determine possible infringement or invalidity during this period or (b) possess the '903 prosecution history prior to this litigation.
6. During this period Ford possessed no potentially invalidating prior art that it ever asserted here.
7. Ford’s specification (and one supplier’s drawings) depicted light illuminating an area on the ground next to the vehicle, rather than illuminating the vehicle’s door.
8. Donnelly designed and made Ford’s first lamp.
B. Period 2
Period 2 begins in March 1999 when Ford disclosed an invalidity defense based on the Kim patent and five other references, and ends when the '903 patent was *810confirmed by the Board of Patent Appeals in February 2002.
The facts I find with respect to this period (and, where relevant, all following periods) are these:
1. Ford offered, in this litigation, challenges to validity which were rejected by the Board of Patent Appeals and were not offered at trial.
2. Ford asserted that its primary defense was non-infringement.
3. Ford sought summary judgment on non-infringement but offered no claim construction to support its argument, and any well-founded argument would require construction of terms which were not addressed within Ford’s motion.
4. Ford unreasonably relied upon its first non-infringement argument — that the '903 patent claim 2 cannot cover any lighting device that shines any type of light on the side of the vehicle — since the patent did not assert that no light from the device would strike the side of the vehicle.
5. Ford unreasonably relied upon its second non-infringement argument— that Krippelz had disclaimed coverage of warning lights that facilitated ingress and egress to the vehicle as to the claim at issue in this case offered a second non-infringement argument that was transparently wrong — since the disclaimer was made with respect to claims not at issue in this lawsuit.
6. Ford added new puddle lamp designs to other Ford vehicles without conducting a technical analysis of those devices to determine whether they might infringe the '903 patent in suit.
7. Krippelz could reasonably conclude that no purpose would be served by seeking a preliminary injunction.
C. Period 3
Period 3 begins with the Board of Patent Appeals decision in February 2002 and ends with Ford’s disclosure of new prior art in April 2007. During this period (and, where relevant, for the final period as well) I find:
1. Ford did not disclose any new prior art on which it relied until 2007.
2. Ford did not pursue any of the prior art that occupied the Board of Patent Appeals.
3. Ford offered no viable prior art defense for five years.
4. Ford did defend on non-infringement grounds.
5. Ford offered non-infringement defenses that required a significantly different claim construction than that employed by the Board.
6. Ford offered non-infringement defenses that were objectively unreasonable.
7. (a) Delay in resolving the question of which Ford lamps were infringing was a consequence of flawed testing procedures used by Ford, and (b) during that delay new infringing devices were used on several Ford models.
D. Analysis of Periods 1-3
When viewed in its totality, Ford’s conduct throughout periods 1-3 clearly and convincingly demonstrates that (1) Ford “acted despite an objectively high likelihood that its actions constituted infringement” of Krippelz’ valid patent, and (2) this objectively-defined risk was so obvious that it should have been known to Ford. Seagate, 497 F.3d at 1371.
Beginning in period 1, Ford had actual knowledge of the Krippelz patent.4 *811Ford had also examined the Donnelly '659 patent, which references the Krippelz patent.5 Despite the fact that Ford licensed the Donnelly patent, it did not investigate the Krippelz patent or look into its prosecution history. Ford began its infringing sales in 1997 and for nearly two years after, Ford failed to produce any allegedly invalidating prior art.
During period 2, Ford sought summary judgment on non-infringement, but failed to offer claim construction to support its argument. See Abbott Labs. v. Sandoz, Inc., 544 F.3d 1341, 1358 (Fed.Cir.2008) (“The first step in most infringement suits is the procedure called ‘claim construction,’ where the scope of the claim is defined by the court.”). Instead, Ford offered two insufficiently supported non-infringement arguments: (1) that claim 2 of the '903 patent cannot cover any lighting device that shines any type of light on the side of the vehicle, even though the patent did not assert that no light from the device would strike the side of the vehicle (essentially ignoring the “conical beam of light” limitation in claim 2), and (2) that Krippelz had disclaimed coverage of warning lights that facilitated ingress and egress to the vehicle, even though the disclaimer was made with respect to claims not at issue in the lawsuit. Furthermore, in 2001, Ford added new puddle lamp de signs to other Ford vehicles without conducting an infringement analysis.
During period 3, Ford continued to argue non-infringement. However this time, rather than ignore claim 2’s conical beam of light limitation, Ford sought a construction of this claim term that would include all of the light that leaves the disputed device. This proffered claim construction was objectively unreasonable at least in part because it required a claim-reading approach significantly different from that employed by the Board of Patent Appeals, which found that a bulb in a hood does not a beam make. Throughout this phase, Ford continued to pursue its non-infringement claim even after its proffered claim construction was rejected by both the Board of Patent Appeals and this court. Again, during this period, Ford continued to add infringing puddle lamps to more vehicles.
While it is true that the law no longer imposes on a potential infringer an “affirmative duty to exercise due care to determine whether or not he is infringing[,]” Underwater Devices, Inc. v. Morrison-Knudsen Co., 717 F.2d 1380, 1389-90 (Fed.Cir.1983), overruled by Seagate, 497 F.3d at 1371, Seagate does not require that I ignore the record. To the contrary, Sea-gate dictates that objective recklessness be *812determined by reference to the record developed in the infringement proceeding. Id. Ford’s conduct throughout periods 1 through 3 indicates a consistently reckless state of mind. One example of this recklessness occurred in October 1998, when Donnelly licensed the rights of the device to Ford. Ford had examined the Donnelly patent; there was no need for Ford to look into prior art or to seek legal advice from counsel to know that there was a high likelihood that its lamps were infringing an existing patent. All Ford needed to do was look at what was already before it, and by failing to do so, Ford acted recklessly. Furthermore, Ford did not produce any allegedly invalidating prior art until nearly two years after its first infringing sales. Ford had no good faith basis to believe that the Krippelz patent was invalid.
A finding of recklessness during periods 1 through 3 is bolstered by Ford’s continuing failure to conduct an infringement analysis during period 2, all the while, adding the infringing device to more and more vehicles. Ford continued to sell the infringing devices even through period 3, when it was clear that Ford had no objectively reasonable non-infringement or operative prior art defense. The fact that Ford consistently sold infringing devices during these periods, without investigation and regardless of the availability of reasonable defenses, demonstrates that Ford acted with the same state of mind throughout these periods, recklessly ignoring an obvious risk.
Ford’s primary ground for defeating a claim of willful conduct during period 2 is that Krippelz waived his right to damages for willful infringement by failing to move for a preliminary injunction. Ford relies on Ball Aerosol & Specialty Container, Inc. v. Limited Brands, Inc., for the proposition that “when a patent holder makes no attempt to stop infringement after litigation has commenced, despite a legal mechanism for doing so, willful infringement should not be found on the basis of such post-filing conduct.” 553 F.Supp.2d 939, 953 (N.D.Ill.2008) (citing Seagate, 497 F.3d at 1374), vacated, 555 F.3d 984 (Fed.Cir.2009).
This argument fails for two reasons. First, there was pre-litigation willful conduct by Ford, which knew of and received a copy of the '903 patent and proceeded in the face of an objectively high likelihood of infringement. Second, the proposition that failure to seek a preliminary injunction constitutes a forfeit of a claim for willful infringement is neither an absolute nor a general rule applicable to all patent cases.
The pursuit of preliminary injunction may fail for reasons other than its lack of merit on the underlying claim. Even a strong claim for injunctive relief will ordinarily fail when an inventor, who does not practice the invention and does not compete with the infringer, sues for injunction. It is a general rule of equity that injunctions should not issue where money damages will suffice because irreparable injury cannot be shown. This principle is applicable to patent injunctions. eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388, 126 S.Ct. 1837, 164 L.Ed.2d 641 (2006). I can envision a rare case where an inventor, while not in direct competition with an infringer, could secure a preliminary injunction if the inventor is entitled to a preliminary injunction, had well-formed plans to compete against the infringer and demonstrated, for example, that the infringer was effectively locking up, for a period of years, a substantial share of customers for the patented device. There is no flavor of that in this case. Krippelz could not have gotten an injunction in this case, and, while only I know that for a certainty, Krippelz could reasonably and *813wisely conclude that he would not win; that it would be a waste of time and resources, his own and the court’s, to make the effort.
Ford suggests that apart from the failure to seek an injunction, its conduct is mitigated by the fact that Krippelz did not actively pursue summary judgment on infringement. This is not a persuasive argument. Summary judgment (sought by both Krippelz and Ford) was delayed, in significant part, by disputed device-testing initiated by Ford, and also by Ford’s persistence in making arguments in contravention of a proper construction of the patent.
Finally, it is difficult to look at the Ford defense during period 3 and find any reasonable basis for what it did as a whole. At least some of it falls into the category of last ditch desperation.
Ford’s testing was deeply flawed and it took a lot of time and effort to resolve the problems created by that testing. I do not rehearse the details of the testing controversy because there is already an inordinate portion of the record devoted to the question. The delay is, in largest part, to be laid at Ford’s door.6
I noted at trial that, assuming Ford’s good faith in its efforts to test the device, which I was willing to do, Ford’s testing “created a needless controversy which delayed the litigation for no good reason. To put it another way, [Ford] used a seat-of-the-pants design to engineer its defense, either because it did not want to bear the cost of better design or because it knew better design would not help its cause.”
By the end of this period, Ford had offered nothing that it could use in defense of the claim. All it had done was try to wriggle out of my Markman ruling by indirect methods. My ruling is not sacred writ and, if it is in error, the judgment against Ford likely will fall. But reasonable defenses are limited to those consistent with that ruling as long as the case is in this court, and that is not a limit that Ford honored. Much, if not all, of what it offered in its defense until April 2007 was impermissible, and what may not have been impermissible was clearly without merit.
I find, by clear and convincing evidence, that Ford willfully infringed the '903 patent throughout periods 1 through 3.
E. Period 4
Period 4 begins with the disclosure of three pieces of prior art and ends with the verdict at trial. With regard to this period, I find that:
1. Ford offered three pieces of prior art in support of its claims of invalidity, to wit, Miazzo, Siegfried and DuBois in support of anticipation and obviousness.
2. Ford declined to offer Siegfried at trial and the patent could not support a finding of invalidity.
3. Miazzo was, as Ford should have known, inadmissible in light of the claim construction ruling and the fact it teaches away from the '903 patent.
*8144. DuBois was admitted, but it barely passed the threshold of admissibility.
5. Dubois did not disclose the “conical beam of light” or the proper placement of the light.
6. Ford offered an inadmissible computer simulation, created in late 2007, in support of its defense.
The case against willful infringement during this period is, ironically, the strongest that Ford has to offer because it had some evidence on which it could rely to conclude that the objectively high likelihood of infringement of a valid patent was not known to it or obvious enough that it should have known. At least, Ford could conclude that Krippelz could not, in these circumstances, prove willfulness by clear and convincing evidence. The law permits a party to defend against the accusation of the state of mind (knew or should have known) element of a willful infringement claim by showing that at certain periods (if not at all periods) the state of mind element of willfulness did not exist.
I conclude that Ford has not been proven to be a willful infringer during this last period. The fact that Ford had some evidence of invalidity on which it could reasonably rely serves to assist Ford in its arguments that willful infringement is not proven with respect to period 4. This also reinforces my earlier conclusion that for periods 1 through 3, Ford was a willful infringer, as it did not have even the faint justification for its conduct that it possessed in period 4.
F. General Observations
One element of Ford’s argument against the claim of willful infringement is that its defenses, while not good enough to win on the merits, were good enough to indicate that the risk of infringement was not so obvious that Ford should have known of the risk. Ford can argue that its positions, even if incorrect, show that its continued infringement was not willful.7 But there was nothing of its defenses in periods 2 and 3 that were more than makeweights. Indeed, if it were not for the requirement that willful infringement must be shown by clear and convincing evidence, the weak defense that they did offer would not have saved them from a finding of willful infringement in period 4.
To the extent that Ford offers evidence, as an infringer may be entitled to do,8 that its heart was pure on the issue of infringement even if its head was empty, it is beyond the boundaries defined by the way this case was tried. Subjective state-of-mind evidence was discussed earlier in the case, but Ford did not offer the testimony of the actual state-of-mind witness that it named in pre-trial proceedings. The lawyers’ arguments referring to what Ford employees and agents believed or did not believe are not evidence and are given no weight.
III. CONCLUSIONS OF LAW
1. The Court has jurisdiction over the subject matter and the parties.
2. Krippelz has proven that from prior to the initiation of this litigation in April *8151998 to April 2007, Ford willfully infringed the '903 patent.
3. Krippelz has not proven that Ford willfully infringed the '903 patent from April 2007 to the close of trial.
I thereby order Plaintiff to prepare a formal draft of complete findings of facts and conclusions of law, if necessary, consistent with this opinion. This draft shall be served upon Ford which must confine its objections to claims that the final draft is inconsistent with this opinion. By so limiting its objections, consistent with this order, Ford shall not be deemed to waive any other objections it may wish to present in post-trial motions. After reviewing Plaintiffs draft and Defendant’s objections, I will issue a final order with respect to findings of fact and conclusions of law. I intend to proceed to determine the scope of remedy for willful infringement concurrently with the process outlined above.
. The facts I find are included in both the numbered sentences and interspersed among the observational passages as well. These facts are those which I deemed important enough to be stated explicitly. These are not all the facts required for the comprehensive order required by Federal Rule of Civil Procedure 52. I will therefore follow my usual practice of issuing a basic set of my findings on specific issues accompanied with an explanation of how I view the evidence I heard. I then order the prevailing party to prepare a more comprehensive set of findings of fact and conclusions of law, consistent with the views expressed in this opinion.
. Ford argues that Krippelz did not produce the letter that accompanied a copy of his patent, from which fact Ford infers the patent was not sent to Ford. What Krippelz did produce was a draft in which he stated he had applied for a patent and would tender the patent application if Ford signed a confidentiality agreement. Implicit in the circumstances here is the possibility that the draft letter may never have been sent, perhaps because Krippelz waited for the patent to issue or it issued before the letter was sent. Once the patent issues there is no need for a confidentiality agreement. I make no finding with respect to the draft letter. However, Krippelz testified that he sent the patent and was neither challenged on cross-examination nor confronted with the draft. I credit Krippelz’s testimony at trial which was offered in my presence. And the sending of the patent makes good sense for Krippelz. He was not in a position to put his invention into commerce, he did not make automobiles or their parts. His market was automakers and their suppliers. It is quite reasonable that he would send his patent, once issued, to an automaker, and that is what he did.
Knowledge of the '903 patent is not the same thing as knowledge of the high likelihood that one’s actions constituted infringement of a valid patent. However, one is far more likely to know of the infringement risk if one knows of the patent.
. I credit Ford’s implicit position that it was not, in fact, interested. Ford had not sold lamps of the type Krippelz was suggesting at the time it received the patent. The lamp that Ford did sell was brought to it in 1993 by Donnelly, a part maker who wanted to sell the part to Ford.
. Krippelz argues that Donnelly was an agent of Ford, and it clearly had knowledge of the '903 patent. The agent’s knowledge can be imputed to the principal, although the case *811law on the point is undeveloped. The premise upon which Krippelz relies is that the knowledge of Donnelly is Ford’s knowledge ”[t]o the extent that Ford relied on Donnelly to investigate intellectual property issues or indemnify Ford regarding Donnelly’s puddle lamp concept!.]” But that "extent” is not clear to me. The evidence shows that Ford requires its suppliers to indemnify it if it must pay infringement damages for use of the supplier’s parts, but this is not necessarily sufficient to find that Ford relied upon the supplier to investigate infringement. Specific reliance on a supplier to investigate infringement is precisely the kind of relationship that can often be deemed as one of principal and agent. The expectation that a supplier's self-interest arising from the obligation to indemnify will cause the supplier to check for infringement may not suffice to create agency. I see no need to sort through the details of the relationship and the fuzzy precedent to resolve this question when I have found as a fact that Ford had direct knowledge of the '903 patent.
. Ford’s examination of the Donnelly '659 patent prompted it to license the device from Donnelly, ending a dispute between Donnelly and one of its competitors that supplied Ford with an allegedly infringing device.
. What Ford can and does argue is that I did not grant summary judgment on infringement until the eve of trial. Part of that is due to the presence of more than this case on my docket, and part of it is due to the manner in which the case was litigated. In opposition to Krippelz’s motion for summary judgment, Ford filed an expert's report which in its conclusions clearly established that there were triable issues on infringement, but only upon very close examination did it become apparent that these conclusions could be reached only by disregarding the way I had construed the patent. It is, ordinarily, not to be expected that an expert finally selected by one party to a patent case would offer opinions inconsistent with the Markman rulings in that case.
. It seems to me that Ford has offered arguments, new in nature and in form. These I ignore as waived.
. I say "may” because it is difficult to see how, in an ordinary case, evidence that an infringer personally believed it was not infringing has real importance under Seagate which holds that the standard is "should have known.” Evidence of actual knowledge of the risk of infringement would be of crucial importance to a plaintiff if it did exist. The absence of actual knowledge is not a meaningful circumstance in the determination of whether an infringer should have known of the risk of infringement.