Hemlock Semiconductor Corp. v. Kyocera Corp.

NOT RECOMMENDED FOR FULL-TEXT PUBLICATION
File Name: 18a0413n.06

Case No. 17-2276

UNITED STATES COURT OF APPEALS
FOR THE SIXTH CIRCUIT
FILED
Aug 16, 2018
DEBORAH S. HUNT, Clerk
HEMLOCK SEMICONDUCTOR )
CORPORATION; HEMLOCK )
SEMICONDUCTOR, LLC, )
) ON APPEAL FROM THE UNITED
Plaintiffs-Appellees, ) STATES DISTRICT COURT FOR
) THE EASTERN DISTRICT OF
v. ) MICHIGAN
)
KYOCERA CORPORATION, )
)
Defendant-Appellant. )

BEFORE: COLE, Chief Judge; CLAY and THAPAR, Circuit Judges.

THAPAR, Circuit Judge. Kyocera Corporation is locked in a long-running bout with

Hemlock Semiconductor Corporation and Hemlock Semiconductor, LLC, which supply Kyocera

with polysilicon that it uses to make solar panels. Kyocera is fighting to get out of certain

obligations under the parties’ contracts. Below, the district court declared victory for Hemlock.

But at this stage, Hemlock has earned only a partial victory. We therefore reverse in part and

affirm in part.

I.

In the mid-2000s, the market for solar panels was taking off. Kyocera needed a steady

supply of quality polysilicon. So it entered into four contracts with Hemlock, in which Kyocera
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

promised to purchase specified amounts of polysilicon from Hemlock at specified prices over the

course of the next ten years or so.

Those contracts contain so-called “take-or-pay” provisions. Under these provisions, the

contracts require Kyocera to “take” a specified quantity of polysilicon from Hemlock each year.

But if Kyocera does not want to take the polysilicon in a given year, Kyocera still has to “pay” full

price for it. This means that Kyocera is on the hook for a certain quantity of polysilicon annually,

whether it takes the polysilicon or not.

The contracts also contain so-called “acceleration” provisions. If Kyocera defaults under

the contracts, these provisions accelerate the amount it owes Hemlock, such that Hemlock can

demand all remaining sums owed. For these acceleration provisions to take effect, Kyocera must

default, Hemlock must serve notice of default, and Hemlock must give Kyocera 180 days to cure

its default. But if Kyocera does not cure, Hemlock can elect to terminate, at which point Kyocera

becomes liable for all remaining payments due—effectively, the sum of the take-or-pay provisions.

Several years into Kyocera and Hemlock’s deal, the Chinese government disrupted the

solar-panel market by subsidizing Chinese solar-panel companies. This intervention reduced the

market price of polysilicon such that the price Kyocera agreed to pay Hemlock was far greater

than the going rate. And so Kyocera sought to renegotiate. Initially, the parties came to a

compromise, temporarily lowering the price of polysilicon under the parties’ deal. But eventually,

Hemlock signaled that it would begin insisting that Kyocera take or pay for polysilicon at the

original (and now inflated) price.

This litigation ensued. Hemlock sought a declaratory judgment that Kyocera had

repudiated the parties’ contracts by indicating that it would not take or pay at the original price. In

response, Kyocera counterclaimed, seeking a declaratory judgment that the “pay” portion of the

-2-
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

take-or-pay provisions is an unlawful penalty, and thus that the acceleration provisions are too.

Kyocera also counterclaimed for breach of contract, alleging that three of the parties’ contracts

obligated Hemlock to expand certain production facilities, which Hemlock had not done. Hemlock

moved to dismiss Kyocera’s counterclaims, and the district court agreed. Kyocera now appeals.

II.

Kyocera first appeals the dismissal of its challenge to the take-or-pay provisions. We

review the district court’s decision de novo. JPMorgan Chase Bank, N.A. v. Winget, 510 F.3d

577, 581 (6th Cir. 2007). In doing so, we accept Kyocera’s well-pled allegations as true and ask

whether Hemlock is nevertheless “clearly entitled to judgment.” Id. (quoting S. Ohio Bank v.

Merrill Lynch, Pierce, Fenner & Smith, Inc., 479 F.2d 478, 480 (6th Cir. 1973)). We view the

facts as alleged in the light most favorable to Kyocera and draw all reasonable inferences in its

favor. See Gavitt v. Born, 835 F.3d 623, 640 (6th Cir. 2016). Our task is to determine whether

Kyocera raises a plausible claim for relief. See Bell Atl. Corp. v. Twombly, 550 U.S. 544, 557

(2007). And Kyocera’s claim is plausible if, assuming the truth of Kyocera’s allegations, a

reasonable factfinder could rule in its favor. Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009).

In assessing Kyocera’s claim, we apply Michigan law—the law of the forum state and that

designated in the parties’ contracts. Erie R. Co. v. Tompkins, 304 U.S. 64, 78 (1938); see Kipin

Indus., Inc. v. Van Deilen Int’l, Inc., 182 F.3d 490, 493 (6th Cir. 1999). As it happens, Michigan’s

courts provide little guidance here. The thrust of Kyocera’s claim is that the take-or-pay provisions

are unlawful penalties in disguise. But there is no case in which a Michigan court has considered

such a claim. The closest we have is a Michigan Court of Appeals decision resolving an earlier

chapter of the parties’ dispute in which Kyocera attempted to invoke a force majeure clause in the

contracts. In dicta, the court referenced the take-or-pay provisions, but only to note their existence

-3-
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

and operation as a means of refuting Kyocera’s force-majeure argument. Kyocera Corp. v.

Hemlock Semiconductor, LLC, 886 N.W.2d 445, 447–48, 453 (Mich. Ct. App. 2015). Hemlock

makes much of this discussion, reading it to suggest that Michigan always enforces take-or-pay

provisions and would do so here. But because Kyocera did not challenge the validity of the take-

or-pay provisions in those proceedings,1 we cannot read the Michigan court’s discussion as setting

out a general rule that it will always enforce take-or-pay provisions or even that it would do so in

this case. And neither of the other cases Hemlock identifies goes so far. See McLouth Steel Corp.

v. Jewell Coal & Coke Co., 570 F.2d 594, 605 (6th Cir. 1978); Attorney Gen. v. Pub. Serv. Comm’n

No. 1, 431 N.W.2d 47, 49 (Mich. Ct. App. 1988).

With no Michigan authority on point, we must look elsewhere to attempt to discern what

path the Michigan Supreme Court might take. Combs v. Int’l Ins. Co., 354 F.3d 568, 577 (6th Cir.

2004) (explaining that “when evaluating an undecided question of [state] law, a federal court

sitting in diversity must make the [sic] ‘the best prediction, even in the absence of direct state

precedent, of what the [state] Supreme Court would do if confronted with [the] question,’”

including considering “jurisprudence from other jurisdictions” (last alteration in original) (first

quoting Managed Health Care Assocs., Inc. v. Kethan, 209 F.3d 923, 927 (6th Cir. 2000); then

quoting Lexington Ins. Co. v. Rugg & Knopp, Inc., 165 F.3d 1087, 1090 (7th Cir. 1999))). Under

the approach followed in other jurisdictions, the key question is whether the take-or-pay provisions

offer Kyocera two viable performance options, on the one hand, or one performance option

coupled with a liquidated damages provision, on the other. See, e.g., Superfos Invs. Ltd. v.

FirstMiss Fertilizer, Inc., 821 F. Supp. 432, 434–35 (S.D. Miss. 1993) (collecting cases); Minnick

1
Hemlock makes no argument that Kyocera is precluded from raising this challenge here, despite not doing
so previously.
-4-
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

v. Clearwire U.S. LLC, 275 P.3d 1127, 1130–31 (Wash. 2012) (en banc); Am. Soil Processing,

Inc. v. Iowa Comprehensive Petroleum Underground Storage Tank Fund Bd., 586 N.W.2d 325,

329 (Iowa 1998); 11-59 Corbin on Contracts § 59.10 (2017); 14 Williston on Contracts § 42:10

(4th ed.). If the former, the take-or-pay provisions are enforceable as written. If the latter, the

question becomes whether the “pay” option quantifies lawful liquidated damages or an unlawful

penalty. If the payment obligation is a penalty, it is unenforceable—regardless of what the parties’

contract labels it. And at least on this point, Michigan law certainly agrees. Mich. Comp. Laws

§ 440.2718(1); Moore v. St. Clair Cty., 328 N.W.2d 47, 50 (Mich. Ct. App. 1982) (“[U]se of the

terms ‘liquidated’ or ‘stipulated’ damages does not necessarily mean that the clause is valid and

not a penalty.”). So if the provisions here are penalties, it is doubtful that Michigan courts would

let them fly by night under the guise of take-or-pay provisions.

Alternative Performance v. Liquidated Damages. First, we ask whether the take-or-pay

provisions offer Kyocera two viable performance options, or one option coupled with liquidated

damages. To make this call, courts consider whether, at the time of contracting, it appears that the

parties intended that the “pay” option present a relatively equivalent (and thus desirable) mode of

performance—and not, as Kyocera claims, a measure to coerce compliance with the “take” option.

See Superfos, 821 F. Supp. at 434; Minnick, 275 P.3d at 1131; Am. Soil Processing, 586 N.W.2d

at 333–34. And at the outset, common sense points to coercion: Why would Kyocera opt to pay

for polysilicon and get nothing in return? See Iqbal, 556 U.S. at 679 (instructing courts to gauge

plausibility by “draw[ing] on [their] judicial experience and common sense”).

Hemlock offers three suggestions. First, Hemlock reasons that Kyocera must have seen

the pay option as a viable alternative, because Kyocera is a big, smart corporation and would never

have agreed to the deal otherwise. But the fact that a sophisticated entity has agreed to pay a sum

-5-
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

does not necessarily mean that the law will always enforce its promise, see, e.g., Lake River Corp.

v. Carborundum Co., 769 F.2d 1284, 1289 (7th Cir. 1985), and Hemlock points to no Michigan

authority holding otherwise. Kyocera’s sophistication at the bargaining table may later become

relevant, but in this posture, it does not doom Kyocera’s claim as a matter of law. So Hemlock’s

first argument swings and misses. Second, Hemlock hypothesizes that Kyocera might have

thought that in certain circumstances it would be willing to pay for polysilicon one year (but not

take it) as a way to keep the contract alive in the event it still wanted polysilicon in the future. But

if Kyocera wanted to keep the contract alive, it could simply purchase polysilicon and take

delivery. It would not need to pay and get nothing in return. So here again, Hemlock swings and

misses.

Hemlock saves its best swing for last. Moving and storing polysilicon is not free. And so

Hemlock theorizes that, if the price of polysilicon tanks (as it did here), it might make sense for

Kyocera to pay for polysilicon without taking it in order to avoid transportation and storage costs.

If the math is right, and the parties intended for the “pay” provision to account for this possibility,

Hemlock may have a point. But nothing in the pleadings suggests that this math influenced the

parties’ negotiations. In fact, Kyocera alleges that there were no such negotiations. Nor do the

pleadings suggest that the price of polysilicon has fallen so much that transportation and storage

costs would justify paying for polysilicon without taking it. At this stage, we consider only

Kyocera’s allegations, viewed in a light most favorable to Kyocera, construing all reasonable

inferences in Kyocera’s favor. See Gavin, 835 F.3d at 640. Accepting Hemlock’s transportation-

and-storage argument at this juncture would disregard those precepts. So we are left with the

common-sense conclusion that paying a lot of money for nothing in return is not a real performance

-6-
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

option under a contract. To be sure, if discovery bears out Hemlock’s point, Kyocera’s claim may

later fail. But for now, Hemlock goes down swinging.

Moreover, two additional allegations further support the conclusion that paying something

for nothing is not a valid performance option. Both have to do with the kind of provisions that

normally appear in lawful take-or-pay contracts. First, the take-or-pay provisions do not contain

“make-up rights” that would credit any money advanced under the “pay” option to Kyocera’s

purchase of polysilicon in future years. Make-up-rights are common in enforceable take-or-pay

arrangements. Superfos, 821 F. Supp. at 436 (collecting cases). Although the absence of make-

up rights may not be dispositive, see World Fuel Servs., Inc. v. John E. Retzner Oil Co., Inc., 234

F. Supp. 3d 1234, 1241 (S.D. Fla. 2017), their absence nevertheless pushes Kyocera’s claim further

into the realm of plausibility, see, e.g., Superfos, 821 F. Supp. at 438–39. Second, while lawful

take-or-pay arrangements often involve the seller (here, Hemlock) bearing construction costs and

associated risks at the outset of a contract, see, e.g., Diamond Shamrock Expl. Co. v. Hodel, 853

F.2d 1159, 1167 (5th Cir. 1988), the contracts at issue required Kyocera to front Hemlock money

for construction and expansion. So Kyocera—not Hemlock—bore these costs and risks from the

start. True, the contracts do require Hemlock to repay this money to Kyocera over time, as a credit

against purchases of polysilicon—meaning that eventually, Hemlock internalizes these costs. But

it is far from clear that the parties envisioned that the pay option would compensate Hemlock for

these costs this late in the duration of the contracts, when construction and expansion are long over

and Hemlock has been providing Kyocera with polysilicon for over a decade. Thus, these

allegations further belie any conceivable purpose for the pay option other than to liquidate

damages.

-7-
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

Liquidated Damages v. Penalty. The fact that the pay option might not be a valid mode of

performance is not the end of the inquiry. If the pay option is a lawful measure of liquidated

damages, then it is nevertheless enforceable. E.g., Superfos, 821 F. Supp. at 440. But if not, it is

a penalty, and Hemlock can seek only its actual damages. E.g., id. at 440–41 & n.8.

Under Michigan law, damages can be liquidated, “but only at an amount which is

reasonable in the light of the anticipated or actual harm caused by the breach, the difficulties of

proof of loss, and the inconvenience or nonfeasibility of otherwise obtaining an adequate remedy.”

Mich. Comp. Laws § 440.2718(1). Otherwise, the clause is “void as a penalty.” Id. Demanding

full price for nothing in return seems unreasonable for one of two reasons. First, if Hemlock has

already produced Kyocera’s polysilicon, requiring Kyocera to pay the full contract price fails to

account for Hemlock’s ability to resell the polysilicon elsewhere. Or, if Hemlock can avoid

producing the polysilicon at all—as Kyocera alleges happened here—requiring Kyocera to pay

full price fails to account for any costs Hemlock would save. Here again, common-sense points

to a problem: Under the pay option, Hemlock can get all the money promised under the contract

but do nothing, thereby making a greater profit than if the parties performed as envisioned.2 See

Iqbal, 556 U.S. at 679.

Our court’s decision in Hemlock Semiconductor Operations, LLC v. SolarWorld Indus.

Sachsen GmbH, 867 F.3d 692 (6th Cir. 2017), does not solve this problem. There, in a dispute

between Hemlock and another buyer, our court affirmed a ruling for Hemlock on summary

judgment that the acceleration provisions in those parties’ contracts were not penalties. Id. at 696–

2
Hemlock points out that a seller has the option to sue for damages under Michigan law even without
reselling its goods. See id. § 440.2703. But in that circumstance, Michigan law only affords “the difference
between the market price at the time and place for tender and the unpaid contract price . . . but less expenses
saved in consequence of the buyer’s breach.” Id. § 440.2708(1). Thus, Michigan law would not permit
recovery of the full contract price.
-8-
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

97, 705–08. Those two pieces of context—the procedural posture and the contractual provisions

at issue—are crucial to understanding why Solarworld is of no help to Hemlock here.

First, Solarworld involved a dispute over acceleration provisions. Id. at 696–97, 705–06.

Those provisions, just like those in Kyocera and Hemlock’s contracts, required the buyer to satisfy

the remainder of the contract in the event of default, i.e., the sum of the remaining take-or-pay

provisions. Id. But unlike in this case, the buyer in Solarworld conceded the validity of the take-

or-pay provisions. See id. at 707. So when the buyer tried to argue that the acceleration provisions

failed to account for Hemlock’s cost savings (as Kyocera does here), the court rejected the

argument. “Although such cost savings might factor into an ordinary breach-of-contract claim,”

the court observed, the acceleration provisions were merely the sum of the unchallenged take-or-

pay provisions. Id. Thus, in Solarworld, the buyer’s cost-savings argument fell flat, but only

because the buyer chose not to challenge the take-or-pay provisions. Had the buyer elected to do

so, that challenge would have been important, just as in the “ordinary” case. See id.

Second, Solarworld reviewed a summary-judgment ruling in favor of Hemlock. Id. at 697,

706. And the summary-judgment record contained facts that we do not have here. As an initial

matter, expert testimony reflected that the buyer had saved (and Hemlock had lost) a large amount

of money early in the contract due to Hemlock pricing the polysilicon well below market. Id. at

707. So the acceleration provisions were something of a fair deal, given that Hemlock was only

making up for what it had lost before. Id. In addition, a forty-one-page expert report established

that calculating Hemlock’s lost profits would have been difficult at the time of contracting. Id. at

707–08. And finally, the Solarworld court was able to conclude the acceleration provisions were

meant to account for Hemlock’s construction costs. Id. at 708. Here, by contrast, the district court

dismissed Kyocera’s claim at the pleading stage—without the benefit of expert testimony or other

-9-
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

record information that might justify the amount owing under the pay provisions. So in this

posture, we cannot say that requiring Kyocera to pay full price for nothing is a reasonable measure

of damages. This is particularly so in light of Kyocera’s allegations that Hemlock could have

adjusted the amount due under the “pay” option to account for cost savings at the time of

contracting, but did not. Down the road, the record may develop in this case such that Kyocera’s

claim falters like the buyer’s in Solarworld. But for now, Kyocera’s claim passes muster, and the

district court erred by dismissing it on the pleadings.3

III.

Next, Kyocera appeals the district court’s decision to dismiss its declaratory-judgment

challenge to the acceleration provisions as unripe. We review that decision de novo. Kiser v.

Reitz, 765 F.3d 601, 606 (6th Cir. 2014). As the district court observed, Hemlock did not seek to

invoke the acceleration provisions at the inception of this case, nor have events come to pass that

would permit Hemlock to do so. Specifically, for Hemlock to invoke the acceleration provisions,

several events would need to occur: Kyocera would need to default, Hemlock would need to serve

notice of default, 180 days would have to pass in which Kyocera could cure, and at the close of

that period, Hemlock would need to elect to terminate the contract. As of right now, none of those

events has occurred. So can Kyocera get a ruling on the validity of the acceleration provisions?

3
The dissenting opinion raises an interesting argument that Kyocera’s failure to plead that specific
performance is available means that its claim must fail. Since Hemlock did not raise this argument before
the district court or on appeal, Hemlock forfeited it. Armstrong v. City of Melvindale, 432 F.3d 695, 700
(6th Cir. 2006). Nor does this argument go to our subject-matter jurisdiction over Kyocera’s challenge to
the take-or-pay provisions. Bell v. Hood, 327 U.S. 678, 682 (1946) (“[I]t is well settled that the failure to
state a proper cause of action calls for a judgment on the merits and not for a dismissal for want of
jurisdiction.”); accord Hrivnak v. NCO Portfolio Mgmt., Inc., 719 F.3d 564, 570 (6th Cir. 2013) (“A bad
theory (whether of liability or of damages) does not undermine federal jurisdiction.” (quoting Gates v.
Towery, 430 F.3d 429, 432 (7th Cir. 2005))); cf. Henderson ex rel. Henderson v. Shinseki, 562 U.S. 428,
435 (2011) (noting the Supreme Court’s effort to “bring some discipline” to courts’ assignment of the
“jurisdictional label,” and explaining that “a rule should not be referred to as jurisdictional unless it governs
a court’s adjudicatory capacity”).
- 10 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

No. To get a declaratory judgment, Kyocera must present a justiciable case or controversy

under Article III. MedImmune, Inc. v. Genentech, Inc., 549 U.S. 118, 126–27 (2007). Specifically,

Kyocera must demonstrate that “the facts alleged, under all the circumstances, show that there is

a substantial controversy, between parties having adverse legal interests, of sufficient immediacy

and reality to warrant the issuance of a declaratory judgment.” Id. at 127 (quoting Md. Cas. Co.

v. Pac. Coal & Oil Co., 312 U.S. 270, 273 (1941)). Kyocera’s challenge fails that test.

The chain of contingencies necessary to trigger the contracts’ acceleration provisions

shows that the controversy here is hardly “immedia[te],” id., and that Kyocera’s prospective

liability under the acceleration provisions is not “certainly impending,” Clapper v. Amnesty Int’l

USA, 568 U.S. 398, 409–10 (2013) (quoting Whitmore v. Arkansas, 495 U.S. 149, 158, (1990));

see MedImmune, 549 U.S. at 128 n.8 (explaining that the “justiciability problem” here “can be

described in terms of standing,” which requires an injury-in-fact). Thus, Kyocera’s situation is not

like MedImmune, where the Court held that there was a justiciable controversy because all that

remained was for a party to refuse to pay royalties. 549 U.S. at 128. In the same vein, damages

under the acceleration provisions are not the equivalent of MedImmune’s royalties. Those

damages are not due and may never be. Kyocera is not faced with paying them or facing suit—

MedImmune’s “dilemma.” Id. at 129. And the circumstances here are not such that, reversing

roles, Hemlock is in a position in which it could sue for damages under the acceleration provisions.

Cf. Surefoot LC v. Sure Foot Corp., 531 F.3d 1236, 1245 (10th Cir. 2008) (Gorsuch, J.)

(confirming jurisdiction to hear declaratory judgment action based on “counterfactual[]”

possibility that defendant could bring “straightforward infringement suit”). That Hemlock has

referenced its rights under the acceleration provisions and invoked them against different buyers

in different circumstances does not negate the contingencies here. So Kyocera’s claim is unripe.

- 11 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

Shifting focus, Kyocera suggests that it has complied with the take-or-pay provisions under

protest only because the acceleration provisions coerced it into doing so. But that argument only

shows why a challenge to the acceleration provisions is unripe. The only immediate injury that

Kyocera faces here is having to pay under the take-or-pay provisions—not having to pay damages

under the acceleration provisions. That injury is traceable to the take-or-pay provisions, not the

acceleration provisions. See Clapper, 568 U.S. at 409 (injury-in-fact must be “traceable to the

challenged action” (quoting Monsanto Co. v. Geertson Seed Farms, 561 U.S. 139, 149 (2010))).

And an order striking down the acceleration provisions would not redress Kyocera’s obligation to

pay under the take-or-pay provisions, so long as those provisions remain in effect. Id. (injury-in-

fact must be “redressable by a favorable ruling” (quoting Monsanto, 561 U.S. at 149)). This means

that Kyocera’s challenge to the take-or-pay provisions must be the first domino to fall. See supra

Part II; cf. Solarworld, 867 F.3d at 707. So long as those provisions are in force, Kyocera must

comply with them. In other words, it is the money owing under the take-or-pay provisions—not

damages under the acceleration provisions—that are the equivalent of MedImmune’s royalties.

Kyocera’s challenge to the acceleration provisions more closely resembles a “declaratory

judgment to litigate a single issue in a dispute that must await another lawsuit for complete

resolution.” Calderon v. Ashmus, 523 U.S. 740, 748 (1998). Considering the validity of the

acceleration provisions in “piecemeal” fashion would not “finally and conclusively resolve” the

parties’ dispute, so long as the take-or-pay provisions remain. MedImmune, 549 U.S. at 127 n.7

(emphasis omitted). Thus, the district court correctly deemed Kyocera’s challenge to the

acceleration provisions unripe.

- 12 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

IV.

The final issue Kyocera raises on appeal is the district court’s dismissal of its breach-of-

contract counterclaim. Kyocera does so conditionally, asking that we reach this issue only if we

validate the take-or-pay or acceleration provisions “because they compensated Hemlock for costs

it had to incur to expand its manufacturing facilities.” Appellant Br. 50. Since we have made no

such ruling, Kyocera’s condition fails, and we need not reach this issue.

In any event, the district court was correct. Kyocera claims that the contracts obligate

Hemlock to expand its facilities. In support of this claim, Kyocera seizes on a sentence fragment

in the contracts that states, “[Kyocera] acknowledges that [Hemlock] will be expanding its

manufacturing facilities.” E.g., R. 89-2, Pg. ID 4163. Emphasizing the “will be” part of that

fragment, Kyocera contends that this line is a contractually enforceable promise. But read in

proper context, this language creates no contractual obligation. In full, the sentence states that

“[Kyocera] acknowledges that [Hemlock] will be expanding its manufacturing facilities (the

“Expanded Manufacturing Facility”) in order to produce the Products to be supplied under this

Agreement.” Id. (emphasis added). And the rest of the corresponding paragraph limits Hemlock’s

liability in the event that its manufacturing expansion causes delay in getting Kyocera polysilicon.

So the purpose of this paragraph is to ward off a potential skirmish resulting from delayed

production—not to obligate Hemlock to build facilities. And that makes sense. The parties’

contracts are for polysilicon, not buildings. Granted, Kyocera made “non-refundable,

unconditional, irrevocable advance payment[s]” so that Hemlock could build facilities to produce

polysilicon. Id., Pg. ID 4162. But the fact that those payments are unconditional only confirms

that the parties did not intend for them to support any enforceable right. Not only that, but those

payments are credited back to Kyocera toward its purchase of polysilicon, meaning the payments

- 13 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

buy polysilicon, not buildings. And so the only reasonable interpretation of the parties’ agreement

is that it obligated Hemlock to do one thing: provide Kyocera with polysilicon. The district court

therefore correctly dismissed Kyocera’s breach-of-contract counterclaim.

* * *

We AFFIRM in part, REVERSE in part, and REMAND for further proceedings

consistent with this opinion.

- 14 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

CLAY, Circuit Judge, dissenting. I am not persuaded that Plaintiff has stated a

declaratory judgment claim for breach of contract. Plaintiff assumes, without explanation, that

Defendant would be entitled to specific performance for breach of the take-or-pay provision.

Under Michigan law, however, the default remedy for a breach-of-contract claim is damages—not

specific performance. Absent a plausible allegation that specific performance is available, the

parties’ “take-or-pay” provision is best interpreted as a “take-or-pay-or-breach” provision.

Because the majority assumes that specific performance is available despite the lack of any

allegation to that effect, I respectfully dissent.

As an initial matter, the majority declines to consider the specific performance issue

because Defendant has not raised it. But Defendant had no reason to raise this issue because

Defendant benefits from Plaintiff’s poorly pleaded declaratory judgment action. Plaintiff is

continuing to buy polysilicon at the contract price pursuant to its “take” obligation out of fear that

a court could order it to specifically perform its “pay” obligation—and to do so under the contract’s

acceleration clause. (See R.127 at PageID #5450 (“Litigation was then stayed through the 2016

calendar year to allow [Plaintiff] to perform on the 2015 contracts.”). This fear, of course, is

unfounded unless specific performance is available, but Defendant would not benefit from saying

so; as long as Plaintiff fears the possibility of being ordered to pay the contract price for nothing,

it will not breach the contract and, significantly for Defendant, Plaintiff might even agree to pay

more than the potential damage award in order to settle the claim. The majority’s assertion that

Defendant has “forfeited” the specific performance issue is therefore misplaced. Where, as here,

a defendant stands to gain more from a plaintiff’s misunderstanding of its rights than from

correcting the plaintiff’s error, the Court cannot rely on the adversarial process alone and may

intervene on behalf of justice, see Dorris v. Absher, 179 F.3d 420, 425 (6th Cir. 1999) (“[T]he

- 15 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

court may choose to entertain arguments not raised by the parties when the failure to do so would

constitute a miscarriage of justice.”); Greenlaw v. United States, 554 U.S. 237, 264 (2008) (Alito,

J., dissenting) (“[T]he interest of the public and the Judiciary in correcting grossly prejudicial

errors of law may sometimes outweigh other interests normally furthered by fidelity to our

adversarial tradition.”).

Furthermore, the Court must address the specific performance issue in this case because,

in the declaratory judgment context, Plaintiff’s failure to state a breach-of-contract claim creates a

jurisdictional defect. As we have previously explained:

Our [jurisdictional] inquiry is . . . whether, absent the availability of declaratory
relief, the instant case could nonetheless have been brought in federal court. To do
this, we must analyze the assumed coercive action by the declaratory judgment
defendant. Federal question jurisdiction exists in a declaratory judgment action if
the plaintiff has alleged facts in a well-pleaded complaint which demonstrate that
the defendant could file a coercive action arising under federal law.

Severe Records, LLC v. Rich, 658 F.3d 571, 581 (6th Cir. 2011) (quoting Stuart Weitzman, LLC v.

Microcomputer Res., Inc., 542 F.3d 859, 862 (11th Cir. 2008)) (internal citations and quotation

marks omitted); see also Sherwin-Williams Co. v. Ins. Co. of Pennsylvania, 105 F.3d 258, 261 (6th

Cir. 1997) (analyzing declaratory judgment action that was “premised on diversity jurisdiction”).

The “pay” provision is the coercive element of the parties’ contract that purportedly gives this

Court jurisdiction to consider Plaintiff’s argument that the “pay” provision is punitive. However,

as further discussed in this opinion, the “pay” provision has no coercive force—rendering the

majority’s analysis of the issue a nullity—because Plaintiff does not allege that the “pay” provision

is enforceable via specific performance. The majority’s refusal to consider the specific

performance issue based on Defendant’s “forfeiture”1 is therefore legally erroneous; “[s]ubject-

1In Footnote 3, the majority opinion also argues that there is no jurisdictional defect because pleading
problems do not “go to” to the issue of subject matter jurisdiction. But pleading problems commonly require dismissal

- 16 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

matter jurisdiction cannot be forfeited or waived and should be considered when fairly in doubt.”

Ashcroft v. Iqbal, 556 U.S. 662, 671, (2009); see also Fed. R. Civ. P. 12(h)(3) (“If the court

determines at any time that it lacks subject-matter jurisdiction, the court must dismiss the action.”).

The majority’s decision is also erroneous on a practical level because it will further delay

Plaintiff’s efficient breach. See Tri Cty. Wholesale Distributors, Inc. v. Labatt USA Operating

Co., LLC, 828 F.3d 421, 429 (6th Cir. 2016) (explaining that “contracting parties . . . have an

inherent right to breach a contract that is no longer advantageous, committing what economists

call an efficient breach”).

Specific performance is an exception to the usual remedy of damages for a breach-of-

contract claim involving the sale of goods. M.C.L. § 440.2716(1); see also Richardson v. Lamb,

235 N.W. 817, 818 (Mich. 1931) (“The general rule is that specific performance is not decreed

where the subject-matter of the contract is personal property.”); Groeb Farms, Inc. v. Alfred L.

Wolff, Inc., No. 08-CV-14624, 2009 WL 500816, at *7 (E.D. Mich. Feb. 27, 2009) (“Specific

performance is an equitable remedy that may be awarded where the legal remedy of damages is

impracticable.”). Under the UCC, which presumably applies to the parties’ contract for the sale

of polysilicon, courts may order specific performance “where the goods are unique or in other

for lack of subject matter jurisdiction. See Gentek Bldg. Prod., Inc. v. Sherwin-Williams Co., 491 F.3d 320, 330 (6th
Cir. 2007) (“Rule 12(b)(1) motions to dismiss for lack of subject-matter jurisdiction generally come in two varieties:
a facial attack or a factual attack. A facial attack on the subject-matter jurisdiction alleged in the complaint questions
merely the sufficiency of the pleading.” (citing Ohio Nat'l Life Ins. Co. v. United States, 922 F.2d 320, 325 (6th
Cir.1990)). The majority relies on several cases, none of which addresses the complexities of jurisdiction in the
context of a declaratory judgment action, to suggest that the specific performance issue in this case merely relates to
the merits of Plaintiff’s legal theory rather than to jurisdiction. But the problem with Plaintiff’s complaint is that there
will never be a controversy over whether the “pay” provision is a penalty as long as Plaintiff may simply breach the
contract and pay damages. As this Court has previously explained, “[i]n order to satisfy the ‘case or controversy’
requirement [of Article III jurisdiction], ‘a party seeking declaratory relief must allege facts to support a likelihood’
that it will incur [the alleged liability].” GenCorp, Inc. v. Olin Corp., 390 F.3d 433, 451 (6th Cir. 2004). The
majority’s argument is therefore inapposite; where, as here, the deficiency of a declaratory judgment complaint renders
the parties’ dispute purely hypothetical, the failure to state a claim is also a jurisdictional defect that the Court may
address sua sponte.

- 17 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

proper circumstances.” § 440.2716(1). The commentary to M.C.L. § 440.2716(1) elaborates on

the meaning of these terms as follows:

The test of uniqueness under this section must be made in terms of the total situation
which characterizes the contract. Output and requirements contracts involving a
particular or peculiarly available source or market present today the typical
commercial specific performance situation, as contrasted with contracts for the sale
of heirlooms or priceless works of art which were usually involved in the older
cases. However, uniqueness is not the sole basis of the remedy under this section
for the relief may also be granted “in other proper circumstances” and inability to
cover is strong evidence of “other proper circumstances”.

§ 440.2716 cmt. 2. Thus, “uniqueness” may refer to either (1) the uniqueness of a product’s source

or (2) the uniqueness of the product itself, and “other proper circumstances” may refer to

circumstances where, at the very least, (1) a breaching seller is the only available source or (2) a

breaching buyer is the only available customer. See id.

In the original context of take-or-pay agreements, specific performance of the buyer’s

“pay” obligation was an appropriate remedy for a breach-of-contract claim. Take-or-pay

provisions had their genesis in the unusual circumstances of the natural gas industry, wherein

pipeline–producer agreements are typically exclusive requirements contracts; the pipeline is often

the producer’s only customer, and the producer typically agrees to sell as much natural gas as the

pipeline is willing to buy. See Colorado Interstate Gas Co. v. Chemco, Inc., 854 P.2d 1232, 1234–

35 (Colo. 1993). Furthermore, as the Supreme Court of Colorado explained:

Long term contracts . . . are prevalent in the natural gas industry. Purchasers, such
as pipelines and industrial consumers, make substantial investments in
equipment for the transportation and consumption of natural gas. Long term supply
contracts ensure supply security for these purchasers during a time of shortage, such
as the shortage that occurred during the 1970’s . . . .

Early gas contracts had no minimum take requirement, permitting the pipeline to
choose the amount taken from each producer. Generally these contracts also
contained an exclusive dedication clause, prohibiting a producer from seeking
another purchaser for any available gas. Thus, pipelines were able to “shut-in” wells
when the demand for gas dropped, effectively utilizing the gas wells as storage

- 18 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

reservoirs for the benefit of the pipelines. Because the demand for natural gas is the
highest in the winter, many wells were shut-in during the summer, and producers
received no revenue from them. Being thus deprived of revenues, the producer “was
often unable to recover the substantial exploration, drilling, and operational costs
associated with wells.”

However, the regulation of natural gas sales in interstate markets placed artificial
ceilings on the price paid to producers of gas . . . . These ceilings limited the
negotiability of price in gas sales agreements. Thus, because supply was limited,
producers sought, and obtained, other economic benefits in their supply contracts.
Among the favorable provisions negotiated by producers in this artificial market is
the take-or-pay clause which, as an incentive for the producer’s contract, became a
part of the price pipelines were willing to pay to insure continued supply.

Id. (citations omitted). As this history demonstrates, take-or-pay agreements between pipelines

and producers will generally qualify for specific performance because a producer generally has

only a single customer (the pipeline) and cannot cover in the event that the pipeline does not

purchase gas for an extended period of time. See id.; § 440.2716 cmt. 2.

In this case, by contrast, Plaintiff does not allege that the contract involves unique goods,

a unique source of goods, or “other proper circumstances.” See § 440.2716. Plaintiff merely

assumes that specific performance is available, perhaps because the contract borrows the phrase

“take or pay” from the natural gas context. But the parties’ mere use of the phrase “take or pay”

does not render their contract enforceable via specific performance. See id.

Moreover, the circumstances of this case are not analogous to those of the natural gas

economy. Whereas a pipeline is typically a natural gas producer’s exclusive buyer, Plaintiff is one

of Defendant’s many customers; Defendant admits that it sells polysilicon to numerous “producers

of solar panels, including [Plaintiff.]” (Def. Br. at 2.) And whereas a natural gas producer is

typically a pipeline’s exclusive source of natural gas, Defendant is one of Plaintiff’s many

suppliers; Plaintiff explains that it “entered into supply contracts with solar polysilicon suppliers

other than [Defendant] between 2004 and 2007, including Wacker AG of Germany, Tokuyama of

- 19 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

Japan, Mitsubishi of Japan, and SGS/REC of Norway and the United States.” (R.144 at PageID

#5963.) Thus, unlike a natural gas producer, which has no ability to cover in the event that a

pipeline does not purchase natural gas, Defendant may cover in the event of Plaintiff’s breach by

selling polysilicon to another buyer. See Colorado Interstate Gas Co., 854 P.2d at 1234–35. The

parties’ agreement appears to be a run-of-the-mill contract for the sale of goods—a far cry from

the exclusive requirements contracts seen in the natural gas economy. As such, Defendant is not

entitled to specific performance in the event of Plaintiff’s breach—at least not for the reasons

applicable to pipeline–producer contracts. See § 440.2716; Colorado Interstate Gas Co., 854 P.2d

at 1234–35.

Plaintiff’s failure to address specific performance is fatal to its declaratory judgment claim.

Plaintiff seeks a declaratory judgment that a court could not order Plaintiff to perform its “pay”

obligation. See Severe Records, LLC v. Rich, 658 F.3d 571, 581 (6th Cir. 2011). But Plaintiff’s

complaint is not so specific. Rather, Plaintiff’s complaint merely asks the district court for a

declaration that the “pay” obligation is “unenforceable.” (Pl. Br. 13.) The term “unenforceable”

is ambiguous; on one hand, the “pay” obligation is probably “unenforceable” via specific

performance, but on the other hand, the “pay” obligation is probably enforceable insofar as

Defendant may seek damages if Plaintiff fails to either “take” or “pay.” Defendant responds to

this ambiguity with an ambiguity of its own: Defendant asserts that the contract is “fully valid and

enforceable under Michigan law,” (Def. Br. 20), because Plaintiff’s “pay” option is one of two

“bargained-for performance obligations.” (Def. Br. 22.) By asserting that the contract is

“enforceable,” Defendant answers Plaintiff’s complaint without addressing the specific

performance issue that motivated Plaintiff to file suit.

- 20 -
Case No. 17-2276
Hemlock Semiconductor Corp. v. Kyocera Corp.

If Plaintiff had properly pleaded its complaint by alleging that the “pay” obligation is

enforceable via specific performance, then Defendant’s statement that the “pay” obligation is

“enforceable” would be a plainly inadequate answer. Even if a court will “enforce” Plaintiff’s

obligation to either “take or pay,” the question remains whether a court could order Plaintiff to

“pay” if Plaintiff refuses to “take.” Michigan courts have not specifically addressed the remedy

for breach of an alternative performance contract, but “the measure of damages for the breach of

such a contract is generally considered to be the value of the alternative least onerous to the

defendant”—not specific performance. 25 Williston on Contracts § 66:106 (4th ed.). Some

jurisdictions make an exception to this rule when “one of the alternatives is to pay a certain sum

of money,” id., but this exception reflects an analytical lapse; the proper term for a payment option

that is (1) included in a contract as an alternative to performing a contractual duty, and (2) subject

to specific performance, is a liquidated damages clause. Specific performance is therefore a

defining feature of a liquidated damages clause and is a necessary element of any claim calling for

a liquidated-damages analysis. Because Plaintiff asked the district court to apply a liquidated

damages analysis without alleging that the “pay” obligation is enforceable via specific

performance, the district court properly dismissed Plaintiff’s claim.

Because the majority finds Plaintiff’s deficient complaint to be adequate, I respectfully

dissent.

- 21 -