AMENDED OPINION '
In our original decision in this case,
1
we reversed the judgment of the trial court, which implemented the jury’s general verdict in favor of plaintiff and overruled defendant’s motion for judgment notwithstanding the verdict (“j.n.o.v.”). We did so upon our conclusion that each of three alternative legal theories of anticompetitive conduct that were presented to the jury were erroneous as a matter of law. We then vacated the award of damages and the injunction issued by the trial court and remanded with instructions to enter judgment in favor of the defendant. The United States Supreme Court directed us to reconsider our original decision in light of
Eastman Kodak Co. v. Image Technical Serv. Inc.,
— U.S. -,
I.
The opinion only briefly revisits the relevant facts. 2 Defendant Prime Computer, Inc. (“Prime”) manufactures and markets computer systems, and provides maintenance services for those systems. Of significance to this lawsuit is one. of its hardware lines, the “50 Series” minicomputer, and one of its applications software products, the so-called Computer-Aided Design/Computer-Aided Manufacturing system (“CAD/CAM”), which can be used with the 50 Series minicomputers.
Ford Motor Company (“Ford”) created and owns a CAD/CAM software design program for use in designing automobiles. Ford’s CAD/CAM program is called Product Design Graphic System (“PDGS”). Ford frequently revises the software program, and requires the automotive design companies with which it does business to use the most recent version of Ford’s PDGS in order to facilitate the transmission of design specifications through CAD/CAM software.
Ford licenses defendant Prime as the exclusive distributor of Ford’s version of PDGS under a year-to-year contract. Ford’s version of PDGS runs only in Prime’s 50 Series minicomputers, but can be translated to other systems at a higher cost.
Prime also distributes software support (i.e., revisions, modifications, updates, and support services) for PDGS software. Prime offers this software support to Ford’s design companies as part of a package that includes hardware maintenance on the Prime 50 Series minicomputers. PDGS revisions may be purchased separately from the hardware maintenance, but only at a prohibitive expense. In'contrast, the general contract to purchase PDGS does not contain a hardware maintenance requirement.
Plaintiff Virtual Maintenance, Inc. (‘Virtual”) brought this antitrust action after unsuccessfully attempting to enter into hardware maintenance contracts with owners of Prime 50 Series computers. Virtual contended that Prime’s package constituted an illegal tying arrangement in violation of § 1 of the Sherman Act, 15 U.S.C. § 1, by conditioning the purchase of software support required by Ford to hardware maintenance from Prime (the tying product).
Virtual’s case was presented to the jury on three alternative theories of liability based on the alleged tie of hardware maintenance services to PDGS updates: (1) a per se claim based on a tying market of all CAD/CAM software; (2) a per se claim based on a tying product market of Ford-required PDGS; and (3) a rule of reason claim alleging that the clause in Prime’s software maintenance package contract requiring the customer to use Prime’s hardware maintenance services created unreasonable and anticompetitive effects in the market for maintaining Prime’s 50 Series computer systems. The jury returned a general verdict in favor of Virtual. The district court denied Prime’s motion for j.n.o.v.' and entered judgment in favor of Virtual.
On appeal, we rejected each of Virtual’s theories, and reversed the judgment of the district court with instructions to enter judgment in favor of Prime. In accordance with the Supreme Court’s directive, we will consider whether any of these prior rulings are impacted by Eastman Kodak.
II.
A.
In
Eastman Kodak,
the antitrust plaintiffs were a group of independent service organizations (ISO’s) that had been servicing Kodak copying and micrographie equipment since the early. 1980s. They brought suit after Kodak began restricting the. sale of replacement parts for its photocopiers and micrographic equipment to only those buyers who also purchased Kodak service or repaired their own machines. Kodak equipment is unique; its parts are not compatible with its competitors’ machines. Because of
*663
Kodak’s restrictive policy, the ISO’s were unable to obtain suitable parts and many were forced out of business. The plaintiffs also offered evidence that some customers who preferred the plaintiffs’ service were forced, as a result of Kodak’s practice, to switch to Kodak’s service. Plaintiffs alleged, inter alia, that Kodak had tied- the sale of service to the sale of parts in violation of section 1 of the Sherman Act.
Id.
at -,
The district court granted summary judgment for Kodak. A divided panel of the Ninth Circuit reversed, finding a genuine issue of material fact as to whether Kodak had sufficient economic power in the tying product market (parts) to restrict competition appreciably in the tied product market (service).
Id.
at -,
Before the Supreme Court, Kodak did not offer any actual data on the proposed markets, but rather urged the adoption of a substantive legal rule that interbrand competition foreclosed finding of monopoly power in derivative aftermarkets as a matter of law.
Id.
at -,
The Court found that the plaintiffs had offered a “forceful reason” why Kodak’s theory might not accurately explain the behavior of the primary and derivative markets for complex and durable goods: “the existence of significant information and switching costs.”
Id.
at -,
Furthermore, the cost of switching products may be high, and consumers who have already purchased the equipment will tolerate some level of service-price increase before switching equipment brands.
Id.
at -,
Also critical to the Court’s conclusion that summary judgment on the ISO’s claims was improper was evidence presented by the plaintiffs that certain parts were available exclusively through Kodak, that Kodak had control over the availability of parts it didn’t manufacture, and that Kodak’s control over its parts market had excluded service competition, boosted service prices, and forced unwilling consumption of Kodak service when plaintiffs’.service would have been preferred.
Id.
at -,
B.
1.
Eastman Kodak
did not involve a rule of reason theory of liability, and we see
*664
no reason to . alter our conclusion that there was insufficient evidence to support a jury-verdict under the rule of reason. Even with Ford-required software support as the tying product market, Virtual failed to demonstrate a substantial threat that Prime would acquire market power in the most narrowly defined tied product market of hardware maintenance of -Prime 50 Series systems.
5
See Virtual I,
2.
Under Virtual’s first per se claim
6
, the district court defined the relevant market as “the sale of software revisions and support for the CAD/CAM industry in general.” We found no error in the legal adequacy of the district court’s definition, but concluded that Prime possessed at most an 11% share of the market, which is insufficient as a matter of law to confer market power.
Virtual I,
957 at 1325 (citing
Jefferson Parish Hosp. Dist. No. 2 v. Hyde,
3.
Our- ruling regarding Virtual’s second per se claim requires much closer scrutiny. The other -market that Virtual urged was defined as “the sale of software revisions and support of software necessary to do business with Ford Motor Company.”
Virtual I,
In response to Virtual’s contention that Ford is not the only customer of Prime’s products, but that all of Ford’s design suppliers made up the customer market for “Ford-required CAD/CAM,” we stated:
Virtual seeks to distinguish these cases by pointing out that Ford is not the only customer of Prime’s products; rather, all of Ford’s design suppliers make up the customer market for “Ford-required CAD/ CAM.” This ignores the fact that Ford requires its suppliers to purchase the software updates for Ford’s benefit. Ford is ultimately the single consumer of its specialized design software because Ford’s requirements define the demand for the software and the updates. But defining the *665 market by Ford’s requirements creates the appearance of market power based only upon the demand side of the market. Defining a market, or “submarket,” on the basis of demand considerations alone is erroneous because such an approach fails to consider the supply side of the market. Philip E. Areeda & Herbert Hovenkamp, Antitrust Law, ¶ 518.1g at 471 & n. 26 (Supp.1990) (citing United States v. Central State Bank,817 F.2d 22 (6th Cir.1987)). The relevant product market cannot be determined without considering the cross-elasticity of supply.
Id.
Virtual countered that Prime has no supply side competition because of its exclusive license for PDGS. We rejected this argument as “confus[ing] the tying product (software support for PDGS) with the interbrand market relevant for antitrust analysis,” id, 8 and ruled that “the relevant tying market is comprised of all CAD/CAM software reasonably interchangeable with PDGS.” Id. Critical to this analysis was the view that PDGS software support is an intrabrand “submark-et”:
Prime has market power in the trivial sense that no one else makes PDGS. But true market power — power sufficient to change and sustain anticompetitive prices — cannot be inferred from this because were Prime to charge exorbitant prices for its software support, its customers would simply switch to some other manufacturer of PDGS-type software. Prime’s lack of market power over the general market for CAD/CAM software thus prevents Prime from controlling the “submarket” for PDGS software.
Id.
Virtual responds that Ford and its design suppliers cannot switch to a new supplier of software support because they are “loeked-in” to Prime as the sole supplier due to their substantial investments in Prime 50 Series computers and other hardware. While Ford might believe it is “locked-in,” this is due in large part to its own decision to purchase Prime’s software •and invest in Prime’s computer systems. *666 But a customer’s initial purchase of a particular manufacturer’s product does not justify a limited market definition. Defining the market by customer demand after the customer has chosen a single supplier fails to take into account that the supplier (here Prime) must compete with other similar suppliers to be designated the sole source in the first place.
Id.
Under Eastman Kodak, our rejection of Virtual’s second per se claim based on a tying product of Ford-required software support was misguided. That Ford had many competitors to choose from when it made its initial decision to grant the exclusive license to Prime cannot, after Eastman Kodak, preclude as a matter of law Virtual’s proposed theory of market power because it ignores information costs. It follows that our rejection of Virtual’s “lock-in” argument, on the basis of an interbrand competitive market in the initial purchase of Prime’s software package and Prime 50 Series minicomputers, was also in error.
Eastman Kodak
also requires us to rethink our characterization' of the tying product market here as merely the preference of one customer. Unlike
Eastman Kodak,
the initial decision to purchase Prime equipment and software was made by a single consumer, Ford. However, by shifting the focus to the derivative aftermarket of software support, there is not a single consumer, but numerous automotive design companies doing business with Ford. In contrast with
International Logistics,
The similarities between this case and
Eastman Kodak
are apparent. As in
Eastman Kodak,
the alleged tie is not between equipment and parts, where interbrand competition would defeat market power and a per se tying claim. Rather, the alleged tie is in the derivative aftermarkets. Like Kodak, Prime is able to exercise control over the sale of software support because of its exclusive distribution license from Ford, and Ford’s requirement that its automotive design suppliers use the most current version of Prime’s software support. Thus, it can be argued that Prime enjoys a significant advantage in the Ford-required software support market by virtue of Ford’s license and the requirement Ford places on the automotive design companies to use the most current version of PDGS. In other words, there is evidence to make the argument that Prime’s tying arrangement bears that “essential characteristic” of an illegal tying arrangement, the ability to exploit control over the tying product to force the buyer to purchase an unwanted tied product.
Jefferson Parish,
Like the ISO’s in Eastman Kodak, Virtual presented evidence of price manipulation. An expert for Virtual testified that although Prime offered software support for sale separately, repurchase of software to obtain updates would cost as much as 900% more than if purchased in the software supporVhard-ware maintenance package. There was also evidence that Prime does not treat all customers equally. Prime apparently allows one customer, Ford Aerospace, to continue to *667 receive software support without purchasing hardware maintenance from Prime. Virtual also offered proof that customers would have preferred Virtual’s service to Prime’s, but were effectively precluded by the tie.
Virtual offered expert testimony concerning lock-in and switching costs. Virtual’s industry experts testified that customers were “locked-in” to the hardware maintenance by the substantial cost incurred for hardware, maintenance, and training, most of which would be substantially worthless if the customer switched to another manufacturer’s system. A Ford employee, called by Prime, testified that Ford itself felt “locked-in” to Prime, stating that Ford could not change from Prime’s computer system and remain economically viable in the automotive industry. Virtual’s expert also opined that Prime could substantially raise its maintenance prices before customers would abandon their investment.
Finally, Virtual made a showing that a “not insubstantial” amount of commerce was affected in that it stood to lose significant profits over a five-year period. Thus, under Eastman Kodak, we conclude that Virtual’s second definition of tying product market was not improper.
C.
Eastman Kodak
dictates that we vacate our holding as to Virtual’s second definition of tying product market on a per se tying claim. All other aspects of our holdings in
Virtual I
are reaffirmed and reinstated. Because the general jury verdict in the first trial provides no indication of the jury’s reliance on this theory in support of the original verdict,'we are obligated to reverse the verdict and remand for a new trial.
See Sunkist Growers, Inc. v. Winckler & Smith Citrus Products Co.,
ADDENDUM
Both Virtual and Prime petitioned for rehearing in this matter. We have reviewed our decision carefully in light of these petitions and responses thereto. Prime’s requests for clarification have been incorporated in the corrected opinion; Prime’s petition for rehearing is otherwise denied. We decline Virtual’s request to apply
Griffin v. United States,
— U.S. -,
Notes
.
Virtual Maintenance, Inc. v. Prime Computer, Inc.,
. For a more detailed history of the case, see id.
. Plaintiffs also claimed that Kodak had unlawfully monopolized and attempted to monopolize the sale of service for Kodak machines in violation of § 2 of the Sherman Act.
. The Supreme Court made clear that
Matsushita Elec. Indus. Co. v. Zenith Radio Corp.,
. Virtual claimed that a jury could find that Prime possessed almost total control over the Prime 50 Series hardware maintenance market.
Virtual I,
. As noted in
Virtual I,
the elements of a per se tying claim are as follows: “(1) There must be a tying arrangement between two distinct products or services; (2) the seller must have sufficient economic power in the tying market to restrain appreciably competition in the tied product market; and (3) the amount of commerce affected must be not insubstantial.”
Virtual I,
. We stated: “An 11% market share, standing alone, is insufficient as a matter of law to confer the 'great market power, evidenced by an exceptional demand for the tying product’ necessary for a per se illegal tie.”
Virtual I,
. We observed that this court previously had refused to ignore interbrand competition by limiting the product market to one manufacturer's product, citing
A.I. Root Co. v. Computer/Dynamics, Inc.,
Kingsport Motors,
. In rejecting Virtual's contention that PDGS constitutes a relevant “submarket,” we relied, in part, on the proposition in
A.I. Root Co. v. Computer/Dynamics, Inc.,
. This view is not inconsistent with dicta in
International Logistics,
