I’ve been busy the last several days reviewing the page proofs for my forthcoming book, Wrongful Patent Assertion: A Comparative Law and Economics Analysis (Oxford Univ. Press 2026), which is due out in late spring or early summer. Now that I have some time to resume blogging, I’ll start by discussing a short (but precedential) Federal Circuit decision from last Friday, Exafer Ltd. v. Microsoft Corp. (opinion by Chief Judge Moore, joined by Judges Taranto and Stoll). It presents an interesting question relating to damages calculation.
Exafer owns two patents in suit relating to “systems and methods for optimizing communication paths between virtual network devices by controlling data forwarding rules at intelligent switches” (p.2). Exafer claims that “Microsoft’s Azure Platform, and specifically the Azure Smart Network Interface Cards (SmartNICs) and Virtual Filtering Platform (VFP) Fastpath technology (Accused Features),” infringe the two patents. Exafer’s damages expert Mr. Blok was prepared to present an opinion concerning the hypothetical royalty the parties would have agreed to ex ante, using as the royalty base the value of certain noninfringing virtual machines (VMs). The theory is that Microsoft’s use of the patented technology enabled “Microsoft to reduce the central processing unit (CPU) usage in Azure servers, freeing up CPU cores to host additional VMs” (p.6); and that the value to Microsoft of using the patented technology is therefore the revenue derived from hosting those additional VMs. The district court agreed with Microsoft’s argument that it was impermissible to use the noninfringing VMs as the royalty base, but the Federal Circuit reverses and remands.
I think the Federal Circuit got it right, though I can understand the appeal of Microsoft’s argument that the royalty should not be calculated using the value of some other, noninfringing product. (Going back in time, one might perceive a similar perspective in cases such as Zenith Radio Corp. v. Hazeltine Rsch., Inc., 395 U.S. 100, 135 (1969), which held that setting royalty rates on the basis of the licensee’s sales of unpatented products constituted patent misuse (while also holding that, if the licensee is not coerced into taking unwanted patents, but instead agrees for convenience to take a portfolio of patents, the arrangement is not misuse).) The premise that the royalty must be related to the use of the patented technology is of course correct, but I think the Federal Circuit is right in finding a sufficient causal connection between that use and the increase in the number of VMs Microsoft can host; and as long as the revenue derived from that increase is a type of benefit that Microsoft would have anticipated ex ante, it stands to reason that the amount Microsoft would have been willing to pay ex ante would have reflected that expected benefit. This is essentially the court’s view, as I read it:
[Exafer’s technical expert] Dr. Congdon opined that the network optimization and efficiency improvements achieved by the claimed inventions “would translate to, among other benefits, the ability to operate more virtual machines on a single CPU or host (i.e., increasing virtual machine density). Accordingly, by increasing virtual machine density, Microsoft would be able to sell more virtual machines without the need for additional network infrastructure.” . . . Mr. Blok’s VM-hour royalty base captured this incremental benefit of being able to offer additional VMs due to operation of the Accused Features within the Azure Platform. . . . This methodology is tethered to the patented invention and does not expand Exafer’s patent monopoly to unpatented technology. Mr. Blok’s testimony therefore satisfies the admissibility standards of Rule 702 (p.8).
Put another way, as long as the additional revenue associated with hosting more VMs was, ex ante, a foreseeable consequence of the use of the patents in suit, a willing licensee would have taken that added benefit into account in determining how much it was willing to pay for a license. By contrast, requiring that the royalty be limited to the immediate benefit of the use (perhaps the cost savings associated with reduced CPU usage for purposes of powering the Azure platform, without any consideration of the next-best use of those otherwise idle CPUs) strikes me as a formalistic constraint lacking in economic substance. That said, if the actual benefit derived ex post from freeing up some of the CPUs were of a type that would not have been foreseeable ex ante, then it should be excluded from consideration; but that is not my understanding of the facts here.
The case does make me think about the connections between the hypothetical bargain construct; the situations in which it might be rational for courts to make use of ex post information (which Norman Siebrasse and I wrote about a few years back); and the doctrine of proximate cause, which limits damages to those that are, inter alia, foreseeable. I may have more to say about this in a future post.
No comments:
Post a Comment