The US Court of Appeals for the Federal Circuit reversed a Patent Trial & Appeal Board decision finding the challenged patent claims not obvious over the prior art. The Court found that the Board, after concluding that no claim construction was required, implicitly construed the claim limitation at issue and did so erroneously. Google LLC v. EcoFactor, Case Nos. 22-1750; -1767 (Fed. Cir. Feb. 7, 2024) (Reyna, Taranto, Stark, JJ.)

Google filed a petition for inter partes review (IPR) challenging claims of an EcoFactor patent related to dynamic climate control systems that factor outside weather conditions and thermal conditions inside the home to balance comfort and energy savings. The challenged claims define a method for reducing the cycling time of a climate control system involving “retrieving a target time at which [the] structure [(e.g., a house)] is desired to reach a target temperature.” The challenged method claims recite a step of “determining a first time prior to said target time at which [the] climate control system should turn on to reach the target temperature by the target time.” The relevant claim limitation reads:

[D]etermining a first time prior to said target time at which said climate control system should turn on to reach the target temperature by the target time based at least in part on [i] said one or more thermal performance values of said structure, [ii] said performance characteristic of said climate control system, [iii] said first internal temperature, [iv] said first external temperature, and [v] the forecasted temperature.

During the IPR proceedings, the parties disputed whether a prior art reference disclosed a method involving determining a first time prior to the target time based on a first internal temperature. Google argued that the prior art taught a calculation of a first time prior to the target time based on thermal performance values (input [i]) calculated from internal temperature values (input [iii]). EcoFactor argued that each of the inputs in the claim limitation was a distinct value not dependent on or calculated from any other input. Based on the claim language, the Board determined that claim construction was unnecessary and concluded that inputs [i] – [v] of the relevant claim limitation were separate inputs using different data. The Board concluded that Google had not shown that the challenged claims were unpatentable, reasoning that Google’s theory of obviousness relied on a single input as the basis for both input [i] and input [iii].

Google appealed. Google argued that although the Board stated that no construction was necessary, it incorrectly construed the claim limitation to require five discrete inputs.

The Federal Circuit agreed with Google, finding that the Board’s assessment of the claim limitation implicitly established the claim scope by requiring inputs [i] – [v] to be completely separate. The Court reasoned that the plain claim language did not provide any indication that none of the listed inputs could be based on any other input(s). Imputing this requirement into the limitation was therefore an act of claim construction.

The Federal Circuit then addressed whether the Board’s construction was erroneous. Google argued that the construction was erroneous because it violated the Administrative Procedure Act (APA), and that limitations introduced by the construction were not supported by the intrinsic record or case law.

Citing its 2021 decision in Qualcomm v. Intel, the Federal Circuit disagreed with Google as to the Board’s noncompliance with the APA. As it held in Qualcomm, the Court reasoned that adoption of a claim construction not advanced by either party does not amount to a violation of the APA. The Court noted that it would have been a violation had the Board, without notice and opportunity for response, changed theories by adopting a construction that neither party requested or anticipated. The Court found that such was not the case here, because the Board’s implicit construction addressed the core issue in dispute (i.e., whether the five inputs were distinct or could be based on common data).

In holding that the Board’s implicit construction was erroneous, the Federal Circuit looked to the intrinsic record: the patent specification and the claims themselves. The Court reiterated that the claims did not contain any proscriptive language excluding methods where an input value is used for two or more of the inputs. The Court also noted disclosure in the specification that contemplated embodiments where an input is calculated using at least one other claimed input. Since the Board’s construction would read such a disclosed embodiment out of the claims, the construction was erroneous. The Court remanded the case back to the Board for further proceedings under Google’s construction of the contested limitation, which “require[s] that each of the five inputs be used at some point during the determination.”

Practice Note: Although separately listed claim limitations may be presumed to define separate and distinct structures, this presumption is rebuttable in view of the intrinsic record and context of the claimed invention. Patent applicants seeking to define claim features as separate and distinct limitations should include clear language to that effect.