Caremark works as a two-step test. The plaintiffs lost at the second step. Caremark requires a board to establish an information system and monitor it so that reports on legal compliance and other key risks actually arrive, and once that system exists and is being watched, the court said, the baseline duty has been fulfilled [7]. What the board then does with what it learns is an evaluation of business risk, which the court called a quintessential board function entitled to deference under the business judgment rule [8]. That framing moves the contest from what the board knew to whether it deliberately chose not to act, and the court found no reasonable inference of deliberate disregard or purposeful lawbreaking, noting that the directors believed they were reasonably performing their duties [9].
Where this bites is the handling of red flags. According to the Fried Frank memorandum, the court took a narrow view of what qualifies, and treated the regular arrival of reports about ongoing safety risks, including violations of law and regulation persisting over several years, as saying something about the company's information system rather than establishing bad faith [10]. A stream of adverse compliance reporting is the ordinary raw material of these complaints. If years of it can be read as the system working, the pleading path that Fried Frank says widened in recent years gets narrower [12].
The board-deck version of this is that documented attention protects directors. Most of the record supports that, since the firm's list of distinguishing facts leads with a years-long effort to overhaul safety systems and culture after the two earlier crashes, and with the absence of any evidence that the board decided not to comply or to delay addressing noncompliance [13]. The list has a third item, and it is not about board process at all. The 2024 failure was dramatic but hurt eight people slightly [5], and Fried Frank names that limited harm as one of the reasons the case is distinguishable [13]. Hundreds of people died in the 2018 and 2019 crashes [4]. Whether the same remediation record would have carried a board through a fatal outcome is a question the decision leaves open.
One pleading-stage opinion is not doctrine, and the memorandum hedges in that same direction, writing that the decision appears to reflect a narrower view and that its analysis and tone suggest a return to the older judicial emphasis on dismissal [12][14]. That caution applies to the decade-long trend, but it does not change the calculus for anyone deciding this quarter whether to fund a derivative complaint or how hard to fight a books-and-records demand, since that calculus turns on the most recent Chancery reading of red flags.
Sequencing matters here more than the holding. The record that carried Boeing's board was assembled in the five to six years between the earlier crashes and the 2024 incident [15], and the dismissal landed roughly two years after that incident [16]. Nothing built after a complaint arrives can do that work, which is the practical cost of the ruling for a board that has been treating oversight documentation as a litigation response rather than a standing obligation.
