That reaction is worth understanding, because the gap between what the headline numbers say and what the market concluded points at something specific about how defense contractors actually make money.
Start with what genuinely went well
The demand picture is not ambiguous, and nothing in the skeptical case disputes it.
Second quarter awards included $7.6 billion for Sentinel, $4.3 billion for restricted programs, $1.0 billion for F-35, $0.8 billion for the Glide Phase Interceptor, and $0.7 billion for Multi-role Electronically Scanned Array. All four segments grew. Sales rose 5% to $10.9 billion.
Aeronautics Systems was the standout, with sales up 13%, or $405 million, on higher B-21 volume, a $106 million increase on the E-130J TACAMO program, and higher volume on B-2, F-35 and E-2. Operating margin there reached 10.3%, and management raised the segment's full-year sales estimate to roughly $14 billion.
Kathy Warden's framing on the call was that the company is fully aligned with U.S. government priorities and sees significant opportunity and increased demand across missiles, missile defense, autonomous aircraft and national security space. Given the award figures, that is a defensible description rather than a promotional one.
Then look at what the beat was made of
Reported EPS of $7.68 cleared a consensus near $6.82, which sounds enormous until the comparisons are cleaned up.
The prior-year quarter reported $8.15, but that included a benefit of $150 million, or $1.04 per diluted share, from the divestiture of the training services business. Strip it out and the year-ago figure was about $7.11, making this quarter roughly 8% growth rather than the double-digit surprise the raw comparison implies.
The prior year also carried a $76 million favorable estimate-at-completion adjustment on Sentinel, and this quarter's earnings per share benefited from a lower effective tax rate.
None of that is unusual or improper. Defense earnings are routinely shaped by contract adjustments, divestitures and tax items. It does mean the operational beat was narrower than the headline, and analysts covering this sector unpack those items as a matter of course.
The number that did not move
Here is the detail that probably explains the share reaction better than anything else.
Full-year EPS guidance went up by $1.20, to a range of $28.60 to $29.10. Revenue guidance rose $250 million.
Adjusted free cash flow guidance stayed exactly where it was, at $3.1 billion to $3.5 billion.
Earnings up, cash unchanged. For a business that investors hold substantially for its cash generation and capital return, that divergence matters more than the EPS line, because a lower tax rate flows to earnings without flowing to operating cash in the same way, and because growth in this industry consumes capital before it produces it.
Which brings up the second half of the same problem.
Growth here is paid for upfront
Capital expenditure runs $1.85 billion this year, with CFO John Greene indicating capex around 4.5% of sales in 2027 and 2028 as the company supports the B-21 production ramp. Northrop won an Air Force contract earlier this year to expand B-21 manufacturing capacity by 25%, with first delivery from the expanded capacity planned for 2027.
Building bomber production capacity is not a software gross margin story. It requires tooling, facilities, supply chain qualification and workforce, all spent years before the associated revenue arrives, and all of it landing in the cash flow statement now.
That is the honest cost of a 1.84 book-to-bill. Converting a record backlog into deliveries requires physical capacity that has to be financed first. The backlog is real. So is the bill for fulfilling it.
The part that decides everything
Management also acknowledged cost pressures on two programs that weighed on segment margins.
That phrase is doing more work than its length suggests, because cost performance is the entire profit mechanism in this industry. A defense contractor does not earn money by winning a contract. It earns money by executing the contract for less than the negotiated cost, or at least in line with it. On fixed-price work, overruns come straight out of the contractor. On cost-plus-incentive work, they reduce the fee. Booking $20 billion of awards tells you what the revenue will be. It tells you nothing about the margin.
This is why Sentinel sits at the center of the investment case in both directions. It is a major growth driver, contributing $7.6 billion of awards in this quarter alone, and it is also the program with a documented history of cost overruns and schedule delays that triggered a formal program restructuring. A ground-based nuclear deterrent replacement is about as complex as procurement gets, and the same scale that makes it valuable makes an execution problem expensive.
Note also that the prior-year quarter included a favorable Sentinel adjustment. Estimate-at-completion revisions move in both directions, and a program with this history will keep producing them.
What actually resolves this
Three things, and none of them is the backlog.
Segment margin trajectory, particularly whether the two pressured programs are contained or representative. Aeronautics at 10.3% with full-year margins guided to the mid-to-high 9% range implies some moderation ahead, and the reasons for it will matter.
Free cash flow against the reaffirmed $3.1 billion to $3.5 billion range, which is the number that was conspicuously not raised alongside earnings.
And Sentinel estimate-at-completion adjustments over the next several quarters, which are the clearest available signal on whether the largest program in the backlog is being executed to plan.
Greene was notably careful on the call about the years beyond 2026, saying he would refrain from giving too many details on the outer years while the strategic plan is finalized, though he added that Sentinel, B-21 and weapons demand should let the company accelerate revenue relative to its history.
That is probably the fair summary of the whole quarter. The demand is as good as it has been in a generation, the revenue growth is coming, and the open question is what it costs to deliver. Backlog is a promise to do work. Whether that work is profitable gets decided later, on factory floors, and this quarter offered a small reminder that it is not automatic.