Inside RTX's Golden Dome Moment: How Defence's Capital Cycle Just Reset

Two years ago, the RTX investment thesis was a turnaround story. The Pratt & Whitney GTF powder metal contamination, disclosed in mid-2023, triggered a fleet-wide inspection programme that required RTX to absorb roughly $5 billion of customer-recovery payments and accelerated maintenance costs through 2024-2026. Free cash flow guidance was repeatedly cut. The dividend was sustained but the buyback was paused. The stock fell from $108 in early 2023 to below $69 by autumn 2023.

The recovery has been faster than consensus expected. Pratt & Whitney's GTF customer-recovery accruals are now largely behind the income statement. Operating margin in the engine segment expanded from a depressed 6.8% in FY2024 to 9.4% in FY2025, with another 100-150 basis points modelled for FY2026. Collins Aerospace, the third leg of the franchise, continues to compound revenue and margin in the high single digits. The Raytheon segment, the focus of the Golden Dome story, has accelerated.

The new dynamic is capacity, not demand. RTX's missile production lines, in particular for Patriot, NASAMS, and the SM-3/SM-6 family, are running at capacity. Lead times have stretched. The capex programme to expand missile-component production capacity has been outlined at $1.2-1.5 billion over the 2026-2028 window. That spend is not in the consensus FCF model yet. The Capital Desk view is that this capex is value-accretive, but it changes the near-term cash conversion math the consensus has been pricing.

Historically, when defence primes hit capacity walls during a budgetary up-cycle, the operating margin expansion gets larger and FCF growth gets faster, with a 12-18 month lag from the start of the capex programme. The 2002-2007 missile-defence buildup is the textbook case. The pattern is now repeating.

The Golden Dome programme, formalised in early April, contemplates a layered missile-defence architecture extending across continental US airspace, with space-based sensor and intercept components. The headline programme cost was sized at $175 billion over a decade, with the first $25-30 billion authorised through FY2027. RTX is named in three architecture work-streams: ground-based missile interceptors, space-based sensors, and command-and-control integration. Lockheed Martin and SpaceX cover the remaining work-streams with overlap on integration.

For RTX, conservatively, the Golden Dome contract flow is worth $40-55 billion over the decade, weighted to the back half as the architecture phase closes and production phase opens. That is roughly equivalent to one full year of Raytheon segment revenue, distributed over ten years, on incremental margins that historically run 200-300 basis points above the segment average. The earnings impact is meaningful but spread out. The capital allocation impact is more concentrated.

Management has signalled, on the Q1 call, that the buyback is reactivating now that the GTF cash drag has been absorbed. The FY2026 buyback authorisation runs at $5 billion (versus near-zero in FY2024-FY2025). At today's $174 print, that is roughly 2.9% of the share count repurchased over twelve months. That is a meaningful per-share earnings tailwind that is not fully reflected in consensus FY2027 EPS estimates.

Dividend coverage has also improved. The $2.72 annualised dividend equates to roughly $3.6 billion of cash outflow against $7.94 billion of FCF, a 45% payout. With FCF modelled at $9-10 billion in FY2026, the payout ratio drops to 36-40%, opening room for above-historical dividend growth. We expect a 7-9% dividend increase at the next declaration cycle.

The overall picture: rising backlog, expanding margins, capex absorbed, buyback restarted, dividend coverage improving. That is the canonical Capital Desk setup for a defence prime in the early innings of a budgetary super-cycle.

Operating income climbed from $6.67 billion in FY2024 to $8.89 billion in FY2025, a 33% increase on the 9.7% revenue line. That is operating leverage in its textbook form: incremental revenue at higher incremental margin, driven by the GTF customer-recovery accruals rolling off and the Raytheon segment volume conversion. The consolidated operating margin moved from 8.3% to 10.0%, with another 80-120 basis points of expansion modelled for FY2026.

Return on invested capital recovered from a depressed 6.5% in FY2024 to roughly 9.0% on the FY2025 base, climbing back through the cost of capital. The Capital Desk anchors heavily on this metric for defence primes. ROIC above the cost of capital is the necessary condition for capital allocation to compound shareholder value rather than simply consume it. RTX's pre-2023 ROIC ran at 11-13%. The recovery path back to that range is now visible.

Net debt sits at $32.1 billion against EBITDA of approximately $13.8 billion, a 2.3x leverage multiple. That is comfortably within the investment-grade range and below management's 2.5-3.0x stated comfort zone. The balance sheet flexibility allows the buyback restart without compromising the dividend coverage or the capex programme. We model net debt holding flat through FY2027 as cash generation funds both the buyback and capex out of operating cash.

Dividend per share is $2.72, up from $2.36 two years ago. The payout ratio at 45% of FCF is the lowest it has been since the United Technologies merger close. The Capital Desk model points to dividend growth in the 7-9% range over the next three years, well above the dividend cluster average of 5-6% across the prime peer set.

Lockheed Martin trades at roughly 18x forward earnings on a flatter growth profile and a more concentrated programme exposure. Northrop Grumman sits at 19x forward earnings with the strongest space exposure but lighter near-term backlog growth. General Dynamics at 17x forward, with a different mix tilted toward Marine Systems and Combat Systems. RTX at 25.5x forward earnings is the highest multiple in the prime cluster, but it is also the franchise with the most concentrated FCF inflection.

The multiple premium is justified. Free cash flow growth at RTX is modelled at 13-15% over the next three years, against 5-7% for Lockheed and 7-9% for Northrop. Backlog growth is comparable across primes, but the conversion rate at RTX is faster because of the missile-production capacity expansion now underway. The Pratt & Whitney recovery, which is essentially a known cash unwinding rather than a cyclical bet, adds an additional 200-300 basis points of FCF growth that the peer set does not have.

The Raytheon segment is the differentiator. With Golden Dome formalised, missile-defence demand is now a structural budget category for the first time since the early Reagan-era SDI investments. Raytheon Missile & Defense generated roughly $26 billion in FY2025 revenue at segment operating margins of 12-13%. Operating leverage on backlog conversion in this segment runs at 25-30% incremental margin. Add the Golden Dome flow on top, and the segment alone could generate $32-35 billion in revenue by FY2028 at 14-15% segment margins. That is a multi-billion-dollar operating income tailwind from one segment alone.

Lockheed and Northrop have similar exposures in adjacent portions of the missile-defence architecture, but their multiples are already pricing those tailwinds. RTX's multiple is paying for them too, but the FCF inflection trajectory is steeper. That is the relative-value case.

The Pratt & Whitney GTF saga has been the dominant narrative drag on RTX for nearly three years. The recovery is now playing out faster than the Q1 2024 disclosure cadence suggested, and the operational data through FY2025 supports a more constructive read on FY2026-FY2027 than the consensus has so far adopted.

The customer-recovery accrual book peaked at roughly $5.2 billion in late FY2024. Drawdown through FY2025 ran at $1.4 billion, leaving roughly $3.8 billion still on the balance sheet at end of period. The remaining drawdown is heavily front-loaded into FY2026, with management guiding to roughly $1.6-1.8 billion of additional cash compensation flowing through the period. Beyond FY2026, the residual accrual is modest. The cash drag from the GTF programme is therefore mostly in the rear-view by end of FY2026.

The operational fix has progressed faster than disclosed. Powder-metal contamination causes have been identified, supply-chain remediation completed, and the inspection cadence on existing fleet shortened. The fleet-availability metrics, disclosed at the segment level on the most recent earnings call, are now within 4-5 percentage points of the pre-recall baseline. That convergence is materially ahead of the trajectory implied by the original FY2025 commentary.

For capital allocation, this matters because every additional billion of cash unwinding from the GTF programme is a billion that becomes available for buyback, dividend, or debt reduction. The Capital Desk model assumes management uses the FY2026-FY2027 unwinding to accelerate the buyback authorisation rather than reduce debt aggressively. At today's $174 print, that is roughly 18-20 million additional shares retired over two years, contributing 1.0-1.5% to per-share earnings growth above the operating EPS pace.