Skunk Works AI Is Real. The Lockheed Martin Dividend Story Is What Actually Matters.

The headline grabs you with AI and fighter jets. The investment case is something much quieter, and arguably more important.

Do you know what scares me more than a world where defense becomes discretionary again? A world where the rearmament cycle that is reshaping the real economy runs out of steam while investors are still pricing defense stocks like cyclical growth stories.

The competitor headline about Skunk Works advancing sensor-powered AI fighter intercepts is real. It's just not the lead.

The AI part: what actually happened

Lockheed Martin's Skunk Works division has been steadily moving artificial intelligence from simulation to live-flight capability. In February 2026, Skunk Works and the U.S. Air Force Test Pilot School completed the Have Remy Test Management Project - a program where autonomous AI agents took direct control of the X-62A VISTA, a modified F-16, and performed missile-evasion maneuvers at the limits of the aircraft's flight envelope. The system completed over 100 test points with what Lockheed describes as robust sim-to-real transfer, meaning the AI behaved in live flight the way it did in simulation. Engineers could then recreate observed real-world behavior back in simulation, debug, and push updated code in hours. That real-to-sim feedback loop is the hard part in autonomy development.

Separately, Skunk Works demonstrated AI-powered anomaly detection for intelligence, surveillance, and reconnaissance platforms in collaboration with Spanish firm Arquimea. The system uses neural networks trained with episodic memory to spot irregularities in electro-optical and infrared sensor data - essentially teaching ISR platforms to recognize threats they've never seen before.

None of this is science fiction. Lockheed's Auto-GCAS system - automatic ground collision avoidance - has saved 13 pilots and 12 F-16s since entering service in 2014. The AI work represents the next iteration of that same philosophy: build autonomy that keeps people alive and extends combat capability.

This matters because it reinforces Lockheed's pricing power. When the Pentagon needs autonomous air dominance, there are three to four primes who can deliver, and Skunk Works is the closest thing to a monopoly on rapid technology prototyping. Pricing power isn't just about raising prices on existing products. It's about being the only supplier of next-generation capability when the customer has no alternative.

The financials: a $230 billion moat

Here's where the story becomes interesting for an income-oriented investor. Lockheed Martin's Q2 2026 results, reported July 23, showed revenue of $20.1 billion, up 11% year over year. Earnings per share were $7.94, which looks enormous compared to $1.46 a year earlier - but much of that gap was artificial, since the prior-year quarter absorbed $1.6 billion in program losses. The real signal is the organic operating margin of 12.4%, up from 10.1% in the first quarter, and management's decision to raise full-year guidance to $80.75 billion.

The number that actually matters is backlog: $230 billion. That is a record. With quarterly revenue around $20 billion, the backlog represents roughly 2.9 years of committed work. It includes a $35 billion multi-year contract for THAAD interceptors - Terminal High-Altitude Area Defense missiles that shoot down incoming ballistic and cruise missiles. If we annualize at current revenue rates, that single contract accounts for nearly 1.75 quarters of sales on a guaranteed basis.

New orders in the quarter were $65 billion. That is not incremental wishful thinking. That is signed contracts with the Department of Defense and allied governments, which historically cancel at very low rates.

Free cash flow for the trailing twelve months is $8.7 billion, up 162% year over year. That cash funds the dividend, buybacks, and working capital for programs that run for decades.

The macro regime: defense as the new utility

Global military spending reached $2887 billion in 2025, up 2.9% in real terms, according to SIPRI. Europe alone surged 14%. U.S. spending dipped in 2025 because no new military aid for Ukraine was approved that year, but fiscal year 2026 appropriations have pushed defense above $1 trillion - a 15% increase. The new NATO spending target of 5% of GDP by 2035 (up from the old 2% floor) adds hundreds of billions in future appropriations across the alliance.

I believe this is structural, not cyclical. Deglobalization, the war in Europe, the Indo-Pacific posture, and the depletion of munitions stocks during active conflicts have converted defense from discretionary spending into mandatory restocking. Governments are prioritizing national security above fiscal restraint, effectively guaranteeing revenue for prime contractors. Lockheed Martin increasingly operates like a highly regulated, government-backed utility rather than a traditional aerospace manufacturer.

The valuation question: is the stock too rich?

Lockheed trades at 21.5 times trailing earnings, 14.0 times EV/EBITDA, and yields 2.33%. Compared to RTX at 37.7 times earnings, General Dynamics at 23.1, and Northrop Grumman at 17.3, Lockheed sits in the middle. It is not cheap, but it is not expensive relative to the quality of its backlog, its cash flow recovery, or its position in the fighter and missile-defense markets.

The stock is up 21.2% year to date but down 6.9% over the past 120 days, suggesting the market has taken profits after a strong run while remaining supportive of the underlying trajectory. The PEG ratio - which factors earnings growth into the multiple - sits at 0.41, implying the earnings growth rate the market expects over the next two years runs well above 50%. Whether that's realistic depends on whether the backlog converts to revenue at the pace management projects and whether fixed-price development programs like the Next Generation Interceptor stay on budget. I don't take that for granted, but I also don't think the current multiple is pricing for perfection - just continued execution.

The dividend profile is the part investors who actually look at payout durability should appreciate. Lockheed has grown its dividend for 22 consecutive years. The TTM payout ratio sits at 65.4%, which is sustainable but not trivially so. Free cash flow of $8.7 billion against dividends of roughly $5.4 billion provides a comfortable cushion. The forward yield of 2.25% puts this squarely on the equity yield curve's growth side: not a yield trap, not a static income stock, but a compounder where the dividend itself grows into a much higher yield on cost over a decade or two.

The risk the headline doesn't mention

Two things worth sitting with. First, Lockheed's balance sheet is leveraged - net debt is $16.8 billion and the debt-to-equity ratio is 234%. That looks alarming on a spreadsheet, but defense contractors routinely carry elevated leverage because their revenue streams are government-guaranteed, their working capital cycles are stable, and their debt is almost entirely long-term investment-grade bonds. The interest coverage ratio is healthy, and the free cash flow growth trajectory means net leverage should decline over time. Still, it's a real variable to monitor, especially if the broader rate environment stays higher for longer.

Second, the AI narrative, while genuine, is still in the development and demonstration phase. Autonomous air-to-air intercepts are impressive technology milestones, but they are not yet line items on a balance sheet. The revenue from these programs will feed into the Next Generation Interceptor, crewed-uncrewed teaming architectures, and the F-35 Block 4 upgrade cycle - all multi-year timelines. The AI story strengthens the moat; it does not rewrite next quarter's earnings.

What this means for the portfolio

This is not a stock you buy for a quick yield. It's a stock that belongs in the income-growth sleeve because the combination of pricing power, backlog visibility, and payout durability supports compounding through whatever the business cycle does. Defense is a TOLL stock in the classic sense: it provides infrastructure that governments cannot function without, and it has the market position to pass through cost increases without losing customers.

The Skunk Works AI developments are important not because they make Lockheed an AI play - they don't - but because they extend the competitive lead that makes the $230 billion backlog defensible over the next decade. When your backlog equals nearly three years of revenue and your customer base has no viable alternative for next-generation fighter, missile, and autonomy technology, you have pricing power that most companies in the S&P 500 can only dream of.

I don't need the market to fall for this setup to make sense. From an income and risk/reward point of view, the appeal is a durable payout on a growing base, a valuation that isn't stretched relative to the defense peer set, and enough structural demand tailwind that the next few years of backlog conversion are highly visible. The question isn't whether Skunk Works will keep advancing its technology. It's whether investors are willing to hold a compounder that grows dividends through a rearmament cycle while everyone else chases yield.