RKLB, ASTS, LUNR, RDW: real reversal or dead cat bounce?

1. What actually happened: two sectors under one label

The first mistake to avoid is treating “space & defense” as if it were a single sector. Over the past ten weeks two very different things have been moving, and anyone who looked at them as one thing understood nothing about what was going on.

On one side, traditional defense: large caps, multi-decade backlogs, real cash flow. It corrected barely at all and printed new highs. On the other, small- and mid-cap space: still lossmaking, valuations built on future revenue, and a sensitivity to rates and liquidity that in 2026 turned out to be brutal. It lost as much as three-quarters of its value in two months.

The drawdown: who paid the bill

The V: four stocks, one trajectory

This is the chart that explains what happened better than any other. I normalized the four names to a base of 100 at the late-May peak. They do not move similarly: they move identically. And when four companies with completely different businesses — launch vehicles, direct-to-cell satellites, lunar landers, space components — trace the exact same curve, the driver is not company-specific. It is flow.

The bounce: seven sessions, emerging-market numbers

The bounce heatmap, day by day

Something important shows up here: the bounce was not a single day of short covering. It was spread across five sessions out of seven, with a final acceleration on August 7. That is behavior different from a pure short squeeze, which typically burns out in one or two sessions.

Daily close-to-close percentage changes. Data source: Yahoo Finance chart API, Merlintrader calculations.

And meanwhile defense was printing new highs

ITA set its 52-week high on August 5, 2026. RTX was one-tenth of a percentage point from its own high on August 7, GD three-tenths. Anyone talking about a “defense sector in correction” is looking at NOC and LHX and mistaking them for the sector: both sit roughly 25% below their March highs for idiosyncratic reasons, not sector-wide ones. More on that shortly.

2. Why it crashed: the four causes, ranked by weight

2.1 The SpaceX effect: the structural cause

SpaceX went public on June 12, 2026 at $135 per share, in the largest IPO in U.S. history. The stock touched $211.39 on June 16, then slipped below its offering price, closing at $133.11 on August 7: -37% from the peak.

The mechanism is the classic one, and it deserves to be said plainly: a raise of that size does not create new money, it moves it. Investors sold existing space positions to fund the purchase of the new listing. July’s selloff was driven more by forced rotation into SpaceX and by rates than by any deterioration in underlying fundamentals. And that is exactly why the four stocks traced the same curve: they were not discounting four company stories, they were absorbing one single liquidity drain.

2.2 The doubt over AI capex

On July 24, Morgan Stanley (Adam Jonas) calculated that at $100 SPCX would imply a zero valuation for the AI division. The Q1 2026 figures cited are merciless in their clarity: Starlink with $1.19 billion of operating income, Space Launch at an operating loss of $662 million, and the AI division at an operating loss of $2.47 billion. Jonas’s comment — “many ascribe zero or negative value to AI given the heavy capex requirement” — sums up the problem. SpaceX had shed roughly $1 trillion of market cap from a $3 trillion peak.

2.3 China recovering a booster

On July 10, 2026, China’s Long March 10B successfully recovered an orbital booster with a net capture: a first for China. On July 13 the sector burned $88.6 billion of market cap in a single session. SPCX -4.4%, RKLB -5.3%, ASTS -7.7% (the worst), LUNR -6.3%, against a Nasdaq 100 down 1.9%. The implicit message the market priced in: the Western technology lead in reusability has an expiration date.

2.4 Rates, which count double for space

The U.S. 10-year yield climbed from 4.53% on July 7 to a six-month high of 4.74% on July 31. Two days after the sector low. That is no coincidence: for companies capitalized in the billions against still-lossmaking revenue and heavy capex, every basis point compresses the present value of future cash flows far more than it does for a Lockheed Martin generating $2.9 billion of free cash flow in a single quarter.

3. Earnings: nine of nine in defense, and that is not a figure of speech

The numbers here are unambiguous. Of the nine defense and defense-tech companies that reported between July 21 and August 4, 2026, all nine beat consensus on revenue and EPS, and all nine raised full-year 2026 guidance. Zero cuts. Zero lowered reiterations.

Backlog: the real story

If I had to keep one data point from the entire season, this would be it. Backlog is not a forecast: it is work already under contract. And this quarter set record after record.

When a beat is not enough: NOC, LHX and the lesson on positioning

Three cases deserve a close read, because they say something important about where expectations sit.

Northrop Grumman beat EPS consensus by 12.3%, pushed backlog to a record $104.7 billion, raised guidance on both revenue and EPS — and the stock fell 7.8%. The reason: segment operating income was down 5% year over year, GAAP EPS fell 6%, and there were charges on two programs. The market looked at earnings quality, not the press release headline.

L3Harris delivered revenue up 8%, EPS +28%, free cash flow +37%, a 1.2x book-to-bill and record backlog. The stock fell. The reasons cited: $800 million authorized for missile production facilities that compresses near-term returns, the missile division IPO pushed back to mid-2027 (removing a catalyst expected this year) and 60 basis points of margin compression in Space.

General Dynamics beat on everything and went nowhere.

The space names that have already reported: triple-digit growth, cash burning

The pattern is readable: revenue accelerating sharply, record backlog almost everywhere, but EBITDA still negative and cash draining. Redwire burned $48 million of free cash flow in the first half. Voyager burns $37.5 million of adjusted EBITDA in a single quarter. Karman is the exception that proves the rule: a 30% adjusted EBITDA margin, positive net income, and it is also the only one in the group that makes critical components rather than platforms.

4. Governments: here the story is structural, not cyclical

This is the part that separates an investment thesis from a bet on a bounce. Flows can turn in a week. A multi-year budget enacted into law cannot.

4.1 United States: the largest request since World War II

The FY2027 request comes to $1.5 trillion: $1,150 billion in discretionary plus $350 billion in mandatory funds via reconciliation, all allocated to a single fiscal year. It is the highest level in any fiscal year since World War II: +24% in real terms over FY2026 discretionary, +38% including reconciliation. Base procurement at roughly $260 billion, RDT&E at $220 billion, shipbuilding at $65.8 billion.

4.2 Golden Dome: where the military space money is going

More than any other program, this is what explains how a Rocket Lab or a Voyager can land contracts worth hundreds of millions. Funding: $24.4 billion from the 2025 reconciliation, roughly $20.5 billion obligated in FY2026, $17.5 billion requested for FY2027 — of which only $400 million is base funding and $17.1 billion is mandatory, meaning 98% depends on reconciliation.

On the program’s total cost there is a divergence worth knowing about: the Department estimates $185 billion through 2035, while the Congressional Budget Office estimated on May 12, 2026 roughly $1.2 trillion over twenty years, of which $720 billion for the space-based interceptor layer alone. General Guetlein rejected the CBO estimate as built on faulty data. No space-based intercept demonstration has ever been conducted: the first ground tests are slated for late 2026, the first flight demonstrations in 2027, and initial operational capability in the summer of 2028.

4.3 Space Force: +80% requested, and NSSL tripled

Procurement would go from $3.6 billion to $19 billion: more than five times. And on July 21, 2026 the NSSL Phase 3 Lane 1 IDIQ ceiling was tripled from $5.6 billion to $17 billion, with the cumulative program ceiling raised to $30 billion over a ten-year window and task order competitions running through FY2029. This is the pool Rocket Lab and the other launch providers will be fishing in over the next few years.

4.4 NASA: the risk from the other direction

Here the direction is the opposite, and it needs saying plainly. NASA’s FY2027 request is $18.829 billion against $24.438 billion in FY2026: -23%, with Science cut from $7.3 billion to $3.9 billion and more than 40 science missions slated for cancellation. The House CJS subcommittee rejected the cut on May 13, restoring NASA to $24.438 billion — the second consecutive year Congress has thrown out the proposed cut. But until there is a law, NASA operates under a continuing resolution.

For LUNR this is the main risk, and the main opportunity at the same time: on one hand NASA has announced more than twenty robotic lunar landings planned through 2029 to build infrastructure at the lunar south pole; on the other, multi-year funding for that program is not locked in. On the positive side: Artemis II completed its lunar flyby on April 7, 2026, breaking the Apollo 13 distance record.

4.5 Europe: 5% of GDP, with all the caveats that deserves

At the Ankara summit on July 7-8, 2026, the European allies plus Canada declared an increase in core defense investment of more than $139 billion in 2025, roughly +20% over 2024, with €70 billion for Ukraine in 2026 and more than $50 billion of new orders announced. For the first time, as of March 2026, every ally is above 2%.

On the instruments front: SAFE has allocated €150 billion of loans to 19 member states — Poland €43.73 billion, Romania €16.68 billion, France €16.22 billion, Hungary €16.22 billion, Italy €14.90 billion. Poland was first to sign and drew €6.6 billion on May 29, 2026, immediately committing roughly $16.5 billion to orders with PGZ on May 30. For Italy, the last verified data point (March 2026) showed access still blocked by the excessive deficit procedure. EDIP adopted a €1.5 billion work program for 2026-2027 on March 30.

A useful counter-example against idealizing Europe: Rheinmetall lost the F126 frigate program (more than €18 billion, cancelled in late June), the stock fell 20% and now trades roughly 45% below last year’s all-time high. Europe’s bottleneck is no longer winning orders: it is production capacity and converting orders into profit.

5. Geopolitics: no “peace dividend” in sight

The data point that matters for the sector

The point most directly relevant to RTX and Lockheed Martin is not a front-page story: Ukraine’s stocks of interceptor missiles for Patriot batteries are running out. RTX builds the GEM-T, Lockheed the PAC-3 MSE. That is structural demand, not a temporary risk premium.

Brent went from $74.16 on July 7 to $83.55 on August 7: +12.7% in a month, consistent with the renewed escalation. No negotiation on Ukraine has produced results: May’s three-day truce collapsed quickly, and June’s deadline for a comprehensive deal passed without a breakthrough. Russia occupies roughly 20% of Ukrainian territory.

The counter-signal on Taiwan

And here is the one piece of news that points the other way. In July there were 190 PLA air incursions into Taiwan’s ADIZ: less than half the July 2024-2025 volume, with six consecutive days without a single crossing of the median line. It looks like a strategic shift, not a seasonal one.

The flip side, though, is qualitative: the Type 076 amphibious assault ship Sichuan — 50,000 tons, electromagnetic catapult, effectively a near-carrier — demonstrated in August the ability to operate crewed aircraft, and on August 1 the PLA released images of amphibious assaults with uncrewed systems integrated into infantry units. Less noise, more capability.

6. The real question: will the next reporters confirm the sequence?

Everything covered so far is history. The test of the thesis lands in the next seventy-two hours, and it is packed together almost unnaturally tightly.

What they need to deliver to confirm

7. Reversal or dead cat bounce: the arguments, from both sides

How long it could last: three descriptive scenarios

8. Merlintrader Health Score

A 1-to-5 score built on five pillars: Balance sheet and runway 30%, Catalyst 30%, Dilution 20%, Liquidity 10%, Execution 10%. It measures the robustness or fragility of the company over a 12-18 month horizon. It is not a buy or sell signal and expresses no view on the price of the stock.

9. Analysts and sentiment

Consensus and price targets

Two observations. First: the $11-75 range on LUNR is extreme dispersion, and it reflects the fact that analysts do not even agree on the order of magnitude. Deutsche Bank reiterated Buy on July 31 while cutting its target from $34 to $20. Second: on RTX and LMT the average target implies 3-7% upside, and on LMT the prevailing rating is Hold with Goldman Sachs at Sell. In large-cap defense, the headroom consensus is willing to acknowledge is thin.

On RDW the post-earnings wave was tight: Cantor Fitzgerald raised its target from $9 to $14 on August 5, Alliance Global Partners from $15 to $16 on August 6, and on August 7 came reiterations from Truist, B. Riley and Roth MKM (the last at $20).

Retail sentiment

The recurring themes, summarized rather than reproduced:

• Anticipation for Monday’s RKLB print — by far the dominant theme, with speculation around the -$0.06 consensus and predictions for Tuesday’s opening price.
• Neutron as the dividing line — those who argue there are no hard deadlines coexist with those expecting a delay and asking how the market would react.
• The “V-shaped reversal” narrative — the idea that the July 29 low was the bottom, with SPCX used as a leading indicator for the group.
• Rotation from the giants to the pure plays — an explicit perception that the small names are outperforming SpaceX on their own catalysts.
• On RDW — enthusiasm on three legs: record backlog, military drones (the NATO Penguin Mk3 contract) and microgravity pharmaceuticals. The price-to-sales comparison against SpaceX recurs as an undervaluation argument.
• The bearish minority — one argument, repeated: valuation, with explicit use of the word “bubble.”

10. Bottom line

The June-July crash was not caused by fundamentals, and that is documentable: forced rotation into the largest IPO in U.S. history, rates at a six-month high in precisely the days around the low, a narrative shock out of China, and profit taking after very large gains. At the same time, sector fundamentals were improving: nine defense companies out of nine beat and raised guidance, backlogs set record after record, and the public money feeding the group is the highest since World War II on both sides of the Atlantic.

But the same reasoning, read in reverse, is the core argument for skepticism: if a sector falls 70% on flows, it can rally 45% for the same reasons. Seven sessions are not a base. And the problems that remain are structural and well identified: cash burn, systematic dilution at every window of strength, retail sentiment at extreme positioning levels, and a U.S. legislative calendar that this fall can generate noise.

Sources