NASDAQ:LUNR Intuitive Machines, Inc.

ISIN: US46125A1007
IndustrialsAerospace & DefenseSpace InfrastructureSELL · all three horizons
Sector: Industrials · Aerospace & Defense · Currency: USD Analysis Status: Stopped
Priced in US dollars. $19.01 is the close of Friday 14 August 2026, the latest print available — this report is dated Sunday 16 August 2026, with markets shut.
$19.01
+33.9% vs 4 Aug ($14.20)
2026-08-16 · Signal v6

What changed since the 4 August 2026 report — a Buy becomes a Sell, and the Pick is Stopped

The shares are $19.01, up 33.9% from the $14.20 at which we rated them short HOLD, medium BUY and long STRONG BUY on 4 August. That gain is not incidental to the downgrade; it is most of the reason for it. Two other things changed, and one of them is our own mistake.

DISCLAIMER: This is a quantitative framework for educational purposes only. It is not financial advice. Always do your own research and consult a licensed financial advisor before making investment decisions.

Intuitive Machines, Inc.

Intuitive Machines (Houston, Texas; founded 2013, NASDAQ-listed since February 2023 via a SPAC merger) builds, connects and operates space infrastructure for US government, national-security and commercial customers. It is the only company to have soft-landed a commercial spacecraft on the Moon twice — IM-1 in February 2024 and IM-2 in March 2025 — although both landers came to rest tipped over, cutting each mission short. Four business lines: Lunar Access Services (the Nova-C landers flown under NASA's Commercial Lunar Payload Services programme), Orbital Services, Lunar Data Services (the NASA Near Space Network Services contract, under which it is building a lunar relay constellation) and Space Products & Infrastructure. The company has been transformed by acquisition: Lanteris Space Systems (the former Maxar Space Systems satellite-manufacturing business) was bought in January 2026 for roughly $851m in cash and stock, and the Goonhilly Earth Station and COMSAT ground-station networks closed on 3 August 2026. Those deals took quarterly revenue from $50m to $206m and turned a NASA-dependent mission-services company into a would-be full-stack space prime. It has never made an operating profit.

HorizonSignalComposite ScoreConfidenceKey Driver
Short-term (1–3 mo)SELL5345%Tape is strong (+40.6% in a month, 50-DMA reclaimed) but the price sits ~3.6× the multiple the anchor warrants
Medium-term (6–12 mo)SELL4945%Backlog and diversification are real; a 4.56× multiple and four hard gates are not survivable at $19.01
Long-term (3–5 yr)SELL4745%Medium-quality business (negative ROIC, loss-making flagship contracts) bought at an Expensive multiple
Next update: 2026-08-30 — default +14d — no dated catalyst inside the window; the IM-3 launch is imminent but undated
Table of Contents
1Five-Pillar Scorecard2Hard Gates & Do-Not-Buy Status3Pillar Detail: Business Quality4Pillar Detail: Valuation Attractiveness5Pillar Detail: Underlying Drivers6Pillar Detail: Economic Alignment7Pillar Detail: Entry/Exit Timing8Economic Event Risk9Multi-Timeframe Technical Analysis10Price Chart (6-Month Daily)11Scenario Summary12Entry / Exit Rules13Position Sizing Context14Calibration Snapshot15Data Sources & Methodology
1

Five-Pillar Scorecard

Five independent scores — each 0–100 with its own confidence. The three fundamental pillars (Quality / Valuation / Timing) set the base BUY/HOLD/SELL via the Decision Matrix; the two context pillars (Underlying Drivers, Economic Alignment) then amplify a BUY to STRONG BUY or a SELL to STRONG SELL when both corroborate.

Business Quality

48
Medium — transformed in scale, unproven in economics
Confidence 45%

Valuation Attractiveness

38
Expensive — 4.56× actual vs 1.28× warranted (3.56×)
Confidence 60%

Entry/Exit Timing

62
Improving — post-print breakout, 50-DMA reclaimed
Confidence 70%

Underlying Drivers

70
Tailwind — US government space & missile-defence budgets
Confidence 65%

Economic Alignment

78
Tailwind — XLI Outperform / Outperform / Strong Outperform
Confidence 70%
2

Hard Gates & Do-Not-Buy Status

Binary safety checks — any TRIGGERED gate is a hard cap regardless of the scores above; CAUTION gates are sizing notes.
Gate 1 — Financial Distress
TRIGGERED — on the interest-coverage arm alone. Trailing-twelve-month operating income of −$134.851m against interest expense of $13.447m gives coverage of −10.03×. The gate fires below +1.5×, so it fires, and we are not writing this cohort a pre-revenue exemption. We do NOT rely on the four-consecutive-quarters-of-negative-free-cash-flow arm, and an earlier draft of this report wrongly asserted it. Q1 2025 free cash flow was positive at about +$13.3m (H1 2025 −$14.0m less Q2 2025 −$27.3m), and we did not re-read the Q3 or Q4 2025 cash-flow statements this run, so that arm is unproven rather than met. The verified quarters are Q2 2025 −$27.3m, Q1 2026 −$62.0m and Q2 2026 −$83.9m, with H1 2026 at −$145.8m. In fairness to the company the classic distress markers are absent: cash of $367.4m against $345.0m of 2.50% convertible notes that do not mature until October 2030 is a small net cash position, the current ratio is 1.66×, and there is no going-concern language in the 10-Q. What the gate is really flagging is that the business cannot yet fund itself from operations.
Gate 2 — Earnings Event Risk
Clear. Q2 2026 was reported on 13 August 2026; the next print is 12 November 2026, 88 days out and well outside the 14-day window.
Gate 3 — Valuation Ceiling
TRIGGERED. The name sits in the Valuation Anchor's Expensive band — an actual EV/Sales of 4.56× against a warranted 1.28×, a ratio of 3.56×, far beyond the 1.40× threshold. This arm has no growth exception. It caps the signal at HOLD; the base matrix had already produced SELL, which is below that cap, so SELL stands.
Gate 4 — Accounting / Dilution
TRIGGERED. On the two-year test, weighted-average Class A shares ran 55,093,365 (Q2 2024) → 117,434,775 (Q2 2025) → 162,172,470 (Q2 2026) — far above the 5%-a-year-for-two-years threshold. On the balance sheet, Class A outstanding went 121,281,880 at 31 December 2025 to 169,529,749 at 30 June 2026, plus 39.8% in six months (issued went 123,472,960 to 171,720,829, plus 39.1%; the difference is 2,191,080 treasury shares held flat across both dates — an earlier draft of this report quoted the issued figure as though it were outstanding). In the June 2026 quarter alone the company sold 8,259,379 shares through a new $500.0m at-the-market facility for $235.2m net of $3.6m of commissions, on top of an 11,574,069-share placement at $15.12 in February and 22,991,028 shares issued for Lanteris; 960,649 more followed for Goonhilly in August. About $261.2m of the ATM authorisation remains unused. Share-based compensation is not the problem here (4.9% of H1 2026 revenue); equity-funded growth is.
Gate 5 — Binary Event
TRIGGERED — and an earlier draft of this report recorded it as caution, which was inconsistent. The gate's third limb reads: “Outcome is genuinely binary (could go either way) and would move the stock >20%.” Unlike Gate 2, which carries an explicit 14-day window, Gate 5 carries no timing qualifier at all. IM-3 is confirmed by primary sources as the company's next launch and confirmed as not yet flown; a lunar landing after two consecutive tip-overs is as binary as an event gets, and we concede it would move the stock more than 20% either way. Having applied strict literalism to Gates 1 and 4 — on the reasoning that where this framework wants a carve-out it spells one out — we cannot then read a timing window into Gate 5 that is not written there. So it fires. What we could not establish, after checking the 3 August 8-K, the 10-Q and NASA's own CLPS event page (updated 31 July 2026, which gives the window only as “2026”), is a dated launch window — recorded as a confirmed absence in the Data Sources block, not as an assumption. The gate caps the signal at HOLD; the base matrix had already produced SELL, so there is no signal effect.
⚠️
Do-Not-Buy Triggers
None fire. Trigger 2(a) was adjudicated and is ruled OUT — a ruling, not an open question. 2(a) fires on a multiple at or above 2.0× warranted "with no exceptional, proven, durable growth." At 3.56× the numeric arm is comfortably met, so the whole thing turns on the growth carve-out, and we rule that the carve-out applies. Three reasons. First, the evidence: revenue +310% year on year, delivered rather than promised; backlog of $1.76bn, about 1.85× guided 2026 revenue; $920m of awards booked in Q2 and another $300m in Q3 to date; national-security revenue from 3% to 30% of the mix in a year. Second, internal consistency: Gate 3's own pre-profit arm is written as EV/Revenue above 20× for a non-hypergrowth company, and this name is at 4.56× forward on growth of that order — the framework's own hypergrowth allowance points the same way. Third, the remaining-performance-obligation schedule gives contractual revenue visibility into 2027 and beyond, so “durable” is supported and not merely hoped for. Firing 2(a) here would output a hard DO NOT BUY on a company whose revenue quadrupled, which the framework's own wording is written to prevent. The durability caveat stands as a monitoring item rather than as a qualification of the ruling: $947.3m of the backlog sits outside firm performance obligations, and much of the growth was bought rather than grown — if either deteriorates, 2(a) is re-adjudicated. Trigger 2(b) needs a live de-rating catalyst: LUNR is not in the AI-concentration cohort (no AI capex or monetisation leverage, not an index heavyweight, and its non-operating items are a drag rather than an inflator), and we identify no structural business-model threat, so 2(b) does not fire. Triggers 1, 3 and 5 do not fire. Trigger 4 (insider selling) could not be settled from Form 4 and Rule 144 metadata and is carried as an unresolved monitoring item, not as a pass.

How the four triggered gates interact

Gates cap the signal at HOLD; they never push it lower. The base Decision Matrix had already produced SELL at all three horizons from Medium Quality (48) × Expensive Valuation (38), so the caps are redundant to the signal. Where they do bite is the exit ladder: a triggered financial-distress or dilution gate is a catastrophic item under the Thesis-Invalidation exit group, which fires on its own without needing a second condition. That is why the exit action below reads Exit even though no stop-loss has been hit and the price is rising.
3

Pillar Detail: Business Quality

A deep dive into the Quality score: business economics, moat, ROIC and the industry benchmark.
Business Quality — Pillar Score
Medium — four times bigger than a year ago, and the part that existed a year ago is smaller.
48
conf 45%

Lifecycle: high-growth, early-commercial — and now, after two acquisitions, a scale business that has never earned an operating profit. Revenue recognition here is contract-milestone based, so quarters are lumpy by construction; the right lenses are backlog, cash runway against burn, award wins and losses, dilution and mission record — not price/earnings, not free-cash-flow yield, not return on equity. Book value is negative on the measure that matters for price-to-book, so that multiple is meaningless too — but say precisely which measure, because the Up-C structure distorts this line the same way it distorts market capitalisation. The −$290.1m deficit and the −$4.22 tangible book value per share are permanent equity: they exclude $1,194.7m of redeemable non-controlling interests, which is economic ownership of the same business sitting in mezzanine equity, above the equity line. That balance is carried at redemption value and marked to the share price, which is why it rose $243.1m in H1 2026 — from $951.5m to $1,194.7m — while the business lost money. So the correct statement is not “there is no equity”; it is that permanent equity attributable to shareholders is negative while a larger economic claim is carried outside it at a mark-to-market value. We do not restate book equity as positive on that basis: a balance that rises because the share price rose is not a measure of what the business is worth.

The headline and the underlying number point opposite ways, and the report leads with that. Q2 2026 revenue of $206.2m was up 310% on Q2 2025's $50.3m. But the 10-Q attributes the increase “mostly” to the January 2026 Lanteris acquisition, which contributed $166.7m in the quarter. Strip it out and the legacy service business ran at $36.7m of service revenue in Q2 2026 against $50.3m in Q2 2025 — down 27%. The 10-Q names the causes: the NASA Near Space Network contract fell $7.3m on schedule delay and an unfavourable estimate-at-completion, OMES III fell $1.8m after NASA cancelled the OSAM task orders, the completed LTV contract fell $5.8m, and IM-4 fell $3.3m on higher estimated contract costs. So: the group is four times bigger, and the part of it that existed a year ago is smaller.

The second thing the filing shows, which our earlier reports did not surface, is that the flagship lunar franchise loses money contract by contract. The 10-Q states plainly that the IM-3 mission contract “became a loss contract in 2021” and that IM-4 “became a loss contract during the second quarter of 2025”. In the six months to 30 June 2026 the company booked a further $6.3m of losses on IM-3 and $16.2m on IM-4. These are fixed-price NASA contracts and the overruns land on Intuitive Machines.

Be precise about what is and is not deteriorating here. The combined additions on the two named contracts were $22.5m in H1 2026 against $21.2m in H1 2025 — essentially flat year on year, not a worsening trend. Underneath that, the two moved in opposite directions: IM-3's additions fell from $18.5m to $6.3m, while IM-4's rose from $2.7m to $16.2m as payload obligations grew. So the level of the drag is real and it legitimately re-scores pricing power — a company absorbing about $22m a half-year of overruns on its flagship product is not a company that can price it. But this report should not, and now does not, describe the drag as accelerating. This is not a new fact either — it is a fact our three prior reports never surfaced, and it is the single largest reason the Quality score falls.

Sub-signal scorecard

Sub-signalReading (period labelled)Score
Revenue trajectoryQ2 2026 $206.2m, +310% YoY; H1 2026 $392.9m, +248%. FY2026 guidance $900m–$1.0bn against FY2025 actual $210.0m. But organic service revenue −27% YoY in Q2 2026.62
Profitability vs peersQ2 2026 gross margin 17.4%; operating margin −22.9%; company-defined Adjusted EBITDA −$13.8m in Q2 2026 and −$11.1m for H1 2026. Guidance is for FY2026 Adjusted EBITDA to be positive — which needs a swing of more than $22m against the H1 run-rate in the second half.48
Cash generationH1 2026 free cash flow −$145.8m (operating −$111.9m, capex −$33.9m); Q2 2026 alone −$83.9m; Q2 2025 was −$27.3m and Q1 2026 −$62.0m. Not uniformly negative, and we say so: Q1 2025 was positive at about +$13.3m, and we did not re-read the Q3 or Q4 2025 statements this run. Free cash flow is heavily negative and getting more so, but the four-consecutive-quarter test is unproven on the filings we have read.25
Balance-sheet healthCash $367.4m (plus $11.7m restricted) against $345.0m face of 2.50% convertible notes due October 2030 — a net cash position of about $22m. Current ratio 1.66×, quick 1.51×. Against that: total shareholders' deficit of −$290.1m and $379.2m of goodwill plus $297.1m of intangibles carried from the acquisitions, giving tangible book value of −$4.22 a share (−$290.082m less $379.216m of goodwill less $297.069m of intangibles = −$966.4m, spread across all 228,924,068 economic shares). An earlier draft of this report published −$5.92, computed on the Class A weighted average — the very share-count error this report exists to correct, made in the opposite direction.52
Backlog (the Industrials organic-demand metric)$1,762.0m at 30 June 2026 against $213.1m at 31 December 2025. Of the $1,549m increase, $612.8m was acquired with Lanteris and $1,340m was new awards. Backlog covers about 1.85× guided FY2026 revenue.72
Backlog quality (the caveat that matters)Only $814.7m is firm remaining performance obligation. The $947.3m difference is mostly $587.0m booked against a multi-satellite programme on which the company has received a $45.0m authority to proceed and recorded an estimated total programme value above $600m, plus $316.0m of funded-value contracts billed as performed and $44.3m of constrained variable consideration.38
Customer diversificationGenuinely better. Q2 2026 mix: commercial 32%, civil 37%, national security 30%. Q2 2025: civil 93%, commercial 4%, national security 3%. A NASA-dependent company has become a three-legged one in twelve months.72
ROIC vs cost of capitalNegative. Operating loss of $47.1m in Q2 2026 on $1,887.3m of total assets. There is no return on invested capital to measure against an 11.13% discount rate.18

Industry benchmark — ROIC vs WACC + backlog growth (Industrials)

ROIC: negative, against a cost of capital of 11.13% — the gap is not close.
Backlog: $1,762.0m at 30 June 2026 vs $213.1m at 31 December 2025, and rising again in Q3 with $300m of further awards booked quarter-to-date.
Rating: MIXED — the framework's “moderate” band (50–84), at its lower edge. One leg of the benchmark is as strong as it gets; the other has not started. Benchmark score: 55 (from 58 in the 4 August report — the backlog leg improved sharply, but the operating loss widened from −$28.6m in Q2 2025 to −$47.1m in Q2 2026, so the return leg got worse).

Competitive Environment (step 7c) — who is attacking, and which way share is moving

Two distinct competitive arenas, and they are moving in opposite directions.

RivalWhere it competesShare trajectory vs LUNR
Firefly AerospaceNASA CLPS lunar delivery — head to head, mission for missionGaining, and on the dimension that matters most. Firefly's Blue Ghost Mission 1 landed fully upright and operated a complete surface mission in March 2025. Both Intuitive Machines landings — IM-1 in February 2024 and IM-2 in March 2025 — came to rest tipped over, ending each mission early. A customer choosing a lander today has a rival with a clean record.
AstroboticCLPS lunar deliveryWeakened — Peregrine failed before reaching the Moon in January 2024 — but still an active CLPS provider bidding task orders.
Blue OriginLunar landers (Blue Moon), NASA Human Landing SystemEncroaching, with a balance sheet Intuitive Machines cannot match.
SpaceXLaunch (supplier), Starship HLS (competitor), and now a listed comparable after its IPOBoth sides of the ledger at once: every Nova-C mission flies on a Falcon 9, so SpaceX is a concentrated supplier and the operator setting the industry's cost curve.
Lockheed Martin, Northrop Grumman, L3Harris, BoeingNational-security space, the arena LUNR just entered with the 18-spacecraft Golden Dome AMDT3 awardIncumbents. LUNR is the challenger here, taking share from a very low base — 3% to 30% of revenue in a year.
Airbus Defence & Space, Thales Alenia, Rocket Lab, MDA SpaceCommercial satellite manufacturing — the Lanteris businessStable-to-competitive. Rocket Lab has bought its way into spacecraft manufacturing and MDA Space announced a $620m acquisition to enter the US defence market in June 2026. The $600m-plus three-satellite GEO award shows LUNR can win here.

How this sets the moat sub-scores. The competitive read is the direct input, not a separate narrative. Pricing power scores 25 because the evidence is unambiguous and audited: IM-3 and IM-4 are loss contracts, with a further $22.5m of losses booked on the two of them in H1 2026 alone. A company that cannot price its flagship product above its own cost has no pricing power, and a credible rival with a cleaner landing record makes the next competition harder, not easier. Cost advantage scores 35 for the same reason, read from the cost side. Switching costs score 60 — real but bounded: NASA periods of performance run long (IM-6 to May 2031) and qualified spacecraft buses plus security clearances are genuine friction, but CLPS task orders are competed mission by mission rather than sole-sourced. Share trajectory: mixed — rising in awards and in national security, eroding in the lunar-landing franchise that gives the company its identity. Threat level: elevated.

Competitive moat scorecard

DimensionAssessmentScore
Pricing powerDerived from the competitive read above: fixed-price CLPS contracts where the overruns land on the company. IM-3 a loss contract since 2021, IM-4 since Q2 2025.25
Network effectsNascent. The build–connect–operate model plus the Goonhilly and COMSAT ground network could become a data-relay network with real pull — a lunar relay constellation gets more useful as more missions fly. Today it is a plan, not an effect.40
Switching costsMulti-year government periods of performance, qualified buses, clearances — but per-mission competition.60
Cost advantageNone demonstrated; the loss contracts are the evidence against it.35
Intangible assetsThe strongest dimension. Six CLPS awards including a sixth won this quarter under the Moonbase programme; prime on NASA's Near Space Network Services; newly prime on the Lunar Reconnaissance Orbiter Camera and ShadowCam; precision-landing intellectual property; ITAR and clearance barriers; and, through Lanteris, heritage on more than 300 spacecraft built.65

Moat score: 45 (from 53 on 4 August). The fall is a re-derivation, not new competitive news: the loss contracts are the audited evidence that pricing power and cost advantage were scored too generously before.

ROIC, capital allocation and skin in the game

ComponentReadingScore
ROIC percentile vs peers (40%)Negative operating profit against aerospace primes running 10–15% ROIC. Bottom of the sector.18
Capital allocation (30%)Mixed, and genuinely arguable both ways. Lanteris ($853.3m of total consideration per the 10-Q — $447.1m of net cash plus 22,991,028 Class A shares) transformed scale, mix and backlog in a single stroke, and Goonhilly plus COMSAT (£37m, half cash and half stock, plus $10.0m for COMSAT) adds ground infrastructure the missions actually need. Against that, it was paid for by issuing stock at $15–18 while the organic base shrank, and the acquired goodwill and intangibles now exceed the entire market value of the Class A float.42
Management skin in the game (30%)High and real: the founders hold 55,692,725 Class C shares, about 24% of total economic shares, alongside 173,231,343 Class A. Set against continuing Rule 144 dispositions through June and July 2026 and $19.3m of H1 share-based compensation.58

Composite: 0.40×18 + 0.30×42 + 0.30×58 = 37.

Business Quality — 48 (from 60 on 4 August)

Weighted: universal sub-signals 30% (46.8) · sector-specific 15% (54.0) · industry benchmark 18% (55) · moat 22% (45) · ROIC and capital allocation 15% (37) = 47.5, rounded to 48. Still squarely Medium (40–64), as it was at 60 — so the Decision Matrix row does not change on this pillar. Three things drive the twelve-point fall, and only the first is new news: the organic revenue decline disclosed in the Q2 filing; the loss-contract status of IM-3 and IM-4, which this report is surfacing for the first time; and a moat re-derivation that follows from it. Pulling the other way, and genuinely: backlog, customer diversification and the cash balance are all materially better than they were a quarter ago. To reach High (≥65) — the threshold that would have changed the matrix row from SELL to HOLD — this business would need positive returns on capital and a flagship franchise that does not lose money contract by contract. It has neither. We deliberately do not list “positive book value” as a third blocker: for an Up-C structure whose largest economic claim is carried in mezzanine equity at a mark-to-market redemption value, book equity is not a clean quality test in either direction.
4

Pillar Detail: Valuation Attractiveness

Sector-appropriate multiples, FCF yield, reverse-DCF implied growth, embedded optionality, and the analyst-consensus cross-check.
Valuation Attractiveness — Pillar Score
Expensive — 4.56× guided 2026 revenue against a warranted 1.28×.
38
conf 60%

First, a correction to our own arithmetic — the enterprise value was wrong

Intuitive Machines has an “Up-C” structure. The Nasdaq-listed Class A shares are only part of the economic ownership; the founders hold Class C shares that convert one-for-one and carry the same economics. The 10-Q cover page dated 6 August 2026 reports 173,231,343 Class A, zero Class B and 55,692,725 Class C shares outstanding — 228,924,068 economic shares in total.

The providers disagree with each other, and it is worth being exact about how, because an earlier draft of this report over-simplified it. FMP reports a market capitalisation of $3.03bn — which at $19.01 implies 159.37m shares, stale even as a Class A count — and an enterprise value of $3.09bn; both are the Class A float alone and understate our $4,351.8m of equity by about 30%. (The 24% figure an earlier draft quoted is a different quantity: it is Class C's share of the total economic count.) Yahoo reports a market capitalisation of $3.05bn on the same Class-A-only basis, but its enterprise value of $4,583,148,032 is higher than ours, because it adds the redeemable non-controlling interests at their $1,194.7m book carrying value rather than marking the Class C shares to the $19.01 market price. So it is not true that every provider ignores the Class C economics — Yahoo picks them up on the enterprise-value line, at book. Our own 4 August report did inherit the error, and published an EV/Sales of 2.6×.

Note the direction, because it strengthens the conclusion rather than weakening it. Reconciling to Yahoo's enterprise value instead of ours gives EV/Sales of 4.82× on guided 2026 revenue and a warranted ratio of 3.77×. We keep our own $4,329.5m — marking every economic share to market is the cleaner construction — but the honest range is 4.56× to 4.82×, ratio 3.56× to 3.77×, and both ends sit deep in the Expensive band.

Corrected: 228,924,068 × $19.01 = $4,351.8m of equity, plus $345.0m of convertible notes at face, less $367.4m of cash = enterprise value $4,329.5m. On a like-for-like basis the 4 August multiple was about 3.5×, not 2.6× — so the honest move this fortnight is 3.5× → 4.56×. That is a multiple expansion, not a methodology artefact, and it is most of what changed.

What the shares actually cost, on every base we can compute

BasisRevenue / value baseMultiple
EV / Sales — trailing twelve months$490.1m (Q3 2025 + Q4 2025 + Q1 2026 + Q2 2026)8.83×
EV / Sales — FY2026E (the primary multiple used)$950m — the midpoint of company guidance of $900m–$1.0bn issued 13 August 20264.56×
EV / Sales — FY2026E on sell-side consensus$935.1m (5 analysts)4.63×
EV / Sales — FY2027E$1,150.0m consensus (4 analysts)3.76×
EV / contracted backlog$1,762.0m at 30 June 20262.46×
EV / firm remaining performance obligations$814.7m at 30 June 2026 — the part that is contractually committed5.31×
Free cash flow yieldH1 2026 FCF −$145.8mNegative — not meaningful
Trailing P/ETTM EPS −$0.94 (Class A basis)−20.24× — not meaningful, recorded only so the currency-coherence check can run

The Valuation Anchor — a warranted multiple, computed rather than compared

The anchor exists to stop the pillar importing a bubble's own baseline as “normal”. It judges the price against interest rates, disciplined growth and sector risk. We considered withdrawing it — the framework permits that for a negative-EBITDA name — and rejected the idea. A multiple does resolve here: EV/Sales on guided and contracted revenue. Withdrawing an anchor because you dislike where it points is how a report ends up flattering a price, so we have run it and followed it.

InputValueSource / discipline
Risk-free rate4.63%A declared deviation from the SKILL, not an oversight. The framework says to read the 10-year from the macro report's market_snapshot.UST10Y. That field is absent from MacroDriver-state-20260812.json altogether, so the instruction cannot be satisfied from the state file, and the macro report's prose carries a different number again — 4.70%. We therefore take 4.63% from FRED series DGS10 at 13 August 2026 and attribute it to FRED. Mitigation: 4.70% would make this name more expensive, not less — the warranted multiple falls to 1.26× and the ratio rises to 3.62× — so the deviation cannot be flattering the conclusion.
Equity risk premium4.50%Framework constant.
Risk add-on+2.00%Beta is 1.829, above the 1.6 threshold that mandates the top add-on.
Discount rate r11.13%4.63 + 4.50 + 2.00.
g (years 1–5)10.0%Consensus implies a 21.4% revenue CAGR from FY2026 to FY2030. The framework haircuts consensus by 25% (to 16.1%) and then caps at what the sector can sustain — 10% for the cyclical/normal bucket that Industrials sits in. Feeding the hype growth in is the banned move, because an optimistic g will warrant any multiple.
g (terminal)3.0%Long-run nominal GDP cap.
Warranted P/E (two-stage)16.89×Below the 23× Industrials guardrail, so the cap does not bind.
Sustainable net margin7.6%The one discretionary input. Anchored to the sell-side's own FY2030 estimate — $154.0m of net income on $2,032.7m of revenue — and corroborated by the aerospace primes, which run 7–8% net. Against a Q2 2026 gross margin of 17.4%, this is a generous steady state, not a punitive one.
Warranted EV/Sales1.28×16.89 × 7.6%. Warranted price/sales converts to EV/sales here because net debt is about −$22m — half a percent of equity value — so the two are the same number to the second decimal.

Result: actual 4.56× ÷ warranted 1.28× = 3.56× → EXPENSIVE

The Expensive band starts at 1.40×. This is 3.56×. It fires Gate 3, the Valuation Ceiling, which has no growth exception.

Sensitivity — because a single discretionary input should not carry a SELL

The obvious attack on the conclusion is that we chose the growth cap and the margin. So here is the whole grid. The question is not whether our central case says Expensive; it is whether any defensible parameterisation does not.

Growth assumption for years 1–5Warranted P/EWarranted EV/Sales at 7.6% net marginRatio vs actual 4.56×Band
10% — the Industrials cap the framework prescribes (central case)16.89×1.28×3.56×Expensive
15% — the secular-growth cap, i.e. treating a space company as if it were software20.58×1.56×2.91×Expensive
20% — the “proven, durable >20% grower” exception, flagged as the framework requires24.93×, capped to 23.00×1.75×2.61×Expensive

Now the margin lever, run backwards. To pull the ratio down to 1.40× — merely to the edge of Expensive, not out of it — the sustainable net margin would have to be 19.3% at the prescribed growth cap, or 14.2% even at the guardrail-capped 23×. A company whose gross margin was 17.4% last quarter cannot carry a 19.3% net margin; the primes it would be competing with earn 7–8%. Neither figure is defensible.

No defensible parameterisation clears the line. Every cell in the grid is Expensive, and two of the three are above 2.0×. That is what makes this a valuation conclusion rather than a choice of assumption.

The sector guardrail — adjudicated by hand, because the automated arm cannot run

The Industrials guardrail line is a P/E of 23×: at or above it, a name is Expensive regardless of the warranted ratio. Our primary multiple is EV/Sales, not P/E, so the deterministic check skips this arm — and a skipped check is not a passed one. Adjudicating it directly: consensus puts FY2028 earnings per share at $0.329, which on the $19.01 price is 57.8×. But that consensus EPS is itself built on a stale share count — $47.0m of FY2028 net income divided by $0.329 implies about 143m shares against 228,924,068 actually outstanding. Correct the denominator and FY2028 EPS is $0.205 and the multiple is roughly 93×, four times the guardrail line, on earnings still two years away. FY2027 consensus EPS of $0.014 does not produce a meaningful multiple at all. The guardrail is breached on every honest construction of it.

Relative cross-checks — they order the name inside the band; they cannot lift it out

LensReadingScore
Warranted-multiple anchor (40%)3.56× — deep in the Expensive band.8
Sector median (20%)Cuts both ways honestly. Against aerospace primes at roughly 1.5–2.5× EV/Sales, 4.56× is dear. Against listed space pure-plays it is not — that cohort trades far higher, which is exactly the “expensive-everywhere-together” problem the anchor exists to catch.55
Own-history decile (15%)The stock is 59% below its intraday high of $46.75, made in the week of 25 May 2026 and sits at the 29th percentile of its 52-week range ($7.78–$46.75). On its own history this is a cheap-ish decile.68
Growth-adjusted (10%)4.56× EV/Sales against a 21.4% forward revenue CAGR is 0.21 on a sales-PEG basis — attractive on that lens alone.72
Analyst consensus (15%)$19.01 against an FMP consensus target of $31.67 (high $41, low $26) and a Yahoo mean of $29.50 (high $43, low $11, 8 analysts) — 40% below consensus, which the table would score 85–100. We cut it to 60. The same panel's earnings estimates use roughly 143m shares against 228.9m outstanding, so these targets are not independent evidence; they are contaminated by the identical share-count error we just corrected in our own work. Grades are 9 buy, 1 hold, 1 sell — 81.8% bullish — with no rating change by any firm in the last 30 days.60

Blended, the five lenses give 40.6. The anchor is supreme, so the score is held inside the Expensive band at 38 — the top of it, which is the honest place for a name where every relative lens is supportive and only the intrinsic one is not.

Embedded optionality — what you get for nothing at $19.01

There is real free upside here and it deserves saying plainly. The $600m-plus three-satellite GEO programme is carried on a $45.0m authority to proceed — if it converts to a funded contract, backlog quality improves in one step. The 18-spacecraft Golden Dome AMDT3 award won in July 2026 is an entry ticket to a missile-defence build-out that could dwarf the lunar business. The Near Space Network lunar relay constellation — with long-lead materials for satellites two through five already bought — is a recurring-revenue business the market is paying nothing for today. And the Goonhilly and COMSAT ground network, with more than 90 antennas, is infrastructure that is hard to replicate. We tilt the Valuation score up a few points for this, within the band. It does not turn a 3.56× multiple into a cheap one — optionality is the reason to keep watching, not the reason the stock is cheap.

What the price implies, and what fair value looks like

Run the anchor backwards: warranted EV/Sales of 1.28× on $950m of FY2026 revenue implies equity of about $1,238m, or $5.41 a share. Run it through earnings instead — the sell-side's FY2030 net income of $154.0m, at the warranted 16.89×, spread over all 228,924,068 economic shares and discounted back 4.4 years at 11.13% — and you get $7.14. We publish $7.00 as the fair-value estimate, taking the more generous of the two constructions because it lets revenue actually grow to 2030 before applying the multiple. Put the other way round: at $19.01 the market is pricing roughly 30% compound revenue growth and a mid-teens sustainable net margin and no further dilution. Any two of those three might arrive. All three, on a company whose organic revenue just fell 27% and whose flagship contracts lose money, is a demanding bet.
5

Pillar Detail: Underlying Drivers

The dominant external force the stock is tethered to, scored 0–100. A context pillar: it does not change the base signal — it feeds amplification (tailwind ≥65 can lift BUY→STRONG BUY; headwind ≤35 can push SELL→STRONG SELL).
Primary Driver
US government space & missile-defence appropriations
70
Tailwind — cannot amplify a SELL

Intuitive Machines is a government-budget stock before it is anything else. Civil (37%) plus national-security (30%) customers were 67% of Q2 2026 revenue, and even the commercial satellite work leans on operators whose own demand is policy-shaped. The driver is US government space and missile-defence appropriations — NASA's Commercial Lunar Payload Services and Moonbase programmes, Artemis, the Near Space Network, and now the Space Force / Golden Dome missile-defence build-out.

HorizonReadingWeightScore
Historical (last 12–24 months)Volatile but net expanding. NASA's lunar budget was squeezed through 2025 and the agency cancelled the OSAM task orders under OMES III, which cost real revenue. Since then the direction has reversed: additional lunar funding was allocated in 2026 and the Moonbase initiative created a new award stream.25%70
Current stateStrong, and specific. In the last quarter the company won a sixth CLPS lander contract under the Moonbase programme, two NASA prime lunar-reconnaissance awards (the Lunar Reconnaissance Orbiter Camera and ShadowCam), and — in July 2026 — an award for 18 spacecraft supporting the Accelerated Missile Defense Tranche 3 “Golden Dome” constellation. National-security revenue went from 3% to 30% of the mix in twelve months. The offset is that the same government cancelled OSAM and let the Near Space Network contract slip.50%75
Forward outlook (6–12 months)Supportive but with a named fiscal risk. Golden Dome is a multi-year programme and Artemis/Moonbase continues. Against that, US fiscal and sovereign-debt strain is one of the macro framework's live drivers, and discretionary space budgets are exactly what gets trimmed when appropriations tighten. The 10-Q lists government shutdowns and budget-process failure among its own principal risks.25%62

0.25×70 + 0.50×75 + 0.25×62 = 70.5, recorded as 70 — a Tailwind (the 65–79 band), nudged up from 68 on 4 August by the Golden Dome award and the sixth CLPS win. Not commodity-leveraged, so the price-trend overlay does not apply.

Amplification role: a driver of 70 is eligible to lift a base BUY to STRONG BUY. It cannot do anything for a SELL — a SELL intensifies to STRONG SELL only when the driver is at or below 35 and economic pressure is a Headwind. Here the driver is a Tailwind and so is the economy, so the two context pillars point the opposite way from the fundamentals. The framework's rule is explicit that this disagreement is recorded as a caveat, not as an upgrade: the base signal stands unamplified. Said plainly — the wind is behind this company, and the shares are still priced past what the wind is worth. Driver confidence 65% (base 70, less 5 because the appropriations link is real but indirect).

6

Pillar Detail: Economic Alignment

How the current economic climate sits relative to this stock, read from the latest Macro-Economic report. Classifies the macro pressure (Tailwind / Neutral / Headwind) — the second amplification input — and frames a long entry as Trend-Following or Contrarian with a 0–100 conviction.
Stance · Pressure
Trend-Following · Tailwind
78
conviction

The macro report of 12 August 2026 reads the regime as energy-shock stagflation: a supply-driven inflation impulse hitting a contracting labour market, with Fed cuts priced out and the debate live between hold and hike. Within that, Industrials (XLI) is one of the two strongest rows on the board — Outperform short, Outperform medium, Strong Outperform long — bettered only by Materials (XLB) at Outperform / Strong Outperform / Strong Outperform, so XLI is second, not first. LUNR does not appear in the macro report's watchlist_forecast, so we take the sector map, as the framework directs.

Two live qualifications. First, the 14 August prints — July retail sales −0.6% against a consensus near +0.1%, and Michigan consumer sentiment at 51.0 against 54.5 expected — post-date the macro report and mark a genuine consumer break. They are not a reason to mark this name down: Intuitive Machines sells lunar landers, satellites and ground stations to NASA, the Space Force and satellite operators, and has no consumer exposure worth naming. Second, and cutting the other way, the same fiscal strain that is driving the macro framework's US-debt driver is the mechanism by which a discretionary space budget gets cut.

Net: Tailwind, conviction 78 (from 76), stance Trend-Following — the sector's own trend is up and this name is levered to the strongest part of it. That is the case for the business. It is not a case for the price.

Source: sector map XLI (O / O / SO) — not in watchlist_forecast · Macro report 2026-08-12

7

Pillar Detail: Entry/Exit Timing

The risk-reward framework, relative strength vs SPY and the sector ETF, the macro overlay, news-derived sentiment, and the catalyst cluster.
Entry/Exit Timing — Pillar Score
Improving — 50-DMA reclaimed on 1.8× volume, MACD histogram positive ten sessions.
62
conf 70%

This is the pillar that improved, and it improved a lot. On 4 August the score was 44 with bearish confluence; it is now 62. The Q2 print landed on the morning of 13 August and the stock rose 3.6% that session and 8.3% on 14 August — +12.2% across the two sessions — closing at $19.01 on volume of 18.9m shares, 1.8× its 20-day average.

ComponentReading (all levels from 14 August 2026)WeightScore
Multi-timeframe trendMonthly and weekly uptrend, daily recovering — the close of $19.01 is back above the 50-day SMA at $18.58 but still just under the 200-day at $19.31, which is the next overhead level. Hourly strongly up; 15-minute weakening. Weighted MTF score 68, confluence Mostly Bullish.30%68
Risk-reward and relative strengthPosition risk is reasonable: nearest daily support at $17.91 is 0.67 ATR below, the next shelf at $16.17 is 1.72 ATR below, and ATR is $1.65 — 8.7% of the price, which is very high. Deduct for proximity to the 200-day resistance 1.6% overhead. Relative strength is genuinely split: over one month LUNR is +40.6% against SPY +4.5%, XLI +4.0% and ITA +9.8%; over three months it is −43.9% against SPY +5.0%, XLI +8.8% and ITA +16.5%. A violent bounce inside a savage downtrend.20%53
Macro overlayXLI Outperform/Outperform/Strong Outperform; Industrials carries the framework's medium macro-sensitivity weight of 0.15.15%70
SentimentEstimate revisions are rising and the print beat. But no analyst firm has changed its rating in 30 days — the most recent action of any kind was a Roth Capital maintain on 28 May 2026 — so the grade panel (9 buy, 1 hold, 1 sell) is a stale “all-maintain” reading dressed up as conviction. FMP's own financial-health rating is C (overall score 2 of 5), with 1-of-5 on discounted cash flow, return on assets, debt/equity, price/earnings and price/book. That divergence from the analyst panel is informative and we side with the ratings snapshot.18%62
CatalystsNothing dated inside 30 days. Q3 earnings are 12 November 2026. The IM-3 launch is imminent but carries no published date, which is its own kind of risk — an undated binary you cannot position around.17%55

0.30×68 + 0.20×53 + 0.15×70 + 0.18×62 + 0.17×55 = 62.0Improving (the ≥55 band). Timing confidence 70%: base 75, less 5 for a name whose average true range is 8.7% of its price and which has travelled $46.75 → $11.29 → $19.01 inside three months.

Why a 62 on timing does not rescue the signal

Timing carries 55% of the short-horizon weighting, which is why the short composite (53) is the highest of the three. But the Decision Matrix is a lookup on the three pillars, not a threshold on the composite: Medium quality × Expensive valuation returns SELL for any timing score. A good tape on an Expensive name is the one combination the framework is most deliberately built to refuse, because it is precisely how momentum talks people into paying up. The honest reading of the strong tape is that it is what created the problem — the stock is up 33.9% since our last report, and that move is the reason the signal has changed.
8

Economic Event Risk

High-impact macro releases in the next 14 days that could swing this stock, plus the last 7 days of surprises.

Upcoming events (next 30 days)

DateEventImpactForecastPreviousRelevant?Why
2026-08-19FOMC MinutesHigh⚠ MediumRate path sets the discount rate in the anchor; a pre-profit, long-duration name is among the most rate-sensitive things in the market
2026-08-25CB Consumer Confidence (Aug)High90.8✗ NoNo consumer exposure — government and satellite-operator customers
2026-08-26Core PCE Price Index MoM (Jul)High0.3%0.1%⚠ MediumFeeds the Fed path and therefore the 10-year, which sets r
2026-08-26Durable Goods Orders MoM (Jul)High0.2%0.3%⚠ MediumThe nearest read on the Industrials capital cycle
2026-09-01ISM Manufacturing PMI (Aug)High55.055.6⚠ MediumSector-level demand signal for XLI
2026-09-04Non-Farm Payrolls (Aug)High12k−23k⚠ MediumLabour-market read drives the hold-vs-hike debate
2026-11-12LUNR Q3 2026 earningsHighEPS −$0.065 · revenue $243.4m✓ YesThe single most important scheduled event: the first test of the FY2026 Adjusted-EBITDA-positive guide

Recent surprises (last 7 days)

DateEventActualForecastSurpriseImpact
2026-08-12CPI YoY (Jul)3.4%3.4%In lineNeutral — no change to the rate path or to r
2026-08-13Producer Price Index MoM (Jul)0.0%0.2%−100%Mildly positive — softer input costs help a fixed-price contractor absorbing overruns
2026-08-14Retail Sales MoM (Jul)−0.6%+0.1%−700%Not applicable — no consumer exposure. Recorded so the omission is deliberate, not an oversight
2026-08-14Michigan Consumer Sentiment (Aug)51.054.5−6.4%Not applicable to this name for the same reason

Nothing on the macro calendar is high-impact for this stock inside 14 days. Intuitive Machines is driven by appropriations and by its own mission calendar, not by CPI. The one indirect channel that matters is the 10-year Treasury: at 4.63% it sets the 11.13% discount rate behind the warranted multiple, and a pre-profit company whose value sits almost entirely in cash flows years away is about as rate-sensitive as an equity gets. The 14 August consumer break is recorded and explicitly set aside — this name has no consumer exposure, and the framework is clear that a weak consumer print is not a reason to mark down a business that does not sell to consumers.

9

Multi-Timeframe Technical Analysis

Trend, RSI and breakout status across monthly / weekly / daily / hourly / 15-minute, with a confluence verdict.
TimeframeTrendDirectionRSIMACDKey S/RBreakoutVol
MonthlyUptrend ↑Bullish48.5+3.73, histogram +0.15S: $7.78 · R: $23.32Resistance breakout0.6×
WeeklyUptrend ↑Neutral48.2−1.17, histogram −1.64S: $14.54 · R: $23.32Resistance breakout1.1×
DailyRecovering →Bullish63.0−0.17, histogram +0.99 and risingS: $17.91 / $16.17 · R: $22.50Resistance breakout1.8×
HourlyStrong uptrend ↑Bullish61.4+0.51, histogram −0.04S: $18.81 · R: $19.61Resistance breakout
15-minuteWeakening →Neutral48.1+0.00, histogram −0.02S: $18.71 · R: $19.41Resistance breakout
Confluence: Mostly Bullish · MTF Score 68

The structure is a powerful counter-trend recovery. The daily close of $19.01 has reclaimed the 50-day SMA ($18.58) with the MACD histogram positive for ten consecutive sessions and volume at 1.8× the 20-day average — a clean technical entry trigger on its own terms. The line that has not been reclaimed is the 200-day SMA at $19.31, 1.6% overhead; until that goes, the daily chart is “recovering” rather than an uptrend, and the weekly MACD is still negative with a falling histogram. The wider context is sobering: the stock made an intraday high of $46.75 in the week of 25 May 2026 and an intraday low of $11.29 in the week of 27 July 2026, so this $19.01 is a 68% rally off the low inside a 59% drawdown from the high. Levels that matter: $19.31 (200-day) then $22.50 as resistance; $17.91 then $16.17 as support, with $14.18 — the low of the week of 10 August — below that. All dates and levels here are taken from get_technical_indicators and Yahoo weekly bars, never from Polygon aggregate labels.

10

Price Chart (6-Month Daily)

A 6-month daily close line with SMA50 and key support/resistance — the visual companion to the MTF table.

126 raw unadjusted daily closes, 13 February to 14 August 2026, with the 50-day simple moving average in orange. The shape is the story: an intraday high of $46.75 in late May, an intraday low of $11.29 in the week of 27 July, and a 68% rally back to $19.01 — which has reclaimed the 50-day average at $18.58 but not yet the 200-day at $19.31. Series pulled with auto_adjust=False; the final close matches the stamped price exactly.

11

Scenario Summary

Twelve-month price paths with named triggers and probability weights.

Bull — $30.00 (25%)

IM-3 lands upright and operates a full surface mission — the first clean Nova-C landing, which repairs the one thing the company is judged on. The three-satellite GEO programme converts from a $45.0m authority to proceed into funded contract value, taking soft backlog into firm remaining performance obligations. The 18-spacecraft Golden Dome award scales into a follow-on tranche. FY2026 Adjusted EBITDA turns positive as guided and FY2027 revenue runs toward the $1.31bn top of the consensus range. On that path the market pays about 6× FY2027 revenue, roughly where it pays today on a smaller number. Note what this scenario requires: near-flawless execution on a mission profile that has failed twice.

Base — $14.50 (45%)

Guidance is met — FY2026 revenue lands inside $900m–$1.0bn and Adjusted EBITDA scrapes positive — and the multiple normalises anyway as the market works through what the Q2 filing actually said: organic service revenue −27% year on year, two loss-making flagship contracts, and $947.3m of backlog outside firm performance obligations. Roughly 2.9× FY2027 revenue, which is still more than double the 1.28× the anchor warrants. That gap is deliberate and worth stating: the anchor is a multi-year convergence estimate, while a twelve-month scenario is where the shares plausibly trade next. A stock does not have to reach intrinsic value inside a year, and this one probably will not.

Bear — $9.00 (30%)

A third consecutive landing anomaly on IM-3 — an outcome the record makes uncomfortably plausible — and the “only company to land twice” story becomes “the company that has never landed upright”, with the next CLPS competition going to Firefly. Add a missed Adjusted-EBITDA guide, which needs a swing of more than $22m in the second half against an H1 that ran −$11.1m, and the remaining $261.2m of ATM capacity gets drawn at a falling price. Cash of $367.4m is about thirteen months of the June-quarter burn rate, which is not long enough to be relaxed about a bad print. The stock breaks the $11.29 low of the week of 27 July 2026 and heads toward its 52-week low of $7.78. This is not a tail — at 30% it is the second most likely path here.

Probability-weighted: 0.25×$30.00 + 0.45×$14.50 + 0.30×$9.00 = $16.73, 12.0% below the $19.01 close. That negative expected value, on a name where the framework already reads Medium quality at an Expensive price, is the arithmetic behind a SELL at every horizon. The systemic tail the macro report carries as armed — the S&P 500 concentration / AI earnings-quality unwind, with its trigger currently receding as breadth broadens — is not inherited into the downside here: Intuitive Machines has no AI capex or monetisation leverage, is not an index heavyweight, and its non-operating items reduce rather than inflate reported earnings. Its downside is idiosyncratic and it is quite sufficient on its own.

12

Entry / Exit Rules

Three independent entry paths (Fundamental · Technical · Catalyst) and three exit triggers (Stop-Loss · Thesis · Profit-Target). Any one entry path is a valid entry — the more that agree, the larger the position the conviction ladder suggests. Exits are graded by severity, not count.

How to read this — the Conviction Ladder

The three entry groups are alternative paths to a buy, not a checklist. A group counts only when all its sub-conditions hold. How many groups are satisfied sets the suggested size — it does not gate whether you may enter: 1 group = Half-Size (a valid starter/scale-in), 2 = Full-Size, 3 = Over-Size (highest conviction); 0 = Wait (no path open yet). A strong overall signal can still read Wait here when the stock is well above its entry zones — that flags "good business, no entry edge right now," not a contradiction. Exits are graded by severity of what is live, not by a count: a hard stop is an Exit on its own.
Entry conviction: Half-Size — VOID for new capital1 of 3 groups met — VOID for new capital: an exit is live (exit_action Exit). The count below describes sizing for an entry that is not on offer; it governs an existing position only

Fundamental (it's cheap / supported) — not MET

The price is 2.7× our fair-value estimate. This path is not close to open.
⛔ Price below the $7.00 fair-value estimate — actual $19.01
✅ No earnings inside 7 calendar days — next print 12 November 2026
✅ Underlying-Driver score at or above 50 — actual 70

Technical (trend turned or at support) — MET

All three sub-conditions hold on the 14 August 2026 close. This is the one open path.
✅ Daily close above the 50-day SMA on volume above 1.5× the 20-day average — $19.01 vs SMA50 $18.58, volume 18.9m at 1.8× the 20-day average (Polygon's 20-bar ratio, which is the window the rule specifies; Yahoo's three-month average volume would give 1.4× and fail the test — we state which we used because the answer differs)
✅ RSI(14) between 35 and 65 — actual 63.0, not overbought
✅ MACD histogram positive for at least 2 consecutive sessions — positive for 10, from 3 to 14 August 2026

Catalyst (an event confirms it) — not MET

The Q2 print was strong but the reaction did not clear the thresholds on the day.
⛔ Post-earnings move within 24 hours above +5% — results were released pre-market on 13 August 2026 and the stock closed +3.60%. It added a further 8.26% on 14 August for +12.2% across two sessions, which we note but which is not what the rule tests
✅ Guidance raised or maintained — FY2026 revenue of $900m–$1.0bn and positive Adjusted EBITDA issued with the print
⛔ Volume above 2× the 20-day average on the print — 20.6m shares on 13 August, about 1.6×

Forecast: Fundamental: requires roughly a 63% fall to $7.00, or the fair value to rise to meet the price — which needs a demonstrated path to a mid-teens sustainable net margin. Neither is a 2026 event; unlikely inside twelve months absent a crash. Technical: already met, and the next thing to watch is the 200-day SMA at $19.31, 1.6% overhead — at the recent trajectory that is days, not weeks, though the average true range of $1.65 means a single session settles it either way. Catalyst: event-driven and not time-projectable. Two candidates: the undated IM-3 launch, and Q3 earnings on 12 November 2026. Read the ladder honestly: one path is open, so the mechanical ladder returns Half-Size. That is the builder computing a size from a group count. It is not a recommendation to open a position — the signal is SELL at all three horizons, four hard gates are triggered, and the exit action is Exit. The ladder is published because the framework publishes it, and because an open technical path on a name we rate SELL is exactly the tension a reader deserves to see rather than have smoothed away.

Exit action: Exita hard trigger is live — exit (stops are non-negotiable)

Stop-Loss — not LIVE

Not live. The price is rising, not breaking down.
⛔ Two consecutive daily closes below $13.90 — the last close was $19.01, 37% above it. The level is set deliberately below the $14.50 modal path: an earlier draft put it at $15.80, which our own 45%-probability scenario would have triggered, making the stop and the central case contradict each other. $13.90 sits under the $14.18 low of the week of 10 August 2026, 3.1 average-true-range units below spot. On a SELL-rated name a stop is not a level to defend but the level whose loss confirms the downside path is running; the first warning shot is the $16.17 shelf

Thesis Invalidation — LIVE

LIVE on the catastrophic arm, which fires alone without needing a second condition.
Catastrophic item — a hard gate has triggered. The framework's catastrophic arm lists financial distress, dilution and going concern, and two of our four triggered gates are on that list: Gate 1 (distress, on trailing interest coverage of −10.03× — operating income of −$134.851m over interest expense of $13.447m) and Gate 4 (dilution, on Class A shares outstanding up 39.8% in six months to 169,529,749). Either one fires this group on its own. Gate 3 (valuation ceiling) and Gate 5 (binary event) are also triggered but are deliberately not counted here, because the catastrophic arm does not list them
⛔ Full-year guidance cut — no; guidance was issued and it is for growth
⛔ Revenue growth decelerating below the sector median — no; +310% year on year
⛔ The primary Underlying Driver has turned to a headwind — no; it is a Tailwind at 70

Profit-Target — not LIVE

Not live.
⛔ Price at the median analyst target of $31.00 — actual $19.01
⛔ RSI above 70 — actual 63.0

Forecast: The exit action is Exit, and it is worth being precise about why, because it sits oddly next to a rising share price and an unbroken stop. It is not the stop and it is not a guidance cut. It is the framework's rule that a triggered financial-distress or dilution gate is a catastrophic thesis-invalidation item in its own right. What would clear it: sustained positive free cash flow — which on the company's own guidance is a 2027-2028 event at the earliest — and an end to equity issuance, with about $261.2m of authorised ATM capacity still outstanding. Neither is likely before the 12 November 2026 print.

Imagine you act at the current price of $19.01 · as of 2026-08-16

What if you bought now?

You would be paying 4.56× guided 2026 revenue for a company that has never earned an operating profit, whose organic revenue fell 27% year on year last quarter, and whose two flagship NASA lander contracts are audited loss contracts. Our anchor warrants 1.28×. You would be buying the strongest tape the stock has shown in three months — and the reason the tape is strong is a genuinely excellent quarter. The bet you are making is that roughly 30% compound revenue growth, a mid-teens sustainable margin and no further dilution all arrive together.

What if you sold now?

You would be selling into strength, 33.9% above where our last report rated it a medium-term Buy on 4 August, and 12.2% above where it traded before the Q2 print. You would be giving up the 25% path to $30.00 in which IM-3 lands cleanly and the GEO programme converts. The probability-weighted outcome across all three paths is $16.73, 12.0% below the current price, and the second most likely single path is the $9.00 one. Selling here banks a re-rating the framework does not think the fundamentals have earned.

13

Position Sizing Context

Illustrative portfolio math (not advice) translating conviction into an allocation given risk-per-share and volatility.

No portfolio allocation or role was specified, so no position size is prescribed. Three context figures a reader should have anyway.

MeasureValueWhat it means
Average true range (14-day)$1.65 — 8.7% of the $19.01 priceA single ordinary session moves this stock nearly 9%. Any sizing built on an equity-market-typical 1–2% daily range will be roughly five times too large.
Realised range, last three monthsIntraday high $46.75 (week of 25 May 2026) to intraday low $11.29 (week of 27 July 2026)A 76% peak-to-trough fall and a 68% rally off the low, inside one quarter.
Beta1.829Above the 1.6 threshold that forces the framework's maximum risk add-on, taking the discount rate to 11.13%.

This is a satellite-sized instrument whatever the signal says. The signal says SELL.

14

Calibration Snapshot

Machine-readable snapshot of every score, level and signal, saved alongside the HTML so the next run can compute deltas.
{
  "ticker": "LUNR",
  "date": "2026-08-16",
  "version": "v6",
  "brand": "",
  "analysis_status": "stopped",
  "exchange": "NASDAQ",
  "exchange_ticker": "NASDAQ:LUNR",
  "isin": "US46125A1007",
  "api_ticker": "LUNR",
  "finder_ticker": "LUNR",
  "finder_exchange": "NASDAQ",
  "company": "Intuitive Machines, Inc.",
  "currency": "USD",
  "reporting_currency": "USD",
  "sector": "Industrials",
  "gics_sector": "Industrials",
  "user_context": {
    "horizon": null,
    "allocation_pct": null,
    "portfolio_role": null
  },
  "user_horizon": null,
  "user_allocation_pct": null,
  "portfolio_role": null,
  "price_at_rating": 19.01,
  "price_asof_note": "close of Friday 2026-08-14, the latest print; report dated Sunday 2026-08-16",
  "signal_short": "SELL",
  "signal_medium": "SELL",
  "signal_long": "SELL",
  "primary_signal": "SELL",
  "short_entry_confirmed": true,
  "composite_short": 53,
  "composite_medium": 49,
  "composite_long": 47,
  "quality_score": 48,
  "lifecycle_stage": "high-growth / early-commercial space infrastructure (Industrials, Aerospace & Defense)",
  "quality_detail": {
    "industry_benchmark_name": "ROIC-vs-WACC + Backlog Growth (Industrials)",
    "industry_benchmark_value": "ROIC negative vs an 11.13% cost of capital; backlog $1,762.0m at 30 Jun 2026 vs $213.1m at 31 Dec 2025",
    "industry_benchmark_score": 55,
    "moat_score": 45,
    "moat_detail": {
      "pricing_power": 25,
      "network_effects": 40,
      "switching_costs": 60,
      "cost_advantage": 35,
      "intangible_assets": 65
    },
    "roic_percentile_vs_peers": 18,
    "capital_allocation": 42,
    "management_skin_in_game": 58,
    "revenue_q2_2026_usd_m": 206.168,
    "revenue_growth_q2_2026_yoy_pct": 309.8,
    "revenue_organic_service_q2_2026_usd_m": 36.677,
    "revenue_organic_service_q2_2025_usd_m": 50.313,
    "revenue_organic_service_yoy_pct": -27.1,
    "lanteris_contribution_q2_2026_usd_m": 166.735,
    "revenue_h1_2026_usd_m": 392.898,
    "guidance_fy2026_revenue_usd_m": "900-1000",
    "guidance_fy2026_adj_ebitda": "positive",
    "adj_ebitda_q2_2026_usd_m": -13.799,
    "adj_ebitda_h1_2026_usd_m": -11.132,
    "gross_margin_q2_2026_pct": 17.4,
    "operating_margin_q2_2026_pct": -22.9,
    "backlog_30jun2026_usd_m": 1761.95,
    "backlog_31dec2025_usd_m": 213.07,
    "backlog_acquired_lanteris_usd_m": 612.8,
    "remaining_performance_obligations_usd_m": 814.7,
    "backlog_outside_rpo_usd_m": 947.3,
    "revenue_mix_q2_2026": {
      "commercial_pct": 32,
      "civil_pct": 37,
      "national_security_pct": 30
    },
    "revenue_mix_q2_2025": {
      "commercial_pct": 4,
      "civil_pct": 93,
      "national_security_pct": 3
    },
    "loss_contracts": "IM-3 (a loss contract since 2021) and IM-4 (since Q2 2025); a further $6.3m and $16.2m of losses booked respectively in H1 2026"
  },
  "cash_runway_months_cash_only": 13.1,
  "cash_runway_months_after_aug_acquisitions": 11.9,
  "cash_runway_months_incl_remaining_atm": 21.2,
  "cash_runway_basis": "cash $367.354m at 30 Jun 2026 against Q2 2026 FCF burn of $83.867m (operating -$59.802m, capex -$24.065m); $35.0m of announced August acquisition cash (GBP18.5m, the cash half of the GBP37.0m Goonhilly consideration, at GBP/USD 1.3491 on 2026-08-14 = $25.0m, plus $10.0m for COMSAT; both subject to post-closing true-up); $261.2m of the $500.0m ATM authorisation unused (the authorisation is gross; $235.2m was raised net of $3.6m of commissions = $238.8m gross). Management asserts >=12 months' sufficiency from issuance (10-Q).",
  "cash_30jun2026_usd_m": 367.354,
  "fcf_h1_2026_usd_m": -145.819,
  "debt_definition_used": "face value of the 2.50% convertible notes due Oct 2030 = $345.0m (carrying $336.4m); Yahoo total_debt of $455.2m is lease-inclusive and was NOT used",
  "net_cash_usd_m": 22.4,
  "tangible_book_per_share": -4.22,
  "tangible_book_basis": "total shareholders' deficit -$290.082m less goodwill $379.216m less intangibles $297.069m = -$966.367m, over 228,924,068 economic shares. An earlier draft published -$5.92 on the Class A weighted average - the same share-count error this report exists to correct, in reverse.",
  "book_value_qualification": "the -$290.082m deficit and the -$4.22 TBV/share are PERMANENT equity only; they exclude $1,194.653m of redeemable non-controlling interests carried in MEZZANINE equity at redemption value, marked to the share price (it rose $243.1m in H1 2026 from $951.536m while the business lost money). Book equity is therefore not a clean quality test in either direction for this Up-C structure, and 'positive book value' is deliberately NOT listed as a Quality-upgrade blocker. Book equity is NOT restated as positive: a balance that rises because the share price rose is not a measure of what the business is worth.",
  "atm_authorised_usd_m": 500.0,
  "atm_drawn_gross_usd_m": 238.8,
  "atm_drawn_net_usd_m": 235.2,
  "atm_remaining_usd_m": 261.2,
  "lease_remeasurement_july_2026_usd_m": 109.1,
  "yahoo_total_debt_gap_unexplained_usd_m": 16.8,
  "gbp_usd_rate": 1.3491,
  "gbp_usd_rate_asof": "2026-08-14",
  "gbp_usd_rate_use": "converts the GBP18.5m cash half of the Goonhilly consideration to $25.0m in the runway calculation; no earnings figure is FX-converted (the company reports in USD)",
  "shares_class_a": 173231343,
  "shares_class_b": 0,
  "shares_class_c": 55692725,
  "shares_total_economic": 228924068,
  "shares_asof": "2026-08-06 (10-Q cover page)",
  "enterprise_value_usd_m": 4329.5,
  "enterprise_value_basis": "228,924,068 economic shares x $19.01 = $4,351.8m equity + $345.0m convertible-note face - $367.354m cash",
  "enterprise_value_yahoo_usd_m": 4583.1,
  "enterprise_value_yahoo_note": "HIGHER than ours - Yahoo adds the redeemable NCI at its $1,194.7m book carrying value rather than marking Class C to market, so it does capture the Class C economics on the EV line. On the Yahoo EV the multiple is 4.82x and the ratio 3.77x - MORE expensive. The FMP $3.03bn market cap implies 159.37m shares, stale even for Class A, understating our $4,351.8m equity by ~30% (the ~24% figure an earlier draft used is the Class C share OF the economic count, a different quantity).",
  "valuation_score": 38,
  "valuation_detail": {
    "fcf_yield": "negative - not meaningful",
    "ev_sales_ttm": 8.83,
    "ev_sales_fy2026e_guidance_mid": 4.56,
    "ev_sales_fy2026e_consensus": 4.63,
    "ev_sales_fy2027e_consensus": 3.76,
    "ev_backlog": 2.46,
    "ev_rpo": 5.31,
    "price_vs_consensus_pct": -40,
    "historical_valuation_decile": "3rd-4th decile of its own range; 29th percentile of the 52-week $7.78-$46.75 band",
    "prior_report_multiple_like_for_like": 3.53,
    "prior_report_multiple_as_published": 2.6,
    "sensitivity_g10_ratio": 3.56,
    "sensitivity_g15_ratio": 2.91,
    "sensitivity_g20_ratio": 2.61,
    "margin_needed_for_ratio_140_at_1689x_pct": 19.3,
    "margin_needed_for_ratio_140_at_23x_pct": 14.2,
    "guardrail_adjudicated_by_hand": "Industrials P/E line 23x. Share-count-corrected FY2028E P/E is ~93x (net income $47.0m / 228,924,068 shares = $0.205 EPS at $19.01). Breached on every honest construction. The automated guardrail arm does not run because the declared basis is EV/Sales, not P/E."
  },
  "warranted_multiple": 1.28,
  "actual_multiple": 4.56,
  "val_multiple_basis": "EV/Sales on FY2026E revenue (company guidance midpoint $950m). An earnings multiple does not resolve: trailing EPS is -$0.94 and FY2027E consensus EPS is $0.014. Derivation of the warranted figure is recorded separately in warranted_multiple_derivation - deliberately kept out of this field, because a basis string naming both an EV/Sales actual and an earnings multiple made the linter compare 4.56 against the 23x Industrials earnings guardrail and pass vacuously (found by the independent auditor, 2026-08-17).",
  "warranted_multiple_derivation": "warranted EV/Sales 1.28x = the two-stage warranted earnings multiple of 16.89x (r 11.13%, g_near 10.0%, g_term 3.0%) multiplied by a 7.6% sustainable net margin. Price/sales converts to EV/sales here because net debt is about -$22m, ~0.5% of equity value. The 23x Industrials guardrail is adjudicated BY HAND in valuation_detail.guardrail_adjudicated_by_hand and in the report body, because the automated arm cannot run on a non-earnings basis; a skipped check is not a passed one.",
  "sustainable_net_margin_pct": 7.6,
  "sustainable_net_margin_source": "sell-side FY2030E net income $154.0m on revenue $2,032.7m = 7.58%; corroborated by aerospace primes at 7-8% net",
  "warranted_pe_two_stage": 16.89,
  "discount_rate_r": 11.13,
  "risk_free_10y": 4.63,
  "risk_free_source": "FRED DGS10 at 2026-08-13 (NOT the macro report, whose prose carries 4.70%; at 4.70% warranted = 1.26x and ratio = 3.62x, so the band is unchanged by it)",
  "risk_addon_pct": 2.0,
  "risk_addon_basis": "beta 1.829 > 1.6 forces the maximum add-on",
  "g_near": 10.0,
  "g_term": 3.0,
  "g_near_basis": "consensus FY2026-FY2030 revenue CAGR 21.4%, haircut 25% to 16.1%, then capped at the Industrials (cyclical/normal) 10% sector-achievable line",
  "warranted_ratio": 3.56,
  "val_band": "expensive",
  "sector_guardrail_line_pe": 23.0,
  "timing_score": 62,
  "timing_detail": {
    "mtf_confluence": "mostly bullish",
    "mtf_trend_score": 68,
    "risk_reward_score": 53,
    "relative_strength_vs_spy": "1-month leader (+40.6% vs SPY +4.5%), 3-month laggard (-43.9% vs SPY +5.0%)",
    "relative_strength_vs_sector": "1-month +40.6% vs XLI +4.0% and ITA +9.8%; 3-month -43.9% vs XLI +8.8% and ITA +16.5%",
    "catalyst_clustering_score": 55,
    "sentiment_score": 62,
    "macro_overlay_score": 70,
    "dynamic_macro_weight": 0.15,
    "atr_usd": 1.65,
    "atr_pct_of_price": 8.7,
    "range_position_52w_pct": 29,
    "sma50": 18.58,
    "sma200": 19.31,
    "rsi14": 63.0,
    "volume_ratio_20day": 1.8,
    "post_earnings_move_24h_pct": 3.6,
    "post_earnings_move_2session_pct": 12.2
  },
  "relative_strength_vs_spy": "1-month leader (+40.6% vs SPY +4.5%), 3-month laggard (-43.9% vs SPY +5.0%)",
  "relative_strength_vs_sector": "1-month +40.6% vs XLI +4.0% and ITA +9.8%; 3-month -43.9% vs XLI +8.8% and ITA +16.5%",
  "driver_score": 70,
  "driver_name": "US government space & missile-defence appropriations (NASA CLPS / Moonbase / Artemis / Near Space Network + Space Force Golden Dome)",
  "driver_commodity_trend": "n/a - not commodity-leveraged",
  "driver_amplifies": "Amplification not applied. A SELL intensifies to STRONG SELL only when the driver is <=35 and economic pressure is a Headwind; here the driver is 70 (Tailwind) and pressure is a Tailwind, so the two context pillars disagree with the fundamentals and the SELL stands unamplified - recorded as a caveat, per the framework.",
  "economic_alignment_stance": "Trend-Following",
  "economic_alignment_conviction": 78,
  "economic_alignment_pressure": "Tailwind",
  "economic_alignment_source": "sector-map XLI (O/O/SO) - LUNR is not in the macro report's watchlist_forecast",
  "macro_report_date": "2026-08-12",
  "consumer_break_applicability": "not applicable - no US consumer exposure; the 14 Aug retail-sales and Michigan prints are recorded in \u00a78 and explicitly set aside",
  "competitive_share_trajectory": "mixed",
  "competitive_threat_level": "elevated",
  "competitors_named": "Firefly Aerospace (Blue Ghost landed upright March 2025), Astrobotic, Blue Origin, SpaceX (supplier and rival), Lockheed Martin / Northrop Grumman / L3Harris / Boeing in national-security space, Airbus DS / Thales Alenia / Rocket Lab / MDA Space in satellites",
  "nonop_pct_of_net_income": 25.0,
  "nonop_direction": "DRAG, not inflator - Q2 2026 other income/(expense) net of -$15.697m (warrant fair value -$11.622m, contingent consideration -$0.890m, net interest -$3.007m) against a net loss of -$62.841m. There is no earnings inflation to strip; step 7b runs the other way here. The material Up-C distortion is on market capitalisation, not EPS: providers count Class A only and understate the equity by ~24%.",
  "clean_pe": "n/a - loss-making, no positive earnings to clean",
  "clean_peg": "n/a - loss-making",
  "eps_trailing": -0.94,
  "trailing_pe": -20.24,
  "eps_trailing_basis": "TTM Class A basic and diluted: Q3 2025 -$0.0594, Q4 2025 -$0.34, Q1 2026 -$0.25, Q2 2026 -$0.29. Recorded so the currency-coherence check can run; a negative P/E is not a valuation multiple.",
  "analyst_consensus_target": 31.67,
  "analyst_target_high": 41,
  "analyst_target_low": 26,
  "analyst_target_median": 31,
  "analyst_target_upside_pct": 67,
  "analyst_targets_caveat": "the same panel's EPS estimates imply ~143m shares vs 228,924,068 outstanding, so the targets are discounted in scoring",
  "analyst_grades_consensus": "Buy",
  "analyst_bullish_pct": 82,
  "analyst_coverage_count": 11,
  "fmp_rating": "C",
  "fmp_overall_score": 2,
  "recent_upgrades_30d": 0,
  "recent_downgrades_30d": 0,
  "overall_confidence": 45,
  "quality_confidence": 45,
  "valuation_confidence": 60,
  "timing_confidence": 70,
  "fair_value_est": 7.0,
  "fair_value_basis": "two constructions: warranted EV/Sales 1.28x on $950m FY2026E revenue gives $5.41; warranted P/E 16.89x on FY2030E net income of $154.0m across 228,924,068 shares, discounted 4.4 years at 11.13%, gives $7.14. We publish the more generous of the two.",
  "stop_loss": 13.9,
  "target_price": 14.5,
  "stop_loss_basis": "set deliberately BELOW the $14.50 modal base path, under the $14.18 low of the week of 2026-08-10, 3.1 ATR below spot. An earlier draft published $15.80, which this report's own 45%-probability scenario would have triggered - the stop and the central case contradicted each other. The $16.17 shelf is the first warning level.",
  "scenario_base_target": 14.5,
  "scenario_bull_target": 30.0,
  "scenario_bear_target": 9.0,
  "scenario_probabilities": {
    "bull_pct": 25,
    "base_pct": 45,
    "bear_pct": 30
  },
  "scenario_weighted_target": 16.73,
  "scenario_weighted_vs_price_pct": -12.0,
  "systemic_tail_inherited": false,
  "systemic_tail_note": "the armed S&P 500 concentration / AI earnings-quality unwind is NOT inherited: no AI capex or monetisation leverage, not an index heavyweight, and non-operating items reduce rather than inflate reported earnings. Its trigger is also receding as breadth broadens.",
  "entry_groups_met": 1,
  "entry_conviction": "Half-Size",
  "entry_group_open": "Technical (was Fundamental on 4 August)",
  "entry_ladder_note": "the Half-Size badge is the builder's mechanical output from one met entry group. It is NOT an entry recommendation: the signal is SELL at all three horizons, four hard gates are triggered (1, 3, 4 and 5) and the exit action is Exit.",
  "exit_groups_live": 1,
  "exit_action": "Exit",
  "exit_action_basis": "Thesis Invalidation is live on its catastrophic arm - a triggered hard gate fires the group alone. Gates 1 and 4 are triggered. No stop-loss has been hit and no guidance has been cut.",
  "hard_gate_state": "caution",
  "gates_triggered": [
    "Gate 1 - Financial Distress: TRIGGERED on the interest-coverage arm ALONE. TTM operating income -$134.851m / interest expense $13.447m = -10.03x, against a +1.5x threshold. The four-consecutive-quarters-of-negative-FCF arm is NOT relied upon: Q1 2025 FCF was POSITIVE at about +$13.3m (H1 2025 -$14.020m less Q2 2025 -$27.317m) and the Q3/Q4 2025 cash-flow statements were not re-read this run, so that arm is unproven. Verified negative quarters: Q2 2025 -$27.3m, Q1 2026 -$62.0m, Q2 2026 -$83.9m. Classic distress markers absent (net cash ~$22m, current ratio 1.66x, no near-term maturity, no going-concern language).",
    "Gate 3 - Valuation Ceiling: Expensive band, actual EV/Sales 4.56x vs warranted 1.28x = 3.56x, against a 1.40x threshold. No growth exception. Cross-check on the Yahoo EV construction: 4.82x / 3.77x, more expensive still.",
    "Gate 4 - Accounting/Dilution: two-year test met on weighted-average Class A shares 55,093,365 (Q2 2024) -> 117,434,775 (Q2 2025) -> 162,172,470 (Q2 2026). Balance-sheet check: Class A OUTSTANDING 121,281,880 (31 Dec 2025) -> 169,529,749 (30 Jun 2026) = +39.8% in six months; ISSUED 123,472,960 -> 171,720,829 = +39.1% (2,191,080 treasury shares held flat across both dates). 8,259,379 shares sold via a new $500.0m ATM in Q2 2026 for $235.2m net of $3.6m commissions, plus an 11,574,069-share placement at $15.12 in February, 22,991,028 shares for Lanteris and 960,649 for Goonhilly in August. SBC at 4.9% of H1 2026 revenue is NOT the trigger.",
    "Gate 5 - Binary Event: TRIGGERED. The third limb (a genuinely binary outcome that would move the stock >20%) carries NO timing qualifier, unlike Gate 2 with its explicit 14-day window. IM-3 is confirmed by primary sources as the next launch and confirmed unflown; a landing after two consecutive tip-overs is binary and would move the stock >20%. Having applied strict literalism to Gates 1 and 4, we do not read a timing window into Gate 5 that is not written there. No dated launch window is public (NASA CLPS page updated 2026-07-31 gives only 2026) - recorded as a confirmed absence, not an assumption. Caps at HOLD; no signal effect, since the matrix already read SELL."
  ],
  "gates_caution": [
    "Do-Not-Buy Trigger 4 - insider selling: could not be established from Form 4 / Rule 144 metadata (Form 4 filings 1, 15 and 29 Jul and 12 Aug 2026; Rule 144 notices 8, 18 and 29 Jun 2026). Not fired; carried as an unresolved monitoring item, not a pass."
  ],
  "do_not_buy_triggers": [],
  "dnb_adjudication": "RULING: Trigger 2(a) is ruled OUT. The numeric arm IS met (3.56x against a 2.0x threshold), so the ruling rests entirely on the no-exceptional-proven-durable-growth carve-out, and we rule that the carve-out APPLIES. Reasons, in order: (1) the growth is exceptional and delivered rather than promised - revenue +310% YoY, backlog $1.76bn at ~1.85x guided FY2026 revenue, $920m of Q2 awards plus $300m in Q3 to date, a sixth CLPS award, national security 3% -> 30% of mix; (2) internal consistency with the framework - Gate 3 own pre-profit arm is written as EV/Revenue > 20x for a NON-hypergrowth company, and this name is at 4.56x forward on growth of that order, so the framework own hypergrowth allowance points the same way; (3) the remaining-performance-obligation schedule gives contractual revenue visibility into 2027 and beyond, which supports durable. Firing 2(a) would output a hard DO NOT BUY on a company whose revenue quadrupled, which the carve-out exists to prevent. The durability caveat is carried as a MONITORING ITEM, not as a qualification of the ruling: $947.3m of backlog sits outside firm RPO and much of the growth was acquired; if either deteriorates, 2(a) is re-adjudicated. Trigger 2(b) requires a live de-rating catalyst - no AI-cohort membership (no AI capex or monetisation leverage, not an index heavyweight, non-operating items a drag not an inflator) and no structural business-model threat identified - so it does not fire. Triggers 1, 3 and 5 do not fire; Trigger 4 is unresolved and monitored.",
  "framework_amendment_proposed": "Gate 1's interest-coverage arm (&lt;1.5&times;) and its four-quarters-of-negative-free-cash-flow arm both fire mechanically on ANY pre-profit company, and Gate 4's dilution arm fires on any company funding growth with equity &mdash; which is the normal and appropriate financing structure for early-commercial capital-intensive names. On LUNR both fire while the classic distress markers are absent: a small net cash position, a current ratio of 1.66&times;, no near-term maturity and no going-concern language. We let Gates 1 and 4 fire, because this SKILL writes exemptions where it wants them (Gate 3 and Do-Not-Buy Trigger 2 both carry a spelled-out pending-takeover carve-out), so an absent carve-out is not an invitation to invent one. PROPOSED AMENDMENT, for the framework owner rather than for this report: (a) add a pre-profit qualifier to Gate 1's interest-coverage arm that reads months-of-runway against burn instead, with a threshold near 12 months; (b) make Gate 4's dilution arm test dilution per unit of revenue or backlog growth rather than raw share count, so accretive equity funding is distinguished from value-destroying issuance; (c) state explicitly whether a mechanically-fired distress gate should also fire the catastrophic Thesis-Invalidation exit item, since on this name that single linkage is what turns exit_action from Hold to Exit; (d) specify how the warranted-multiple anchor should be instantiated for a pre-profit name, because warranted EV/Sales = warranted P/E &times; sustainable net margin is a construction we have had to derive rather than read off. This is the same amendment raised on MNO.TO this run. CLAUSE (e), added after the independent audit: Gate 5 carries NO timing qualifier while Gate 2 carries an explicit 14-day window. On this name that asymmetry decided whether an imminent-but-undated binary mission fired the gate or merely cautioned it. We resolved it by firing the gate - literalism, consistent with our treatment of Gates 1 and 4 - but the framework should state explicitly whether Gate 5 has a scheduling window and, if so, what it is.",
  "post_audit_fixes_round1": "Audit sustained the SELL x3, all four triggered gates and the DNB 2(a) ruling; no blocker. Seven MAJOR fixes, none of which changed a score or a signal: TBV/share -5.92 -> -4.22 on the economic share count; book-value claims qualified for the $1,194.653m mezzanine redeemable NCI; the loss-contract deterioration framing withdrawn (combined additions $22.5m H1 2026 vs $21.2m H1 2025 - flat; IM-3 fell 18.5 -> 6.3, IM-4 rose 2.7 -> 16.2); Gate 1 four-quarter FCF arm DELETED (Q1 2025 FCF was +$13.3m), leaving interest coverage to carry the gate alone; Gate 4 issued-vs-outstanding corrected (+39.8% outstanding, not +43%); the provider-EV claim reconciled (Yahoo EV $4,583.1m is ABOVE ours and implies 4.82x / 3.77x, more expensive; understatement ~30% not ~24%); Gate 5 now FIRES on its literal text, which carries no timing window, removing an interpretive asymmetry against Gates 1 and 4 - the point is added to framework_amendment_proposed as clause (e). Nine MINOR fixes including stop_loss 15.80 -> 13.90 (it sat above the 45%-probability modal path), ATM remaining 264.8 -> 261.2, the $109.1m July lease remeasurement disclosed, the ~$16.8m Yahoo debt gap flagged unexplained, the 24-May intraday-high date aligned to the week of 25 May, XLB (O/SO/SO) acknowledged as the strongest sector row rather than XLI, Lanteris consideration $853.3m, and the FRED-vs-macro risk-free deviation declared (market_snapshot is absent from the 12 Aug state file altogether; 4.70% would make the name MORE expensive).",
  "delta_vs_prior": "Signals flip to SELL at all three horizons, from HOLD / BUY / STRONG_BUY on 4 August, and analysis_status moves from donatien-pick to stopped because no horizon now carries a BUY. The valuation score fell 72 -> 38: the 4 August calibration recorded the warranted-multiple anchor as 'na' while asserting an 'attractive' band, and it computed EV/Sales on a Class-A-only enterprise value; correcting the share count to all 228,924,068 economic shares and running the anchor properly gives 4.56x actual against 1.28x warranted, a ratio of 3.56x, which is Expensive and fires Gate 3. The quality score fell 60 -> 48, driven by the Q2 2026 filing's disclosure that organic service revenue declined 27% year on year, and by the moat score falling 53 -> 45 once the audited loss-contract status of IM-3 and IM-4 is reflected in pricing power and cost advantage; the industry benchmark score fell 58 -> 55. The timing score rose 44 -> 62 on the post-print breakout. The driver score rose 68 -> 70 on the Golden Dome AMDT3 award and a sixth CLPS win. Gates 1, 3, 4 and 5 are all newly triggered - the 4 August calibration had an empty gates_triggered array and carried the IM-3 mission as a caution. Because only distress and dilution are on the catastrophic Thesis-Invalidation list, Gates 1 and 4 are what move exit_action from Hold to Exit. short_entry_confirmed moves from false to true, because the Technical entry group is now met. entry_groups_met is 1 for a second consecutive run and entry_conviction is Half-Size for a second consecutive run, but the open path swapped from Fundamental to Technical: the price is no longer below our fair value, and the tape has turned. Price +33.9% since 4 August, which is itself most of the reason the signal changed.",
  "next_update_date": "2026-08-30",
  "next_update_basis": "default +14d from the 2026-08-16 run date - no dated catalyst inside the window; the IM-3 launch is imminent but undated, so it is re-checked each cycle. Q3 earnings 2026-11-12.",
  "next_check_date": "2026-08-30"
}

Every score, level and gate above, machine-readable, so the next run computes deltas against it rather than against memory. Note in particular the corrected share count, the recorded warranted multiple where three prior runs recorded 'na', and the framework_amendment_proposed note on Gates 1 and 4.

15

Data Sources & Methodology

Audit trail of every data source: fully available (✓), fallback (⚠), or failed (✗), plus provenance-based confidence haircuts.
Data Source Status
SEC 8-K + Exhibit 99.1 (filed 13 Aug 2026, accession 0001628280-26-056476) The Q2 2026 earnings release, read in full from EDGAR. Source for revenue, backlog, cash, Adjusted EBITDA, free cash flow, guidance and every award claim. PRIMARY SOURCE.
SEC Form 10-Q for the quarter ended 30 Jun 2026 (filed 13 Aug 2026) Read in full. Source for the loss-contract disclosure on IM-3 and IM-4, remaining performance obligations of $814.7m, the revenue mix by customer type, the ATM programme, the convertible-note terms, the subsequent-events note and the cover-page share counts. PRIMARY SOURCE.
SEC 8-K + Exhibit 99.1 (filed 3 Aug 2026, accession 0001193125-26-330778) Goonhilly and COMSAT completion. Also the primary evidence that IM-3 had NOT yet flown as of 3 August: it names “the upcoming IM-3 and Altus-1 missions”.
10-Q cover page, share counts as of 6 Aug 2026 173,231,343 Class A, 0 Class B, 55,692,725 Class C = 228,924,068 economic shares. This is the figure that corrects the enterprise value.
Mission-manifest / Gate-5 check — Gate 5 TRIGGERED; CONFIRMED ABSENCE of a dated launch window Established this run from primary sources, per the Live-Verify discipline. Checked: (1) the 3 Aug 2026 8-K, which calls IM-3 “upcoming”; (2) the 10-Q of 13 Aug 2026, which says IM-3 “readies as the Company's next launch” and gives a period of performance running through March 2027; (3) NASA's own CLPS event page for the IM-3 flight, last updated 31 July 2026, which states the launch window only as “2026”. Result: IM-3 is confirmed as the next launch and confirmed as not yet flown, and no dated launch window is public — that absence is confirmed, not assumed. What the confirmed absence covers is the window, not the gate: Gate 5 is TRIGGERED, on a third limb that carries no timing qualifier at all. An earlier draft of this report drew the opposite conclusion from the same three sources — “no dated window, therefore caution” — which was the very inference the independent audit rejected. We also explicitly reject a third-party aggregator claim that IM-3 launched on 1 June 2026 — it is contradicted by both company filings above.
Price — Yahoo quote and get_technical_indicators $19.01 is the close of Friday 14 August 2026 and is confirmed by three independent reads (Yahoo quote, get_technical_indicators dated 2026-08-14, and the multi-timeframe daily series). Every dated claim, swing high and swing low in this report comes from get_technical_indicators or Yahoo weekly bars. No date is taken from get_stock_prices, whose Polygon aggregate labels are known to run one session early.
FMP get_income_statement — USED WITH ONE FIELD OVERRIDDEN Revenue, operating income and share counts reconcile to the release. But FMP reports Q2 2026 depreciation and amortisation of $75.3m where the company's own statement of operations says $14.9m. That error makes FMP's derived Q2 EBITDA of +$16.9m wrong; the company's Adjusted EBITDA was −$13.8m. We use the release throughout.
Provider enterprise values — RECONCILED, not simply rejected FMP reports enterpriseValueTTM of $3.09bn on a market capitalisation of $3.03bn; that capitalisation implies 159.37m shares at $19.01, stale even as a Class A count, and it understates our $4,351.8m of equity by about 30%. REJECTED. Yahoo reports enterprise_value of $4,583,148,032, which is above our figure because it adds the redeemable non-controlling interests at their $1,194.7m book carrying value instead of marking the Class C shares to market — so Yahoo does capture the Class C economics on that line, at book. We compute our own from the 10-Q cover-page counts: 228,924,068 × $19.01 + $345.0m − $367.4m = $4,329.5m, and cross-check against Yahoo: EV/Sales 4.82×, ratio 3.77× — more expensive than our own 4.56× / 3.56×, so the choice of construction is not flattering the conclusion.
Debt definition used (trap 3c) We use the $345.0m face value of the 2.50% convertible notes due October 2030 (carrying value $336.4m). We do not use Yahoo's total_debt of $455.2m, which is lease-inclusive. Reconciling the gap explicitly: $455.2m less $345.0m of notes less $93.4m of operating and finance lease liabilities ($25.1m current plus $68.3m non-current operating, plus $0.05m finance) leaves about $16.8m we cannot account for from the balance sheet — flagged as unexplained rather than assumed away. Using the lease-inclusive figure would have overstated enterprise value by roughly 2.5% and understated net cash. One item bears directly on this adjudication and belongs on the record: in July 2026 the company amended its Palo Alto lease, extending the term by 17 years to July 2043, and remeasured the right-of-use asset and lease liability to about $109.1m. That is a post-quarter obligation roughly the size of the entire current lease book. It does not enter our enterprise value, because the framework strikes the multiple on financial debt — but a reader comparing our $345.0m against a lease-inclusive provider figure should know the lease side is about to roughly double.
FMP get_analyst_estimates — USED FOR REVENUE, CORRECTED FOR EPS Revenue estimates are used as published (FY2026 $935.1m from 5 analysts, FY2027 $1,150.0m from 4, FY2030 $2,032.7m from 1). The EPS estimates are internally inconsistent with the net-income estimates: FY2028 net income of $47.0m against EPS of $0.329 implies about 143m shares, versus 228,924,068 outstanding. We recompute per-share figures on the true count. FMP's forward ebit and ebitda estimates are also incoherent (deeply negative EBIT alongside positive net income) and are not used.
get_price_target_consensus + get_yahoo_analyst_targets FMP: high $41, low $26, median $31, consensus $31.67 — a dispersed, non-degenerate panel, so no fallback was required; Yahoo cross-checked at mean $29.50, median $29.00, high $43, low $11, 8 analysts. Both are discounted in the scoring because the same panel's per-share estimates carry the share-count error above.
get_grades_consensus + get_stock_grades 9 buy, 1 hold, 1 sell (81.8% bullish). Zero upgrades and zero downgrades in the last 30 days; the most recent action of any kind was a Roth Capital maintain on 28 May 2026. Treated as a stale all-maintain reading.
get_ratings_snapshot C, overall score 2 of 5, with 1-of-5 sub-scores on discounted cash flow, return on assets, debt/equity, price/earnings and price/book. This diverges sharply from the analyst panel and corroborates our Quality and Valuation scores rather than theirs.
get_multi_timeframe_analysis + get_technical_indicators All five timeframes returned. Daily bar dated 2026-08-14, close $19.01, SMA50 $18.58, SMA200 $19.31, RSI 63.0, ATR $1.65, volume ratio 1.8× the 20-day average.
Relative strength — Yahoo weekly bars for SPY, XLI, ITA One month: LUNR +40.6% vs SPY +4.5%, XLI +4.0%, ITA +9.8%. Three months: LUNR −43.9% vs SPY +5.0%, XLI +8.8%, ITA +16.5%. Total-return series are used for these performance comparisons only, never for a price-versus-moving-average test.
Chart series — yfinance daily closes, auto_adjust=False 126 raw unadjusted sessions from 13 February to 14 August 2026, with a 50-day simple moving average computed from the same raw series. Final close $19.01 matches the stamped price exactly. LUNR pays no dividend, so adjustment would be inert here, but the raw series is used as the framework requires.
Risk-free rate — DECLARED DEVIATION from the SKILL The SKILL says to read the 10-year from the macro report's market_snapshot.UST10Y. That key is absent from MacroDriver-state-20260812.json entirely (verified this run), so the instruction cannot be satisfied from the state file, and the macro report's prose carries 4.70% rather than the 4.63% used across this batch. We use 4.63% from FRED series DGS10 at 13 August 2026 and attribute it to FRED. Mitigation: 4.70% moves the warranted multiple to 1.26× and the ratio to 3.62× — more expensive — so the deviation cannot be doing the conclusion any favours.
Macro report MacroDriver-state-20260812.json Run date 12 August 2026. XLI Outperform / Outperform / Strong Outperform. LUNR is NOT in the report's watchlist_forecast, so the sector map is used, as the framework directs.
get_economic_calendar (US, high impact) Next 30 days and last 7 pulled. Nothing high-impact for this specific name inside 14 days; the 14 August consumer prints are recorded and explicitly set aside as inapplicable.
get_earnings_calendar Next report 12 November 2026, consensus EPS −$0.065 on revenue of $243.4m. 88 days out, so Gate 2 is clear.
Insider activity — INCONCLUSIVE, and recorded as such EDGAR shows Form 4 filings on 1, 15 and 29 July and 12 August 2026 and Rule 144 notices on 8, 18 and 29 June 2026. We could not establish that three or more C-suite officers each disposed of more than 25% of their holdings within 60 days outside a 10b5-1 plan, which is what Do-Not-Buy Trigger 4 requires, so the trigger is not fired. This is an unresolved monitoring item, not a clean result.
Competitor evidence (step 7c) Firefly's Blue Ghost Mission 1 upright landing (March 2025), the Astrobotic Peregrine failure (January 2024) and the IM-1 and IM-2 tip-overs are drawn from the public mission record and from the company's own disclosure of its landing history. The MDA Space $620m US-defence acquisition is from a June 2026 news item. Market-share percentages are not published for CLPS task orders, so the share trajectory is a qualitative read from the award and mission record rather than a measured number.
Impact on scores: Confidence: 45% overall — the weakest link is Quality at 45%, and it deserves to be weak. Two acquisitions closed inside eight months mean there is no clean comparable period, no peer whose reported numbers are structurally similar, and only two analysts modelling 2028. The Valuation pillar is at 60%: the multiple itself is solid (guided revenue, corrected share count, primary-source cash and debt), but the warranted figure rests on one discretionary input, the 7.6% sustainable net margin, which is why §4 publishes the full two-way sensitivity rather than a single number. Timing is the strongest at 70%.

Layer-1 author self-audit — what we found in our own work. (1) Corrected a defect in our own prior reports: the 3 July, 20 July and 4 August calibrations all recorded warranted_multiple: "na" and warranted_ratio: "na" while simultaneously asserting val_band: "attractive" — a band with no anchor behind it — and computed EV/Sales on a Class-A-only enterprise value. Both are fixed here, and the fix is most of the 34-point fall in the Valuation score. (2) Surfaced a material fact three prior reports missed: IM-3 and IM-4 are audited loss contracts, with $22.5m of further losses booked on them in H1 2026. That is the largest single input to the Quality and moat downgrades. (3) The uncomfortable tension we have not smoothed away: the conviction ladder returns Half-Size on a met Technical entry group while the signal is SELL at all three horizons and the exit action is Exit. Both are correct outputs of their own rules; §12 says so explicitly rather than hiding one of them. (4) Cash runway is the number that matters most here and it is stated three ways rather than blended: about 13 months on the 30 June cash of $367.4m at the June-quarter free-cash-flow burn of $83.9m; about 12 months after the $35.0m of announced August acquisition cash (£18.5m — the cash half of the £37.0m Goonhilly consideration — at GBP/USD 1.3491 on 14 August 2026 = $25.0m, plus $10.0m for COMSAT; both subject to post-closing true-up); and about 21 months if the $261.2m of remaining ATM capacity is drawn — which is dilution, not free cash. Management's own assertion in the 10-Q is sufficiency for at least twelve months from issuance. The falsifiable tension: H1 Adjusted EBITDA was −$11.1m against a full-year guide of positive, so the second half needs a swing of more than $22m. (5) The absence of a dated IM-3 launch window is confirmed, not assumed — three sources named above. Gate 5 fires regardless, on a third limb that carries no timing qualifier. (6) Unresolved: the insider-selling test could not be settled from Form 4 metadata alone.

Corrections made after the independent audit of this report (round 1). The audit sustained the verdict, all four triggered gates and the Do-Not-Buy ruling, and found no blocker; it found seven material defects, every one of which is fixed above and none of which changed a score or a signal. (a) Tangible book value per share was published at −$5.92 on a Class A weighted average; corrected to −$4.22 on all 228,924,068 economic shares — our own flagship correction, made in reverse. (b) The Up-C distortion had been corrected on the market-capitalisation side but not the book-value side; §3 now states that the −$290.1m deficit is permanent equity and excludes $1,194.7m of redeemable non-controlling interests held in mezzanine at a mark-to-market redemption value, and “positive book value” is no longer listed as a Quality blocker. (c) The loss-contract passage said cost estimates “rose again”; the filing shows combined additions of $22.5m in H1 2026 against $21.2m in H1 2025 — flat — with IM-3 falling from $18.5m to $6.3m while IM-4 rose from $2.7m to $16.2m. The level re-scores pricing power; the deterioration framing was unsupported and is withdrawn. (d) Gate 1 had asserted the four-quarter free-cash-flow arm. Q1 2025 was positive at about +$13.3m and Q3/Q4 2025 were unread, so that arm is deleted; Gate 1 now rests on interest coverage alone, which carries it unaided at −10.03×. (e) Gate 4 quoted 171,720,829 as shares outstanding when that is the issued figure; outstanding was 169,529,749 against 121,281,880 at 31 December 2025, so six-month growth is +39.8%, not +43%. (f) The claim that every provider ignores the Class C was wrong — Yahoo captures it at book, above our own figure — and the understatement is about 30%, not 24%. Reconciled, and the Yahoo construction is shown to be the more punitive of the two. (g) Gate 5 now fires. Its text carries no timing window, unlike Gate 2; recording it as caution while applying strict literalism to Gates 1 and 4 was an interpretive asymmetry inside one section. No signal effect — it caps at HOLD and the matrix already read SELL. Nine minor items were also fixed, of which two mattered: the published stop of $15.80 sat above the $14.50 modal scenario, so our own central case would have triggered our own stop — moved to $13.90; and remaining ATM capacity is $261.2m, not $265m, because the $500.0m authorisation is gross while the $235.2m raised was net of $3.6m of commissions.

Two Layer-0 findings, recorded because they were failures of the checker rather than of the report. First, our earlier “clean” lint was partly vacuous on the check that mattered most here: val_multiple_basis mentioned the warranted earnings multiple in its derivation, the guardrail arm matched on that substring, and it then compared our EV/Sales of 4.56 against the 23× Industrials earnings line and passed. The derivation now lives in a separate calibration field so that arm skips honestly and we adjudicate the 23× line by hand instead. Second, the check that should have caught our three prior calibrations — a real band asserted over an uncomputed anchor — was gated on the inverse condition and returned zero errors against all three files.
DISCLAIMER: This is a quantitative framework for educational purposes only. It is not financial advice. Always do your own research and consult a licensed financial advisor before making investment decisions.