Artemis 3–5 · Gateway · World Programs — First-Principles Audit (2026-07-09)

1. Executive snapshot (what is really going on)

Artemis II

DONE — launched 1 Apr 2026, ~10-day crewed lunar flyby

Crew: Wiseman, Glover, Koch, Hansen (CSA). Orion + SLS human-rated on the real trajectory. Foundation for everything below.

Artemis III

NOT a landing — LEO HLS docking demo, NET 2027

Crew named Jun 2026: Bresnik (CDR), Parmitano ESA (PLT), Douglas, Rubio. Apollo-9 analogue for commercial landers.

Gateway

PAUSED 24 Mar 2026 — not flying as NRHO station

Official language: “paused in its current form.” Hardware real; architecture abandoned. Pivot to ~$20B surface Moon Base.

One-sentence truth The program you remember (Orion → Gateway in NRHO → lander down to south pole) is dead as a critical-path architecture. The program that exists is: LEO lander dress-rehearsal (A3) → direct lunar-orbit landing without Gateway (A4/A5) → build surface base with money and some hardware that used to be Gateway. NASA.gov still shows mixed legacy copy; the controlling acts are Feb 27 architecture reshuffle + March 24 “Ignition” / national space policy event.

source-backed Artemis III is no longer the first landing. NASA mission page and Jun 9, 2026 crew release: LEO rendezvous/docking with SpaceX and/or Blue Origin lander pathfinders; first south-pole landing reassigned to Artemis IV, early 2028.

source-backed Gateway as a lunar-orbit station is not happening under current policy. Administrator Isaacman, Mar 24, 2026: pause Gateway in current form; focus on surface infrastructure. Reuters and SpaceNews characterize as cancel / halt of the orbital station deployment plan.

conjecture “Paused” ≠ reversible on the old design. Corrosion on HALO/I-Hab, partner MOU renegotiation, and explicit $20B surface redirection make a full NRHO Gateway revival the low-probability path without a future political reverse.

2. Why NASA’s own pages read stale

Both pages below still contain substantial pre-2026 architecture copy. That is not accidental ignorance — it is partial rebranding lag plus deliberate editor notes.

Source	What it still says	What policy now says (2026)
nasa.gov/…/artemis/	Has been updated for A3 LEO demo, A4 first landing, A5 late-2028 cadence. Still links Gateway as “Deep Space Exploration / first space station around the Moon.”	Gateway is paused; Moon Base is the sustained-presence centerpiece. Mission numbers match new plan; Gateway framing does not fully.
nasa.gov/…/gateway-about/	Full NRHO station pitch: PPE+HALO Falcon Heavy, I-Hab on A4, Lunar View on A5, UAE airlock on A6, Canadarm3, etc. Has editor note that website is being updated for Feb/Mar 2026 policy.	That assembly sequence is not the plan. Hardware status sections (HALO arrived AZ Apr 2025, PPE build, etc.) remain useful as inventory, not as flight plan.

Working rule for this artifact Prefer dated news releases and architecture announcements over undated program encyclopedias when they conflict. Prefer physical hardware milestones over artist concepts. Prefer partner agency statements over NASA spin when interests diverge.

3. Artemis III · IV · V — current plan

3.1 Timeline (controlled by Feb 27 + Mar 24 2026 decisions)

2022-11

Artemis I — uncrewed SLS/Orion test. Complete.

2026-04-01

Artemis II — first crewed Artemis flight; lunar flyby. Complete.

2027 (mid / NET)

Artemis III — crewed LEO demo: Orion docks with Blue Origin and/or SpaceX lander pathfinders. ~2 weeks. No Moon.

early 2028

Artemis IV — first crewed lunar surface landing (south pole region). Orion + commercial HLS. No Gateway required.

late 2028

Artemis V — second surface mission; “expected to begin building its Moon base.” Cadence target ~1/year after.

3.2 Artemis III — detail

Role

Risk reduction for first landing (explicit Apollo 9 analogue for commercial HLS).
First flight of Orion docking system.
Prove multi-launch choreography with heavy commercial vehicles in LEO before cislunar complexity.
Drive HLS readiness under Administrator pressure (Isaacman pushing both providers for 2027).

Crew (named 2026-06-09)

Randy Bresnik — Commander (NASA)
Luca Parmitano — Pilot (ESA) — first ESA astronaut on an Artemis flight assignment
Andre Douglas — MS (NASA)
Frank Rubio — MS (NASA); record US single-flight duration (371 days)
Bob Hines — backup

Planned A3 ops sequence (NASA Jun 2026 description)

Blue Origin lander pathfinder launches first; loiters weeks in orbit.
Crew launches on SLS/Orion to LEO.
~2 days docked with Blue Origin article: interfaces, software, propulsion, comms; crew enter lander.
Undock; await SpaceX Starship pathfinder.
~1 day docked with Starship pathfinder for checkouts.
Undock; splashdown Pacific. Total ~two weeks (flexible to ops).

NASA language allows “one or both” providers; public narrative is both pathfinders. Exact order/duration is ops-dependent.

3.3 Artemis IV — first landing

Target: early 2028 (held since mid-2025 landing target; renumbered when A3 became LEO demo).
Architecture: Orion to lunar orbit → transfer to commercial HLS → ~1 week near south pole for two surface crew → return to Orion → Earth.
Lander choice: whichever provider is ready — SpaceX Starship HLS and Blue Origin Blue Moon both still in play for surface missions.
SLS change: standardize configuration; replace ICPS with a new second stage approach; Exploration Upper Stage and Mobile Launcher 2 are no longer planned (explicit Feb 2026 architecture note).
Not required: Gateway docking, I-Hab delivery, NRHO station assembly (legacy Gateway-about page still describes I-Hab co-manifest — treat as obsolete for flight plan).

3.4 Artemis V

Target: late 2028 surface mission on standardized SLS.
Role shift: second landing + start of Moon Base build (robotic + early infrastructure cadence). Not “deliver Lunar View to Gateway.”
Subsequent missions ~once per year (aspirational cadence; not yet demonstrated).

3.5 What changed vs the “classic” Artemis narrative

Element	Classic (≈2020–2024 briefing)	Now (mid-2026)
First landing mission number	Artemis III	Artemis IV
Artemis III purpose	South pole landing via Starship HLS, often via Gateway	LEO dual-lander docking demo
Gateway on critical path	Yes for sustained ops; often for early landings	No — paused; surface-direct landings
SLS Block 1B / EUS / ML-2	A4+ backbone	EUS/ML-2 cancelled from plan; alternate upper stage path
Sustainment goal	Gateway + Artemis Base Camp later	Moon Base three phases; Gateway hardware maybe repurposed
Cadence ambition	~annual later	Explicit push: A4 and A5 months apart; then ~annual

4. Gateway — the real status

Was it cancelled? Functionally yes as a lunar-orbit station deployment plan. Formally “paused in its current form.” On March 24, 2026, at the national space policy / “Ignition” event, Administrator Jared Isaacman announced NASA is pausing Gateway and redirecting focus and funding (~$20B over ~7 years cited for lunar base development) to a sustained surface presence. Multiple outlets (Reuters, SpaceNews, Space.com) report cancel/halt language for the orbital station; NASA spokespeople repeated “not cancelled, paused and repurposed.” For analysis: treat NRHO Gateway as off the critical path and unlikely to assemble as designed.

4.1 Why the pivot (first principles, not press-release)

Stated drivers

Race framing vs China crewed landing ~2030.
Simplify architecture; put humans on surface sooner without orbital intermediate.
Political mandate: enduring physical foothold on Moon before end of administration term language.
Budget trade: surface infrastructure vs dual-track orbital + surface.

Structural drivers (derived)

Commercial HLS (especially Starship-class) reduces need for a logistics/aggregation depot in NRHO for early landings.
SLS co-manifest Gateway assembly tied to delayed Block 1B path — now abandoned.
Years of Gateway critical-path risk (integration, thrusters, international schedule) vs surface-robotics path that can fail and retry cheaper.
Hardware already late + corrosion discovery in Apr 2026 made “finish and fly as designed” less attractive.

4.2 What “pause” does and does not mean

Claim	Status
PPE + HALO launch on Falcon Heavy to NRHO before A4	off plan
Crew live on Gateway for A4/A5	off plan
Physical modules exist / advanced build	true — see inventory
International MOUs (ESA, JAXA, CSA, UAE) still valid as written	under renegotiation / study
Some hardware repurposed to surface base or other missions	stated intent
Congress fully locked final disposition of every contract	open — budget & diplomacy still matter
PPE reassigned to nuclear-electric Mars demo	assigned — SR-1 Freedom, NET Dec 2028

4.3 Partner impact (who is “left holding the bag”)

ESA: I-Hab, Lunar View (ex-ESPRIT refuel/window), Lunar Link, plus Orion European Service Modules (ESM still central — A2 used ESM; more under contract). Jun 2026: continue I-Hab through CDR then decide repurpose; slow Lunar View. Public friction over dependence on US architecture decisions.
JAXA: ECLSS/batteries for I-Hab; HTV-XG logistics concept; separately stronger on surface (pressurized rover / Lunar Cruiser with Toyota; LUPEX with India). Surface pivot may help Japan more than pure Gateway partners.
CSA: Canadarm3 — advanced robotics without a station to attach to is an awkward product; potential surface or LEO reuse studies.
UAE (MBRSC): Crew & Science Airlock commitment — early-stage relative to HALO/PPE; more design flexibility, more diplomatic uncertainty.

5. Module inventory — did they ever take shape?

Yes, for the core US/Italian pressure vessels and major PPE structure. Gateway was not vaporware. Several elements reached advanced manufacturing, delivery, or mid-build before the policy pivot. None have flown.

PPE
Power & Propulsion

Lanteris (ex-Maxar) commercial 1300-class bus + AEPS thrusters + giant ROSA arrays. Central cylinder complete; tanks installed; thruster acceptance testing at Glenn; solar arrays at Redwire Goleta. Power system started up for the first time in 2025 (NASA update, Jan 2026). Most complete “spacecraft-like” element — and the first to get a new job: assigned March 2026 as the propulsion element of SR-1 Freedom, NASA’s nuclear-electric Mars demo, launch NET Dec 2028.

advanced build
SR-1 Freedom

HALO
Hab / Logistics

Northrop Grumman; primary structure welded Italy (Thales Alenia), delivered Gilbert AZ 1 Apr 2025 for final outfitting. Intended command hub. Corrosion / surface metallurgical issue raised publicly by the Administrator at the 22 Apr 2026 House Science Committee hearing; Thales: “well-known metallurgical behavior” with ISS precedent, fix targeted by end Q3 2026. Isaacman questioned whether repair is warranted under surface pivot. A candidate for surface-habitat repurpose after remediation — NASA OIG (ML-26-002, Jun 2026) lists HALO among systems whose termination-or-repurposing disposition is not finalized.

delivered US
corrosion
surface candidate

Lunar I-Hab
ESA + JAXA

Thales Alenia Turin; ~10 m³ habitable / ~36 m³ pressurized. Primary structure in production; mockups for human factors; less severe version of same metallurgical issue; still in Europe as of spring 2026. ESA (Jun 2026): continue to CDR, then assess repurpose (surface lab / LEO / other). Not handed to NASA.

mid build
corrosion lighter
post-CDR decision

Lunar View
ESA (ex-ESPRIT)

Refuel + logistics + windows. Earlier design maturity than I-Hab flight article. ESA: “slow down the pace” after pause. Unlikely to fly as Gateway refueler soon.

slowed

Lunar Link
ESA comms

Telecom package for HALO. Powered on / tested around Apr 2025; path was ship to AZ for HALO install. Useful lunar-relay tech independent of full station.

hw tested
comms useful

Canadarm3
CSA / MDA

Detailed design / construction / test phase from 2024. Advanced robotics without Gateway host is open problem. Possible free-flyer, LEO station, or surface crane applications — none selected publicly as of this audit date.

in development

UAE Airlock
MBRSC

Implementing arrangement Jan 2024; crew + science airlock. Early relative to HALO/PPE. Most flexible / least sunk flight hardware.

early

DSL / Logistics
SpaceX contract

Gateway Logistics Services award to SpaceX (2020). Without Gateway, cargo architecture shifts to surface delivery + commercial LEO/cislunar as needed.

host gone

source-backed Modules took shape. HALO is a physical module in Arizona. PPE is a physical spacecraft under integration. I-Hab is a physical pressure shell program in Turin. This was not PowerPoint — which is exactly why the pause is expensive diplomatically and financially.

conjecture Corrosion is real but not the root cancel cause. ESA/Thales pushed back on framing corrosion as primary delay driver; issue characterized as known metallurgy with ISS precedent. Policy pivot (surface race + budget) is the root; corrosion is an aggravating fact that weakens the case to finish the orbital design.

6. Replacement architecture — Moon Base (surface-first)

Announced March 24, 2026; phased plan fleshed out through May–June 2026 NASA materials.

Phase 1 — Now–2029

Learn, Test, Build

Robotic south-pole scouting
Tech demos, landing reliability
CLPS / commercial cargo ramp
Prep for A4/A5 crew landings

Phase 2 — 2029–2032

Early Habitation

Semi-permanent infrastructure
Power, nav, comms constellations
Surface prep; early logistics
Gateway hardware integration attempts

Phase 3 — 2032+

Sustained presence

Longer crew stays
Heavier cargo HLS deliveries
Routine rotations
Mars-prep ops

Also announced in the March policy package: SR-1 Freedom, a nuclear-electric demonstration spacecraft that pairs a >20 kW fission reactor with the repurposed Gateway PPE, targeting a Mars launch NET December 2028 with the Skyfall payload (three Ingenuity-class helicopters to scout landing sites). Distinct from the Moon Base but part of the same “stop orbiting the problem, go somewhere” framing — and the first concrete answer to what happens to Gateway hardware.

7. Rest of the world

7.1 China + ILRS (primary competitor track)

International Lunar Research Station (ILRS) — China-led, Russia partnered, growing roster of smaller partners. Surface-centric research station (south pole focus), not a US-style NRHO gateway clone.

Phase / asset	Target	Notes
Chang’e-6	2024 (done)	Farside sample return — completed; prestige & science win
Chang’e-7	2026	South pole survey; lander + flyer + international payloads
Chang’e-8	~2028	ISRU tech verification; basic ILRS building block
Basic ILRS	by ~2035	South pole basic facility
Expanded ILRS	~2045	Networked surface + orbital hub concepts
Crewed landing (national goal)	~2030	Race metric US officials cite constantly

ILRS deliberately courts non-Artemis-Accords states. Parallel diplomatic track to Artemis Accords (60+ signatories; Botswana noted as 68th in 2026 NASA news).

7.2 Russia

Luna-Glob / Luna-Resurs line (Luna-25 failed 2023; later orbiters/landers delayed). Luna-26 polar orbiter still on long-range charts but schedule credibility is weak post-sanctions and failures.
ILRS partnership with China is the strategic human-lunar path; independent Russian crewed lunar program is not competitive with Artemis or CNSA on current evidence.

7.3 Europe (ESA)

Still critical to Artemis transport: European Service Module for Orion (proven on A1 uncrewed, A2 crewed).
Gateway contributions disrupted: I-Hab / Lunar View / Lunar Link — continue selectively while studying reuse.
Political lesson (2026): European press and ESA leadership stressing reduced structural dependence on US architecture U-turns (Gateway pause + Mars Sample Return pain in same news cycle).
Parmitano on A3 keeps European astronauts on the human flight line even without Gateway crew slots.

7.4 Japan (JAXA)

Gateway: ECLSS/batteries/HTV-XG path disrupted.
Surface (stronger hand): Pressurized rover (Lunar Cruiser, Toyota) under NASA–Japan implementing arrangement — NASA provides delivery + Japanese surface crew seats. Still the flagship non-US surface system.
LUPEX (with ISRO) lunar polar exploration continues as robotic science/ISRU pathfinder track.

7.5 Canada, UAE, India, others

Canada: Canadarm3 orphaned from host; CSA remains Artemis partner via Accords + robotics expertise.
UAE: Airlock deal and astronaut flight opportunity under review with architecture change.
India (ISRO): Chandrayaan successes; LUPEX with Japan; human spaceflight (Gaganyaan) LEO-first; not yet a full lunar crew competitor but capable robotic partner.
Commercial US: SpaceX, Blue Origin landers; Astrobotic / Firefly / Intuitive Machines CLPS; Firefly Blue Ghost successful landing Mar 2025 — commercial surface economy is real, not theoretical.

7.6 Comparative scoreboard (qualitative, mid-2026)

Capability	US / Artemis	China / ILRS	Others
Crewed deep-space flight (beyond LEO)	Proven 2026 (A2)	Not yet	ESA via Orion ESM only
Crewed lunar landing	Target 2028 (A4)	Target ~2030	—
Robotic lunar soft landing / sample	CLPS mixed; US commercial success 2025+	Mature (CE-3/4/5/6)	India Chandrayaan-3 success
Orbital lunar station	Paused / cancelled path	Later ILRS phase concept	—
Surface base program	Now centerpiece	ILRS centerpiece	Partner contributions
Heavy commercial crew lander	SpaceX + Blue Origin race	State lander path	—

8. First-principles audit

8.1 What problem was Gateway solving?

Original theory of the case:

Aggregation node — assemble lander, crew, cargo in a stable cislunar orbit without requiring every piece to meet perfectly on one launch.
International political glue — ISS-like partnership structure for the Moon.
Deep-space systems testbed — life support, autonomy (VSM), high-power SEP, radiation science outside Van Allen belt.
Mars rehearsal — distant, delayed-comms habitat ops.

8.2 Which of those still hold after Starship-class HLS + surface pivot?

Need	Still requires Gateway?	Notes
First landings	No	Direct Orion↔HLS in lunar orbit is the A4 plan.
High cargo mass to surface	No	Super-heavy commercial landers + CLPS beat NRHO relay for bulk cargo.
International partnership vehicle	Partially	Need exists; Gateway was the vehicle; surface modules + seats must replace it or partners hollow out.
Deep-space habitation R&D	Partially	Surface habitats are different environment (1/6 g, dust, thermal). LEO commercial stations ≠ cislunar. Some gap remains.
SEP / high-power bus demo	No	PPE can demo without a full station if flown free.
Science in NRHO	No unique requirement	Specialized free-flyers can do radiation/heliophysics cheaper.

8.3 Architecture logic of the new plan

deduction — stated assumptions If commercial HLS can repeatedly deliver crew and cargo to the south pole, an orbital waystation is optional for early Artemis objectives. Assumptions: lander reliability, propellant logistics (especially Starship refueling chain), abort modes without Gateway safe haven, and political will to accept surface-first risk. A3 exists to stress-test docking pieces of that assumption in LEO.

deduction — stated assumptions If the binding constraint is calendar race with China, any critical path element that is not surface boots is a candidate for deletion. Gateway failed that test. ESM, SLS/Orion, suits, landers, and surface power do not.

conjecture The 2028 double landing + base start cadence is schedule-aggressive relative to historical NASA human-spaceflight slip rates and dual-HLS maturity. A3 as Apollo 9 is wise. Compressing A4/A5 into the same year after that is the stress point.

8.4 Failure modes of the new architecture (not of the old brochure)

Lander readiness: If neither HLS is crew-ready for early 2028, A4 slips; A3 may have “passed” docking with pathfinders that are not flight-final vehicles.
No orbital safe haven: Contingency options for stranded lunar-orbit crew are thinner without Gateway.
Partner erosion: ESA/JAXA/CSA/UAE sunk costs without flight roles → Accords coalition softens while ILRS recruits.
SLS upper-stage transition: Killing EUS/ML-2 while raising cadence is a production bet that can bite A4+.
Repurpose fantasy: Orbital modules are not surface habitats without major redesign (dust seals, thermal, landing loads, foundation). “Repurpose HALO” can mean years of re-engineering, not a forklift move.
Budget fiction: $20B surface base + existing Artemis stack + nuclear path must survive Congress; pause language may re-expand if appropriators defend Gateway jobs.

8.5 What we would need to see to restore confidence in each claim

Claim	Falsifier / confirmer
A3 in 2027	SLS core stacking + both pathfinder flight articles + docking system integrated; public joint test schedule
A4 landing 2028	Uncrewed HLS lunar landing demo; suit certification; South Pole site downselect with comms/power plan
Gateway truly dead	Appropriations zero out NRHO assembly; partner MOUs rewritten; PPE flown to non-Gateway mission
HALO becomes surface hab	Public redesign contract, landing system interface, and environmental re-qual plan — not just speeches
US beats China to crew landing	A4 success before CNSA crewed landing attempt

9. Open questions / watch list

Near-term (to end-2026)

A3 hardware: Orion docking system, heat shield, SLS core engines install, spacer vs upper stage design
Blue Origin + SpaceX pathfinder readiness statements
Chang’e-7 launch outcome
ESA post-CDR decision on I-Hab
HALO corrosion remediation by ~Q3 2026 claim
FY budget language: Gateway line items vs Moon Base

Structural

Who lands first on A4 — SpaceX or Blue Origin?
Starship orbital refueling cadence demonstrated end-to-end
JAXA rover delivery mission number under new base plan
Whether Canadarm3 finds a host
Whether SR-1 Freedom holds its Dec 2028 Mars window — the PPE disposition is now the test case for every other “repurposed” Gateway element

Anomaly / inconsistency log (compiled 2026-07-09)

NASA Gateway-about page still describes full A4–A6 NRHO assembly while editor note admits architecture change — dual messaging continues mid-2026.
NASA Artemis main page still sells Gateway as “central” / “first space station around the Moon” after pause.
Wikipedia and some outlets say “cancelled”; NASA says “paused”; Reuters said “canceled plans to deploy” — all three can be true under different definitions.
A5 “begin Moon base” vs Phase 1 robotic base work already claimed for 2026–2029 — terminology overlap between crewed mission labels and base program phases.
Corrosion narrative diverged: Administrator skepticism vs contractor “known issue, fixable like ISS.”

10. Primary sources & further reading

NASA Artemis hub (partially updated)

NASA Gateway about (legacy + editor note)

NASA: Adds Mission, Updates Architecture (announced 2026-02-27)

NASA: Strengthens Artemis (posted 2026-03-03) — A3 LEO demo added

NASA: Gateway power system first start-up

NASASpaceflight: SR-1 Freedom unveiled (2026-03) — PPE to Mars demo

NASA OIG ML-26-002 (2026-06-24) — disposition of terminated/repurposed Artemis systems

NASA A3 crew release (2026-06-09)

NASA Artemis III mission page

NASA Artemis IV mission page

NASA national space policy / Moon Base (2026-03-24)

NASA Moon Base phases

NASA Moon Base hub

SpaceNews: Gateway halt (2026-03-24)

Reuters: cancel orbital station / $20B base

Space.com: Gateway out, bases in

ESA next steps on Gateway contribs (2026-06)

ESA on corrosion claims

Payload: Europe left holding the bag

ESA Lunar I-Hab

ILRS overview

NASA/JAXA pressurized rover