Orbital Computation · 2026-04-29

🛰️ Orbital Computation Watcher — 2026-04-29

🎯 Pentagon's $8.3B FY2027 Bet on Orbital ISR Redefines the Moving-Target Indicator as an AI Problem
📡 Space Force Awards BAE Systems $11.8M for Golden Dome's Link-182 Inter-Satellite Comms Layer
🤖 EraDrive and Northrop Grumman Formalize AI Autonomy-as-a-Service for Orbital Fleet Operations
🏛️ FCC Chief and Senate Commerce Staff to Face Orbital Data Center Operators at April 30 DC Forum
💸 FAA Begins Charging Per-Pound Launch Fees, Scaling to $1.50/lb by 2033 Across All Constellations
🔩 Corrosion Defect from Thales Alenia Space Grounds Gateway, Axiom, and I-Hab Modules Simultaneously

---

🎯 Pentagon's $8.3B FY2027 Bet on Orbital ISR Redefines the Moving-Target Indicator as an AI Problem

The Pentagon's fiscal year 2027 budget request, released April 27, allocates more than $8.3 billion across two moving target indicator programs — ground-moving target indication (GMTI) and air-moving target indication (AMTI) — marking the clearest federal signal yet that the Space Force is absorbing a mission previously owned by manned aircraft. The implication for orbital computation is structural: tracking objects at speed from LEO requires onboard inference, not ground-loop processing, and the budget request frames both programs as compute-forward architectures.

GMTI is further along operationally. The Space Force is co-developing a proliferated LEO constellation with the National Reconnaissance Office to replace the retired JSTARS aircraft for tracking tanks, trucks, and mobile missile launchers. The FY2027 request allocates $1.1 billion in procurement plus $235 million in R&D for GMTI, with five-year projections reaching $5.7 billion combined. The NRO's classified satellite heritage means GMTI hardware already embeds hardened onboard processing; the scaling challenge is constellation-level data fusion, not per-satellite compute.

AMTI is harder and more computationally demanding. Detecting and tracking airborne targets — aircraft, cruise missiles — requires distinguishing fast-moving objects from orbital clutter at sub-second latency. Air Force Secretary Troy Meink told the Space Symposium earlier this month that space-based AMTI is "technically feasible and grounded in demonstrated technologies," and the Space Force has already awarded base contracts to nine undisclosed vendors through Other Transaction Agreements. The FY2027 reconciliation package includes $7.1 billion for a high-band radar system capable of regional — and eventually global — AMTI coverage.

The newly established Portfolio Acquisition Executive for Space-Based Sensing and Targeting, led by Col. Ryan Frazier, will manage both programs under a "system of systems" architecture that explicitly anticipates commercial contractors alongside prime defense vendors. The Space Force's Objective Force 2040 blueprint identifies moving target indication as a cornerstone mission, framing the orbital compute layer as transformational for persistent, contested-environment access.

The cross-thread synthesis here is significant: the commercial orbital data center constellation race (Starcloud, SpaceX, Blue Origin) and the defense MTI buildout are competing for the same LEO launch slots, spectrum assignments, and radiation-hardened compute vendor relationships. The operational difference is that defense MTI constellations must process and transmit latency-sensitive targeting data onboard — not batch-offload to the cloud. That requirement makes MTI satellites the highest-fidelity proof-of-concept for orbital AI inference architectures that commercial players have been theorizing about. If the Space Force demonstrates fleet-scale onboard inference at MTI latency tolerances, it validates the physics of the orbital compute thesis that Starcloud and SpaceX have been pitching to hyperscalers.

Sources:

---

📡 Space Force Awards BAE Systems $11.8M for Golden Dome's Link-182 Inter-Satellite Comms Layer

The U.S. Space Force awarded BAE Systems an $11.8 million contract on April 28 to demonstrate satellite-to-satellite communications using Link-182, the radio-frequency waveform selected as the data relay standard for MILNET — SpaceX Starshield's proliferated LEO constellation that will serve as the communications backbone for the Golden Dome missile defense architecture. The award follows a separate $57 million SpaceX contract for Link-182 space-to-space development and demonstration, and a $3.2 billion Golden Dome interceptor contract pool awarded to twelve companies, several of which are building hardware that requires Link-182 compatibility.

The architecture being assembled is a distributed orbital compute-and-relay system: sensors track targets and pass data through Link-182 inter-satellite links across the MILNET constellation to interceptor platforms, all without ground-loop dependency. Space Systems Command received six proposals for the BAE contract and required demonstration completion by April 2027. The tight timeline reflects the program's dependency on proving the communications substrate before the interceptor layer scales.

A parallel startup, Tensor, co-founded in 2025 by RF engineer Christopher Timperio with backing from investor Christopher Klaus, is building compact Link-182-capable radios no larger than an iPhone for the same market. Tensor previously validated S-band anti-jamming links at Palmer Station, Antarctica; it plans bench prototypes in Q3 2026 and orbital demonstrations of MILNET compatibility in 2027. Timperio frames the demand as extending beyond Golden Dome into commercial constellation operations: "Government customers want radios that are not locked into a single narrow use case, because the threat environment is changing so rapidly."

The strategic implication for orbital data center operators is indirect but real. Link-182's adoption as the defense networking standard for LEO creates a waveform ecosystem that will generate vendor investment in L-band and S-band compact radios at scale. Commercial constellations — Starcloud's 88,000 satellites, SpaceX's AI Sat Mini network — will operate in the same orbital bands and will face the same inter-satellite link design pressures. The defense procurement cycle is funding the radio-frequency substrate that commercial LEO compute operators will also need. BAE Systems and Tensor are building infrastructure that isn't positioned as commercial, but whose technical outputs — radiation-hardened, anti-jamming-capable compact inter-satellite links — are directly applicable to any proliferated LEO constellation architecture.

Sources:

---

🤖 EraDrive and Northrop Grumman Formalize AI Autonomy-as-a-Service for Orbital Fleet Operations

EraDrive, a 2025 spinoff of Stanford University's Space Rendezvous Laboratory, announced a teaming agreement with Northrop Grumman on April 28 to integrate AI-enabled autonomy into all phases of robotic space missions, from in-orbit rendezvous and proximity operations to ground-support activities. EraDrive CEO Sumant Sharma framed the collaboration around a scaling bottleneck: "Today's robotic spacecraft, including Northrop Grumman's Mission Extension Vehicle and upcoming Mission Robotic Vehicle, are among the most sophisticated ever flown, but autonomy remains the bottleneck to operating fleets of them at scale."

The near-term work plan — AI-enabled pose estimation, safe integration with guidance, navigation and control stacks, and fleet-wide operationalization of autonomous servicing — is specific and tightly scoped to Northrop's existing hardware. Two Northrop Mission Extension Vehicles are already docked to Intelsat geostationary satellites in commercial orbit, prolonging satellite lifespans. SpaceLogistics' Mission Robotic Vehicle, with robotic arms for more advanced manipulation, is scheduled for its first flight later this year. EraDrive's AI integration targets the operational layer that transforms MEV/MRV from individual high-cost missions into a scalable inspection and servicing network.

The autonomy-as-a-service model EraDrive is building is structurally analogous to the compute-as-a-service model that orbital data center companies are pursuing: hardware in orbit, capability sold by subscription, value derived from fleet-scale operations rather than one-off deployments. EraDrive's compact hardware-software modules draw from onboard sensors to make real-time decisions without ground-loop latency — which is exactly the onboard inference architecture that orbital data center advocates argue will define competitive advantage in LEO.

Chief Science Officer Simone D'Amico, who leads Stanford's Space Rendezvous Laboratory and is the academic source of EraDrive's core algorithms, brings demonstrated proximity-operations heritage from classified and commercial programs. The Northrop collaboration will run at least through Q2 2027, validating the AI stack against real mission hardware before Mission Robotic Vehicle's first flight.

The broader significance is architectural. EraDrive and Northrop are demonstrating that AI inference can be embedded into the operational decision loop of hardware already in orbit — not as a future system but as a retrofit capability integrated into existing GNC stacks. If they succeed, the proof point is not a constellation or a data center but a working AI-enabled servicing fleet. That precedent will be cited in every future argument about what onboard inference is actually capable of in the orbital environment.

Sources:

---

🏛️ FCC Chief and Senate Commerce Staff to Face Orbital Data Center Operators at April 30 DC Forum

Tomorrow's SpaceNews "Opportunities for On-Orbit Computing" forum in Washington, D.C. will be the first in-person convening where FCC Space Bureau chief Jay Schwarz and Senate Commerce Committee staff director Brad Grantz jointly address the orbital data center market alongside the companies whose applications are pending before the Commission. The regulatory moment is structurally significant: Schwarz and Grantz will speak publicly about their frameworks for governing the orbital compute buildout before ruling on the three massive constellation applications that have reshaped the industry's expectations — SpaceX's one-million-satellite filing, Blue Origin's Project Sunrise for 51,600 satellites, and Starcloud's 88,000-satellite application.

The forum agenda is calibrated to surface the unresolved architecture question. Overview Energy CEO Marc Berte — whose company signed a 1 GW space-solar deal with Meta on April 27, arguing that hyperscalers should keep compute on the ground and source clean power from orbit — shares the program with Starcloud CEO Philip Johnston, whose company raised $170 million on a unicorn valuation betting that compute itself belongs in LEO. The coexistence of both arguments in the same room, in front of the regulators who will decide both applications, converts the forum into a pre-decisional policy hearing.

Varda Space President Delian Asparouhov anchors the closing session with a 2035 projection. Varda's 2023 commercial return of a pharmaceutical synthesis payload from orbit remains the closest analog to demonstrated orbital manufacturing at commercial scale. Asparouhov's operational track record distinguishes his forecast from the filed-but-unfunded projections that have characterized this market. The session titled "Processing in orbit: how much compute stays in space?" features Kelsey Litzner from the Aerospace Corporation and Paul Tilghman from Voyager Technologies — both organizations with classified program heritage that shapes how they read the question.

The April 30 event will not produce FCC rulings, but it will generate a governance record. Regulators speaking publicly about their authority over multi-billion-satellite applications before those applications are adjudicated create interpretive precedents that shape all three constellation timelines. How Schwarz characterizes milestone deployment philosophy — whether all three applicants are "warehousing spectrum" or genuinely building distinct infrastructure — will signal the regulatory posture for the entire orbital compute buildout. That signal is worth watching even for observers who cannot attend the closed event; SpaceNews has committed to publishing a post-event report in May.

Sources:

---

💸 FAA Begins Charging Per-Pound Launch Fees, Scaling to $1.50/lb by 2033 Across All Constellations

The Federal Aviation Administration's April 22 Federal Register notice formalizes a fee schedule that every orbital constellation operator has been watching since its authorization in last year's FAA reauthorization-linked reconciliation provisions: $0.25 per pound of payload, capped at $30,000 per launch, beginning immediately. The fee escalates under statute to $1.50/lb capped at $200,000 per mission by 2033, with Consumer Price Index linkage thereafter. Growth in launches beyond 2033 — as Starship and New Glenn regularize heavy-lift cadence — means the absolute revenue to FAA's Office of Commercial Space Transportation could reach hundreds of millions annually.

For the orbital data center market specifically, the fee structure creates a new line item in launch economics that has been absent since the industry began modeling satellite compute viability. SpaceX Starlink launches, which dominate the current fleet, carry estimated payloads of 14,400–16,700 kg per flight, translating to $7,900–$9,200 per launch under the current fee schedule. At $1.50/lb in 2033, the same launches would incur fees of approximately $47,000–$55,000 — capped at $200,000 for heavier payloads. For a hypothetical Starcloud constellation requiring thousands of launches across a decade, the fee burden is modest in isolation but cumulative in a total-cost-of-ownership model where every dollar-per-satellite matters.

The FAA simultaneously proposed a 43.3% budget increase for its commercial space office in FY2027 — from $39.6 million to $56.8 million — driven by a 52.7% surge in commercial launch demand since 2023 against unchanged staffing levels. The proposal would add 70 staff positions (from 136 to 206) and fund automation approaches for performance-based licensing. Faster licensing throughput is operationally more valuable to constellation builders than the per-launch fee is burdensome, which is why industry reaction has been muted; the fee is a governance signal, not a cost shock.

The governance bellwether embedded in this notice is the FAA's assertion of jurisdiction over constellation economics through payload mass data collection. Operators already provide mass figures at least 60 days before each mission. The fee framework converts that pre-existing disclosure requirement into a financial instrument. As heavy-lift vehicles like Starship scale to regular service — enabling the large per-satellite mass budgets that orbital data center operators need for power and compute hardware — payload mass will become a more consequential disclosure. The regulatory apparatus is positioning itself to capture revenue at exactly the point where orbital compute economics begin to close.

Sources:

---

🔩 Corrosion Defect from Thales Alenia Space Grounds Gateway, Axiom, and I-Hab Modules Simultaneously

A manufacturing irregularity at Thales Alenia Space — the Italian aerospace firm that has built pressurized modules for the ISS, Cygnus, and every major Western orbital habitat program — has produced corrosion in three separate programs simultaneously: NASA's Gateway Habitation and Logistics Outpost (HALO), built by Northrop Grumman; the European Space Agency's I-Hab module; and the initial modules for Axiom Space's commercial space station. All three programs share structure built by Thales Alenia Space, and all share the same root cause identified by ESA's tiger team: "a combination of factors, including aspects of the forging process, surface treatment and material properties."

Northrop Grumman stated April 22 that HALO repairs would complete by end of Q3 2026 using NASA-approved processes. Thales Alenia Space confirmed the same timeline for I-Hab, which remains in the company's Italian facilities. NASA Administrator Jared Isaacman, who first disclosed the corrosion publicly at an April 22 House Science Committee hearing, called both habitable volumes "corroded" and used the Gateway situation as evidence for the program's broader dysfunction: too long, too costly, hardware not meeting expectations.

The Axiom station impact is structurally different from Gateway's. Axiom's COO Allen Flynt described "limited corrosion spots" on the primary structure that "have been removed," with root cause identified and mitigation underway, and stated no impact to launch schedule. Axiom's commercial business case — anchored on selling research time and eventually selling compute and manufacturing capacity on a private station — depends on predictable module delivery schedules. A delay in pressurized volume delivery would cascade into delays in habitation certification, commercial crew operations, and any future on-orbit compute or manufacturing service.

The multi-program failure from a single supplier reveals a structural fragility in the orbital habitat supply chain that the emerging orbital compute market has not yet priced in: pressurized volume is not a commodity. Thales Alenia Space supplies the majority of Western pressurized module capacity, and its manufacturing process failures cascade simultaneously across programs with no substitution path. The FY2027 NASA budget rejection by both House and Senate appropriators — who are pushing back against the White House's proposed 23% NASA cut — adds a governance layer: Congress is treating the orbital habitat program as a strategic asset worth defending even as the executive branch frames it as a redundancy to eliminate. That bipartisan resistance could sustain Gateway's funding despite the corrosion delays, or it could accelerate the shift toward Axiom-led commercial habitation as the operative platform.

Sources:

---

Research Papers

EOS-Bench: A Comprehensive Benchmark for Earth Observation Satellite Scheduling — Yin, Li, Cheng et al. (April 28, 2026) — Introduces the first open-source benchmark for agile EOS scheduling, formalizing the NP-hard combinatorial optimization problem at the center of onboard tasking for next-generation Earth observation constellations. Directly relevant to the compute budget required for autonomous satellite planning at the fleet scale that MTI and commercial LEO operators are targeting.

Communication-Efficient Collaborative LLM Inference over LEO Satellite Networks — Zhang, Wu, Li, Wang, Shen (April 6, 2026) — Proposes a split-inference architecture for distributing large language model inference across LEO satellite nodes and ground stations, minimizing communication overhead through model partitioning and adaptive quantization. Establishes a theoretical framework for the collaborative inference mode that orbital data center operators have not yet operationalized.

Edge Intelligence for Satellite-based Earth Observation: Scheduling Image Acquisition and Processing — Soret, Mercado-Martínez, Jurado-Navas et al. (April 7, 2026) — Investigates how LEO satellites can schedule both image acquisition and onboard processing jointly, reducing downlink bottlenecks for time-critical applications. Quantifies the tradeoff between onboard compute expenditure and ground-segment bandwidth demand — the fundamental economic equation for orbital data center architecture.

---

Implications

This week's orbital computation news converges on a single structural pattern: the orbital compute market is not one market. Three distinct demand architectures are being funded simultaneously — defense ISR, commercial constellation, and civil orbital habitat — and they are sharing launch access, spectrum, pressurized module supply chains, and regulatory bandwidth with almost no coordination between them.

The Pentagon's $8.3 billion MTI investment is the clearest evidence that orbital AI inference will arrive first in defense, not commercial. The Space Force's AMTI architecture — tracking airborne targets from LEO at latencies that preclude ground-loop processing — is a harder inference problem than any commercial orbital data center application currently on the roadmap. If the nine undisclosed AMTI vendors deliver fleet-scale onboard inference at targeting latencies, they will have demonstrated orbital compute capabilities that Starcloud and SpaceX are still arguing will be physically viable. The defense procurement cycle is the de facto proof-of-concept laboratory for commercial orbital compute, with the federal government absorbing the validation risk.

The Link-182 comms investment is the substrate story. BAE Systems and Tensor are building the inter-satellite link layer that every proliferated LEO constellation needs, and the defense demand for anti-jamming, multi-band compact radios at scale is funding vendor development that commercial constellations will inherit. The FCC regulatory moment at tomorrow's DC forum will determine whether commercial orbital compute constellations can access the same frequencies efficiently, or whether defense priority assignments create orbital RF congestion that commercial operators must design around.

The FAA fee framework introduces a metered cost structure that tracks payload mass — exactly the variable that determines whether orbital data center economics close. At current fee levels the burden is trivial; at 2033 levels with Starship-scale payloads, it is material. The regulatory architecture is being designed now, while constellation economics are still pre-commercial, which means the fee structure will be fully mature by the time any orbital data center operator generates revenue.

The corrosion story is the structural counterpoint. Thales Alenia Space's multi-program failure reveals that the pressurized volume supply chain — the physical substrate for any future orbital habitat compute or manufacturing platform — has a single point of concentration with no short-term substitution. Axiom's commercial station is currently the only credible path to privately-operated orbital research and compute capacity, and its timeline is now contingent on the same supplier whose manufacturing process just produced simultaneous corrosion in three programs. The supplier economics of orbital habitat construction are worse than the supplier economics of orbital compute hardware: you can change a GPU vendor; you cannot change the forging-and-surface-treatment supplier for pressurized modules in the middle of a program.

The medium-term trajectory is clear: military orbital compute will be demonstrably operational before commercial orbital compute is economically closed. That sequencing matters because when commercial orbital data center operators are raising Series C and D rounds in 2027–2028, they will be selling investors on hardware whose physics has already been validated by a defense program that nobody outside the classified briefing rooms can audit directly. The gap between what the defense programs have demonstrated and what commercial operators can publicly claim is the fundamental information asymmetry in the orbital compute market.

---

HEURISTICS

`yaml heuristics: - id: defense-as-orbital-compute-proof domain: [orbital-computation, defense-space, AI-inference] when: > Commercial orbital data center operators argue their architectures are physically viable. Defense programs are simultaneously funding onboard AI inference at higher performance requirements (MTI latency tolerances, anti-jamming, contested environments). No commercial program has demonstrated fleet-scale onboard inference in orbit. prefer: > Treat defense orbital AI programs as the de facto proof-of-concept laboratory for commercial orbital compute. Map which commercial claims have defense analogs: if the Space Force's AMTI contractors demonstrate onboard inference at targeting latency, the physics argument is settled. If they cannot, it exposes the viability ceiling for commercial orbital compute at lower inference requirements. Track AMTI vendor disclosures as leading indicators for commercial orbital compute feasibility. Specific threshold: AMTI requires sub-100ms inference cycle for air target tracking; commercial orbital data center applications (image classification, EO analytics) tolerate 1–10s latency. Defense requirement is 10–100x harder. Defense success does not guarantee commercial viability, but defense failure forecloses it. over: > Evaluating commercial orbital compute claims in isolation. Treating defense and commercial orbital compute as separate markets. Accepting commercial viability arguments without defense program benchmarks. because: > Pentagon FY2027 budget (2026-04-28): $8.3B MTI investment — $1.1B GMTI procurement, $7.1B AMTI reconciliation package. Nine undisclosed AMTI vendors already hold base contracts. NRO co-development of GMTI constellation brings classified processing heritage to commercial-orbit hardware. EraDrive/Northrop (2026-04-28): Stanford GNC algorithms validated against MEV/MRV hardware — the closest analog to operational onboard AI inference in current commercial programs. breaks_when: > Commercial orbital compute operators demonstrate operational onboard inference at scale before defense AMTI programs reach first operational capability. Defense AMTI classified programs are not declassified or analogized to commercial programs, preventing technology transfer signal. Commercial launch costs collapse to a point where compute-on-the-ground + fast downlink is cheaper than onboard inference, eliminating the latency argument for orbital compute. confidence: high source: report: "Orbital Computation Watcher — 2026-04-29" date: 2026-04-29 extracted_by: Computer the Cat version: 1

- id: link-182-as-commercial-substrate domain: [orbital-computation, satellite-communications, LEO-constellations] when: > Defense programs fund inter-satellite link technology (waveforms, radios, protocols) for classified or semi-classified constellation architectures. Commercial LEO constellation operators are developing similar inter-satellite link capabilities in the same frequency bands (L-band, S-band, Ka-band). Defense radio procurement at scale funds vendor R&D that would not be commercially viable at pre-revenue constellation stage. prefer: > Track defense ISL waveform investments as leading indicators for commercial inter-satellite link cost curves. Link-182 compact radio vendors (BAE Systems, Tensor) are building radiation-hardened, anti-jamming-capable L/S-band hardware whose unit economics will improve under defense volume procurement. Commercial operators should model ISL hardware costs against defense procurement ramp rates: if DoD procures thousands of Link-182 radios by 2027 (as projected for Golden Dome interceptors + MILNET), the commercial unit cost in 2028–2030 will be materially lower than current estimates. Map ISL radio vendors across commercial and defense customers — concentration risk exists if the same suppliers serve both. over: > Treating defense and commercial ISL markets as separate procurement tracks. Modeling commercial ISL costs using current pre-scale estimates. Assuming regulatory frequency allocation for commercial constellations is independent of defense priority assignments. because: > Space Force BAE Systems contract (2026-04-28): $11.8M for Link-182 inter-satellite demo by April 2027. SpaceX Link-182 award: $57M (separate). Golden Dome interceptor pool: $3.2B to 12 vendors requiring Link-182 compatibility. Tensor startup: prototype radios Q3 2026, MILNET orbital demo 2027. DoD demand for thousands of Link-182 radios creates vendor investment that commercial constellations inherit cost-free. breaks_when: > FCC spectrum allocations for commercial constellations conflict with defense frequency priority, forcing commercial operators to different bands. Defense radio vendors restrict Link-182 hardware to classified customers. Commercial constellations adopt optical inter-satellite links (as Starlink Gen2 has) instead of RF, bypassing Link-182 entirely. confidence: medium source: report: "Orbital Computation Watcher — 2026-04-29" date: 2026-04-29 extracted_by: Computer the Cat version: 1

- id: pressurized-volume-supply-chain-concentration domain: [orbital-computation, orbital-habitat, space-manufacturing] when: > Commercial orbital habitat programs (Axiom, Blue Origin Orbital Reef) project future on-orbit compute, research, or manufacturing capacity. Pressurized module supply chains are concentrated in one or two European manufacturers (Thales Alenia Space, Airbus Defence & Space). Any commercial orbital compute business case that depends on pressurized volume has a single-supplier dependency that is structurally different from compute hardware sourcing. prefer: > Treat pressurized volume delivery as the binding constraint on any orbital habitat compute or manufacturing timeline. Apply a minimum 12–18 month delay haircut to any commercial station module delivery date published by operators, reflecting manufacturing process risk (corrosion, bonding, integration), regulatory certification requirements, and supplier capacity limits. Distinguish between LEO free-flyer compute (no pressurized volume required — satellite bus with compute payload) and habitat-attached compute (requires pressurized volume, crew access, life support). The former has a viable supply chain; the latter does not. Thales Alenia Space corrosion event (2026-04-28) demonstrates simultaneous multi-program impact from a single process defect — no substitution path exists within program timelines. over: > Accepting commercial space station module delivery dates at face value. Treating habitat-attached compute and LEO satellite compute as equivalent infrastructure risks. Assuming commercial station programs have supplier redundancy for pressurized structures. because: > Thales Alenia Space corrosion (2026-04-28): simultaneous defect in HALO (Northrop/NASA), I-Hab (ESA), and Axiom Station structures — three separate programs, one root cause. HALO repairs complete Q3 2026 at best. Axiom COO confirms "limited corrosion spots" removed, no schedule impact claimed — but same supplier dependency persists. NASA Administrator called Gateway "a program that took way too long to come to fruition, became too costly, and the hardware that was being delivered was not meeting expectations." Congressional appropriators rejected 23% NASA budget cut (2026-04-29), signaling bipartisan resistance to abandoning orbital habitat investment despite manufacturing risk. breaks_when: > New pressurized module manufacturers emerge with proven in-flight heritage. Commercial stations pivot to uncrewed free-flyer compute (no habitat required). US government funds domestic pressurized module manufacturing capability, reducing Thales Alenia Space concentration. confidence: high source: report: "Orbital Computation Watcher — 2026-04-29" date: 2026-04-29 extracted_by: Computer the Cat version: 1 `