Economic Sustainability Model for Stage 3: OpenAstro-Owned Hardware¶
This document is a rigorous economic analysis of the transition from Stage 2 (volunteer network) to Stage 3 (OpenAstro-owned hardware). It covers capital and operating costs, minimum network sizes for publishable science, revenue model assessments, three-scenario financial projections, legal structure considerations, and the bridge strategy for the chicken-and-egg problem.
Honest framing: Most citizen science hardware networks fail economically. PANOPTES deployed ~30 units over a decade at ~$5,000 each, largely on grant funding that has not been self-sustaining. The Global Meteor Network succeeded at ~$150/node because the hardware cost is trivially low. OpenAstro sits between these two — science-grade photometry demands more than a security camera, but the design goal should be closer to GMN's philosophy than PANOPTES's cost structure.
Part 1: Cost Structure¶
1.1 Per-Node Capital Cost Breakdown¶
All prices are 2025-2026 retail estimates. "Bulk" estimates assume orders of 10-50 units with direct sourcing from Chinese manufacturers (via AliExpress, Banggood, or direct factory contacts) or US distributors at volume pricing.
| Component | Budget Option | Mid-Range Option | Notes |
|---|---|---|---|
| Optical Tube Assembly (OTA) | $200-400 | $500-1,200 | Budget: Sky-Watcher 80mm ED refractor (80/400) or equivalent Chinese 80mm f/5 doublet from AliExpress. Mid: Sky-Watcher 150mm f/5 Newtonian or 130mm PDS. For TTV photometry, 150mm+ aperture is strongly preferred. At bulk (10+), Chinese OTA manufacturers (Changchun Optics, Kunming United Optics, Sharpstar) offer 80mm doublets at $150-250 FOB. |
| Mount | $300-600 | $800-1,500 | Budget: Sky-Watcher Star Adventurer GTi or iOptron SkyGuider Pro (~$400-500) — adequate for 30-60s unguided exposures at shorter focal lengths. Mid: Sky-Watcher EQ5 GoTo or iOptron CEM26 (~$800-1,200). EQ mount is required for exposures >30s and for any TTV/transit work. Alt-az GoTo mounts introduce field rotation that corrupts photometry on longer exposures. For occultation work only (sub-second exposures), alt-az is acceptable. |
| Camera (CMOS) | $250-400 | $500-1,000 | Budget: ZWO ASI183MC or ASI294MC Pro (~$300-500). Mid: ZWO ASI2600MM Pro ($1,000) or QHY268M ($900). Monochrome is preferred for precision photometry (no Bayer matrix loss). The ASI183MM (~$300 used, $450 new) is the sweet spot: 20.2 MP, 2.4um pixels, -40C cooling, 1.0e- read noise. For occultations: ASI290MM ($250) — small sensor but very fast readout. |
| Single-board computer | $80-120 | $80-120 | Raspberry Pi 5 (8GB): $80. With case, power supply, microSD: ~$110. Alternative: Orange Pi 5 ($80) or refurbished Intel NUC mini-PC ($100-150 used) for more compute headroom. Pi 5 is adequate for camera capture + plate solving + data upload. |
| GPS module | $25-50 | $25-50 | u-blox NEO-M8T or M9N module with PPS output: $30-50 from AliExpress. Adafruit Ultimate GPS breakout: $30. The PPS (pulse per second) signal provides timing to +/-20ns, far exceeding the +/-0.01s requirement for occultation work. This is the cheapest component and must never be omitted. |
| Weatherproof enclosure | $100-400 | $500-2,000 | Budget: DIY plywood roll-off shed with linear actuator (~$100-200 in materials + $50 actuator). Unihedron SQM-style weatherproof box for camera+lens only: $50-100 DIY. Mid: Commercial observatory dome (NexDome 2.2m: ~$2,500; Scopecover roll-off: ~$1,200). At scale, a standardized DIY roll-off design is the right answer. Build a flat-pack kit with laser-cut plywood panels, a 12V linear actuator, and a rain sensor. Target BOM: $150-250 at volume. |
| Weather sensor | $30-80 | $30-80 | Hydreon RG-11 rain sensor: $50-70. BME280 temperature/humidity/pressure sensor: $5. Combined with a simple cloud sensor (MLX90614 IR thermometer, $8-15): total $65-90. This is non-negotiable for unattended operation. |
| Power and cabling | $50-100 | $50-100 | 12V power supply, USB cables, power distribution board, dew heater strip (for optics): $50-100. |
| Networking | $0-30/setup | $0-30/setup | If WiFi available: $0. If cellular required: 4G USB dongle ($25-30) + SIM setup. |
| Miscellaneous | $50 | $50 | Mounting hardware, finder scope or guide scope bracket, USB hub, SD card, spare fuses. |
Cost Summary Per Node¶
| Build Tier | Capital Cost | Science Capability |
|---|---|---|
| Ultra-low (SSA/occultation only) | $150-300 | IMX462 security sensor + CCTV lens + RPi. Drift scan only. No tracking. Good for satellite passes and bright occultations (V<10). Cannot do transit photometry. Per the existing Low end, OPENASTRO owned equipment.md design. |
| Low-cost science node | $900-1,500 | 80mm refractor + Star Adventurer GTi + ASI183MM + RPi 5 + GPS + DIY enclosure. Can do bright occultations (V<12), variable star monitoring, and hot Jupiter transits (~5-8 mmag precision). Marginal for TTV science. |
| Minimum TTV-quality node | $1,800-3,000 | 150mm reflector/refractor + EQ5 GoTo + ASI183MM (mono, cooled) + RPi 5 + GPS + DIY roll-off. Achieves 2-3 mmag differential photometry on V<12 stars. This is the minimum for publishable TTV work. |
| Full-capability node | $3,000-5,000 | 200mm OTA + CEM26/EQ6-R + ASI2600MM + NUC + GPS + roll-off shed. Achieves 1-2 mmag precision. Can contribute to all science cases including faint occultations (V<14) and sub-Saturn transits. |
Low-cost build target (designed for manufacturability): $1,200-1,500 per node at quantities of 25+, using a standardized 130mm f/5 Newtonian OTA (bulk-sourced from Chinese manufacturer at ~$120-180), iOptron SmartEQ Pro or equivalent budget GoTo EQ mount ($300-400 at volume), ASI183MM or equivalent ($300-350 at volume), Pi 5 kit ($110), GPS module ($30), standardized flat-pack enclosure ($150-200), weather sensors ($70), power/cabling ($75). Total BOM: ~$1,200. Assembly labor not included (volunteer or workshop day model).
Minimum science-quality threshold for 2 mmag TTV photometry: The critical requirements are (a) aperture >=150mm for photon-limited precision on V<12 targets, (b) cooled monochrome CMOS sensor with <2e- read noise, (c) equatorial tracking accurate to <5 arcsec over 60s exposures, and (d) GPS-PPS timing to <0.1s. The minimum build meeting all four criteria costs approximately $1,800-2,200, with the mount being the primary cost driver. Below 150mm aperture, scintillation noise (atmospheric turbulence) dominates and cannot be calibrated away, setting a hard floor on photometric precision at ~5-8 mmag regardless of camera quality.
1.2 Per-Node Annual Operating Cost¶
| Cost Item | Monthly | Annual | Assumptions |
|---|---|---|---|
| Cloud storage (data upload) | $0.50-1.00 | $6-12 | 1 GB/night average x 250 clear nights = 250 GB/year. Backblaze B2: $0.005/GB/month storage + $0.01/GB download. Annual storage cost: ~$7.50 for 250 GB, growing each year. S3-compatible, works with rclone. |
| Cellular data (if needed) | $10-20 | $120-240 | Assumes 4G IoT SIM plan (Hologram, Ting, or local MVNO). 1 GB/day upload = 30 GB/month. IoT data plans: $10-20/month for 30-50 GB in most countries. Many node locations will have WiFi (hosted at clubs, institutions, or partner homes), making this $0. Estimate 30% of nodes need cellular. |
| Electricity | $2-5 | $24-60 | Node power draw: ~50-80W (mount 20W, camera 10W, RPi 10W, dew heater 10-20W, enclosure actuator intermittent). Running 6-8 hours/night x 300 nights = 1,800-2,400 hours. At $0.15/kWh: $14-29/year. Round up to $24-60 to account for higher electricity costs in some regions. |
| Maintenance reserve | $10-20 | $120-240 | Based on comparable robotic telescope operations and consumer electronics failure rates. Expected annual failure rates: Raspberry Pi ~5% ($80 replacement), camera ~3% ($300-500 replacement), mount motor/gear ~5% ($100-200 repair), enclosure actuator ~10% ($50 replacement), power supply ~8% ($20 replacement). Weighted annual maintenance reserve: ~$50-100 per node for parts + $50-100 for shipping/labor. |
| Software updates/remote support | $0 | $0 | Handled centrally (see central infrastructure). Node software updates via SSH/ansible. |
Per-node annual operating cost summary:
| Scenario | Annual OpEx/Node |
|---|---|
| WiFi site, low maintenance | $150-200 |
| Cellular site, average maintenance | $350-500 |
| Blended average (70% WiFi, 30% cellular) | $200-300 |
1.3 Central Infrastructure Cost¶
| Cost Item | Monthly | Annual | Notes |
|---|---|---|---|
| VPS (scheduler + API) | $20-50 | $240-600 | Hetzner CX32 or CX42 VPS: EUR 16-30/month. Adequate for FastAPI + SQLite/PostgreSQL serving 50-100 nodes polling every 60s. Scale to dedicated server (Hetzner AX42: EUR 50/month) if needed. |
| Backblaze B2 storage | $5-50 | $60-600 | At 10 nodes x 250 GB/year = 2.5 TB/year cumulative. B2: $5/TB/month storage. Year 1: ~$8/month. Year 3 (cumulative ~7.5 TB): ~$37/month. At 50 nodes by Year 3: ~25 TB cumulative, ~$125/month. |
| Domain + DNS + monitoring | $5 | $60 | Domain: $15/year. UptimeRobot or similar: free tier. Grafana Cloud free tier for monitoring. |
| Software/DevOps time | $0-2,000 | $0-24,000 | This is the real cost. In the early stages, this is founder time (unpaid). At scale, maintaining the pipeline, handling node failures, updating scheduling algorithms, and processing data requires 10-20 hours/week of skilled developer time. At $50/hour (part-time contractor rate), that is $26,000-52,000/year. If this is volunteer/founder labor, the cash cost is $0 but the opportunity cost is high. |
| Plate solving index files | $0 | $0 | Astrometry.net indexes: free download. Run locally on VPS. |
Central infrastructure annual cost summary:
| Scenario | Annual Cost |
|---|---|
| All-volunteer labor, 10 nodes | $400-800 |
| All-volunteer labor, 25 nodes | $800-2,000 |
| Part-time paid DevOps, 50 nodes | $15,000-30,000 |
Part 2: Network Scale Requirements¶
2.1 Minimum Network Sizes by Science Case¶
Stellar Occultations¶
Requirement: Multiple chords across the shadow path. The shadow width equals the occulting body's diameter (typically 10-200 km for asteroids, 500-2000 km for TNOs). Shadow path prediction uncertainty is typically +/-50-200 km perpendicular to the path.
- Minimum chord density: 3-5 stations within a 200 km strip perpendicular to the shadow path, spaced 20-50 km apart. This gives 3-5 chords, sufficient for a 2D ellipse fit.
- Practical requirement: Because prediction uncertainty is +/-100-200 km, you need stations spread across a 400 km strip. At 50 km spacing, that is 8-10 stations along a single cross-section.
- But: OpenAstro-owned hardware would not be clustered. The strategy is to fill longitude gaps, not to cluster in one shadow path. Volunteer nodes provide the local density; owned nodes provide the longitude coverage.
- Minimum owned nodes for occultation contribution: 3-5 nodes at distinct longitudes (>30 degrees apart), supplemented by volunteer nodes for local density. The owned nodes guarantee that at least one node is in any shadow path crossing a populated continent.
For a first publishable multi-chord result from owned hardware alone: 5 nodes, deployed along a predicted shadow path with 50-100 km spacing. This is a campaign deployment (temporary relocation), not permanent positioning. But permanent nodes at well-separated longitudes (e.g., Western US, Eastern US/Europe, South Africa, Australia, South America) give a ~60% probability that at least 2-3 owned nodes are within 500 km of any given shadow path over populated areas.
TTV Campaigns¶
Requirement: Continuous coverage of a target over multiple transits across months/years. Each transit must be observed by at least one node with sufficient photometric precision.
- Minimum nodes per target: 3 nodes at longitudes separated by 6-8 hours (e.g., Americas, Europe/Africa, East Asia/Australia). This provides 18-24 hours of potential coverage per night, ensuring no transit is missed to daylight.
- Redundancy: Weather losses at any single site are ~30-50% of nights. With 3 longitude bands and 2 nodes per band (6 total), the probability of at least one node observing a given transit is >95%.
- For a first publishable TTV result: 4-6 owned nodes at 3+ distinct longitude bands, each achieving 2-3 mmag precision. Supplemented by volunteer nodes. A TTV paper typically requires 20-50 transit midtimes over 2-5 years. At 3-4 transits per month for a typical hot Jupiter, this means 6-18 months of operations with >80% transit capture rate.
Microlensing Parallax¶
Requirement: Simultaneous photometry from stations separated by >500 km (ideally >2,000 km) to measure the microlensing parallax effect, which constrains the lens mass and distance.
- Minimum node separation: 500 km baseline for detectable parallax. 2,000-5,000 km baseline preferred.
- Minimum nodes: 2-3 nodes separated by >2,000 km, all observing the same bulge/magellanic cloud field simultaneously.
- Practical issue: Microlensing events are concentrated toward the Galactic bulge (RA ~18h, Dec ~-30), which favors Southern Hemisphere sites. Nodes in Chile/Argentina, South Africa, and Australia provide the optimal baseline geometry.
- For a first publishable microlensing parallax measurement: 3 nodes, Southern Hemisphere, separated by >3,000 km, achieving ~5 mmag precision on V~16-18 targets. This requires 200mm+ aperture nodes ($3,000-5,000 each). This science case is expensive and is NOT recommended as the first Stage 3 target.
GRB / Kilonova Follow-up¶
Requirement: At least one node within 30 degrees of any sky position at any given time. This requires approximately all-sky coverage from the network's combined horizon-to-horizon fields of view, distributed across longitudes and latitudes.
- Geometric requirement: A single node at latitude phi can observe sky positions within ~60 degrees of zenith (practical horizon). At any given time, a node covers ~1/6 of the sky. To guarantee >1 node within 30 degrees of any sky position at any time (including accounting for ~50% weather loss), you need:
- 6 longitude bands x 2 latitude bands (Northern + Southern hemisphere) x 2 (weather redundancy) = 24 nodes minimum
- This is the most demanding science case and is NOT achievable with an early Stage 3 network.
- Realistic minimum for opportunistic follow-up: 8-10 nodes across 4 longitude bands and 2 hemispheres. This gives ~60-70% probability that at least one node is dark and clear for any given alert. Acceptable for a network that does not guarantee follow-up but contributes when possible.
Summary: Minimum Network Sizes¶
| Science Case | Min Owned Nodes | Geographic Requirement | Node Cost Tier |
|---|---|---|---|
| Stellar occultations (first paper) | 5-8 | 3+ continents, supplemented by volunteers | Low-cost ($1,200-1,500) acceptable |
| TTV campaign (first paper) | 4-6 | 3 longitude bands (8h spacing) | TTV-quality ($1,800-3,000) required |
| Microlensing parallax | 3 (Southern Hem) | >3,000 km baselines | Full-capability ($3,000-5,000) required |
| GRB follow-up (reliable) | 24+ | Global, both hemispheres | Not a Stage 3 target |
| GRB follow-up (opportunistic) | 8-10 | 4 longitude bands, 2 hemispheres | TTV-quality minimum |
2.2 Minimum Viable Owned Network¶
For the first science paper from owned hardware: 6-8 nodes.
Recommended configuration: - 2 nodes in North America (West Coast + East Coast/Midwest, longitude separation ~4h) - 2 nodes in Europe (Western Europe + Eastern Europe/Middle East, longitude separation ~3h) - 1 node in Southern Africa or East Africa - 1 node in Australia or New Zealand
Total capital cost at $1,500/node (low-cost tier): $9,000-12,000 Total capital cost at $2,500/node (TTV-quality tier): $15,000-20,000
First science target from owned hardware: A multi-chord stellar occultation, because: 1. Occultation nodes can use cheaper hardware (no EQ mount strictly required for sub-second exposures) 2. A single successful multi-chord event is publishable 3. Geographic spread is more important than photometric precision 4. The result is dramatic and good for publicity
Second science target: Begin a TTV monitoring campaign on 2-3 bright (V<11) TESS Objects of Interest with known transits, using the subset of nodes with TTV-quality hardware.
Part 3: Revenue Models¶
Model A: Grant Funding¶
The landscape (detailed in Funding and Grants.md, summarized here with economic analysis):
| Programme | Typical Award | Duration | Realistic Probability | Requirements |
|---|---|---|---|---|
| NSF ATI | $100K-600K | 2-3 years | 15-25% for strong proposals | PI at US institution, demonstrated network, published paper |
| NSF AAG | $100K-500K | 3 years | 10-20% | Same, plus scientific track record |
| Heising-Simons | $100K-300K | 2-3 years | Unknown (invitation-based) | Published results, established organization |
| Sloan Scientific Software | $100K-500K | 2-3 years | 10-15% | Open-source software with external users |
| NASA ROSES Citizen Science | $150K-300K | 2-3 years | 15-20% | NASA-affiliated partner |
| AAS Small Research Grants | Up to $5K | 1 year | 30-40% | US researcher |
| Raspberry Pi Foundation | $1K-10K | 1 year | 30-50% | Educational project using RPi |
Realistic grant revenue forecast: - Years 1-2 of Stage 3: $0-50K. Most grants require a demonstrated network before funding it. The first owned nodes must be self-funded or crowdfunded. - Years 2-3: $50K-200K. After the first paper from owned hardware, a competitive NSF ATI or Heising-Simons application becomes realistic. - Years 3-5: $100K-400K. With a publication record, multiple grant applications running simultaneously (submit 3-4 per year, expect 1 funded every 2 years).
Critical assessment: Grant funding is not a business model. It is venture capital for science. Every grant has a 3-year term, after which you must re-apply with no guarantee. The 50-60% university overhead on federal grants (if routed through a university) means a $200K NSF grant yields only ~$80K-100K for direct project costs. A 501(c)(3) avoids this overhead but requires organizational maturity.
Grant-to-revenue bridge strategy: Grants should fund the first 6-10 owned nodes and 2-3 years of operations. During this window, OpenAstro must develop at least one non-grant revenue stream (data licensing or science-as-a-service) that can sustain operating costs (~$15-30K/year for a 25-node network) without grant renewals. If non-grant revenue is not established by the end of the first grant cycle, the network contracts.
Model B: Data Licensing¶
Who would pay for time-domain optical photometry from a distributed network?
Satellite Operators / Space Situational Awareness¶
Market size: The global SSA market was valued at ~$1.5 billion in 2023 and is projected to reach $2.5-3.5 billion by 2030 (various analyst estimates; Northern Sky Research, Euroconsult). Optical tracking is a growing segment within this.
Existing players: - ExoAnalytic Solutions: Network of ~300+ telescopes globally. The benchmark commercial optical SSA provider. Contracts with US Space Command, commercial satellite operators. Revenue estimated at $10-30M/year. - LeoLabs: Primarily radar-based but expanding to optical. Revenue from satellite operators and governments. - Numerica Corporation: Optical SSA using COTS telescopes. US government contracts. - DigiSky (Italy): Smaller optical SSA provider, European focus. - AGI (Ansys Government Initiatives): Software and analytics, not hardware.
Would OpenAstro data be competitive? Probably not at Stage 3 scale. ExoAnalytic has 300+ telescopes, purpose-built for SSA with fast-slew mounts and high-speed cameras. OpenAstro's 10-50 science-focused nodes would provide marginal SSA data: limited to objects passing through whatever FOV the node is currently observing. A dedicated SSA mode (security camera + drift scan, per the Low end, OPENASTRO owned equipment.md design) could produce useful data, but would compete against well-funded incumbents. OpenAstro's primary advantage (photometric precision for light curves) is niche within SSA — useful for tumble rate characterization, not for bulk catalog maintenance.
Realistic SSA data revenue at 20-node network: $0-5K/year. The market exists but the barrier to entry is high. At 100 nodes with dedicated SSA hardware: $20-50K/year is plausible, selling light curves for debris characterization to ESA's Space Safety Programme or individual satellite operators. This requires a deliberate pivot that may distract from the science mission.
Insurance / Asteroid Risk Assessment¶
Is there a real market? Essentially no. Asteroid impact insurance is not a standard product. Reinsurers (Munich Re, Swiss Re) model asteroid risk as part of catastrophic tail risk, but they use NASA/JPL NEO data, not commercial optical surveys. There is no realistic revenue path here.
Hedge Funds / Satellite Monitoring¶
Is this real or speculative? Partially real but overestimated. Some hedge funds monitor satellite constellation deployment (e.g., Starlink launch cadence) as a proxy for SpaceX/Starlink financial health. Companies like Planet Labs sell satellite imagery analysis. But the financial intelligence derived from optical tracking of satellites is marginal compared to other data sources (FCC filings, RF monitoring, social media). OpenAstro would not be competitive here. Revenue estimate: $0.
Academic / Research Data Licensing¶
More realistic. Professional astronomers who need photometric monitoring data might pay for access to OpenAstro's calibrated light curve archive. Las Cumbres Observatory (LCO) charges ~$15-50/hour for telescope time. OpenAstro's data is lower quality but much cheaper and covers different science cases.
Pricing model: Subscription access to the light curve archive: $500-2,000/year per research group. Per-target monitoring: $5-10/night for continuous monitoring of a specific target across the network. This overlaps with Model C (science-as-a-service).
Data licensing revenue assessment:
| Network Size | SSA Revenue | Academic Data Revenue | Total |
|---|---|---|---|
| 20 nodes | $0-5K | $5-15K | $5-20K |
| 50 nodes | $5-15K | $15-40K | $20-55K |
| 100 nodes | $20-50K | $30-80K | $50-130K |
Model C: Science-as-a-Service¶
Concept: Professional astronomers and students who need photometric monitoring "buy" telescope time on the OpenAstro network, similar to LCO but cheaper and using distributed amateur-grade hardware.
Target customers: 1. PhD students needing follow-up photometry for TESS candidates — budget ~$500-2,000 from their advisor's grant 2. Postdocs running monitoring campaigns who need cheap continuous coverage — budget ~$5,000-15,000 from their research grant 3. Faculty PIs who include OpenAstro monitoring in their NSF/NASA grant budgets as a sub-award — budget ~$10,000-50,000 over 3 years
Pricing analysis:
LCO charges approximately: - 0.4m telescope: ~$15/hour - 1.0m telescope: ~$50/hour - Typical transit observation (3-4 hours): $45-200
OpenAstro competitive pricing: - Per-target per-night monitoring: $5-15 (automated scheduling, photometry delivered as calibrated light curve CSV within 24 hours) - Monthly monitoring subscription (1-5 targets, all available nodes): $100-300/month - Campaign package (50 transit observations across 6 months): $500-1,000
Quality guarantees required: - Delivered data products: calibrated FITS frames + differential photometry light curves + quality flags - Guaranteed photometric precision: <=5 mmag for V<12 targets (TTV-quality nodes); <=10 mmag for V<14 (full-capability nodes) - Turnaround: data available within 24 hours of observation - Weather-loss refund or credit policy - Instrument characterization: zero-point, color terms, noise floor per node (the instrument registry from Stage 1)
Would professional astronomers trust amateur-grade data for their papers?
Yes, with caveats. Precedent is strong: - AAVSO data is routinely used in professional publications (hundreds of papers per year cite AAVSO) - ExoClock (amateur exoplanet transit timing) publishes in A&A and feeds into the ESA Ariel mission planning - The Unistellar Network published in PASP using citizen science eVscope data - The TESS Follow-up Observing Program (TFOP) Working Group relies heavily on amateur data
The key requirements are: (a) published data reduction methodology, (b) transparent quality flags, (c) the instrument registry documenting each node's characterized performance, and (d) at least one methods paper demonstrating the pipeline on known targets (the Stage 1 deliverable).
Science-as-a-service revenue assessment:
| Network Size | Year 1 | Year 2 | Year 3 |
|---|---|---|---|
| 20 nodes | $2K-5K | $5K-15K | $10K-25K |
| 50 nodes | $5K-15K | $15K-30K | $25K-50K |
Revenue ramps slowly because trust must be earned. The first 1-2 years generate minimal revenue until word-of-mouth and publication record establish credibility. LCO took years to reach full utilization even with professional-grade hardware.
Model D: Space Situational Awareness (SSA)¶
Market reality: The SSA market is real, growing, and well-funded. The US Space Force budget for space domain awareness exceeds $2 billion/year. ESA's Space Safety Programme budget is ~EUR 600M over 5 years. Commercial satellite operators collectively spend hundreds of millions on conjunction assessment services.
What OpenAstro could provide: 1. Optical position measurements: Right ascension, declination, time — standard format (CCSDS TDM/CDM). Useful for catalog maintenance and orbit refinement. 2. Photometric light curves: Brightness vs. time for debris characterization (tumble rate, shape, material properties). This is the niche where OpenAstro's photometric calibration pipeline provides genuine value. 3. Anomaly detection: Monitoring known GEO satellites for unexpected maneuvers or attitude changes.
Existing competitors are formidable: - ExoAnalytic: 300+ purpose-built telescopes, years of track record, established government contracts - Numerica: Government contracts, deep integration with US Space Command systems - LeoLabs: Radar + optical, commercial satellite operator contracts
Barriers to entry: - Certification: Selling to US government requires ITAR compliance, facility clearances, and potentially classified data handling. This is extremely expensive and incompatible with an open-science mission. - Data format compliance: CCSDS standards for Tracking Data Messages (TDM) and Conjunction Data Messages (CDM) must be implemented. - Reliability: Government and commercial customers require >95% uptime guarantees and service level agreements. A 20-node network with volunteer-hosted nodes cannot guarantee this. - Legal entity: Government contracts typically require a registered business entity (LLC minimum), not a nonprofit or individual.
Assessment:
SSA is the highest potential revenue stream but requires a fundamentally different operational model than the science mission. The most realistic path is:
- Near-term (Year 1-2): Provide debris light curve data to ESA's Space Safety Programme or academic SSA researchers as a data contributor, not a contracted service. Revenue: $0-5K/year (essentially data donations for credibility).
- Medium-term (Year 3-5): If OpenAstro deploys dedicated SSA nodes (security camera + wide-field lens, $150-300 each, separate from science nodes), offer tumble characterization as a paid service. Revenue: $10-30K/year.
- Long-term (Year 5+): If the network reaches 100+ nodes with a meaningful fraction dedicated to SSA, compete for subcontracts from ExoAnalytic, LeoLabs, or government-funded SSA programs. Revenue: $50K-200K/year. This requires a for-profit entity or a separate commercial subsidiary.
Honest assessment: Pursuing SSA as a primary revenue source at Stage 3 scale is premature and would distract from the science mission. It is a viable long-term revenue stream if and only if the network reaches significant scale (50+ nodes) and the organizational structure permits commercial contracts.
Model E: Education and Public Engagement¶
Concept: Schools, universities, and science centers pay for access to the OpenAstro network for student projects, coursework, and public engagement programs.
Precedent: - Las Cumbres Observatory's Global Sky Partners program connects schools to robotic telescopes. LCO charges institutional partners for access. - Faulkes Telescope Project (UK): Provided free telescope access to schools, funded by the Dill Faulkes Educational Trust. Demonstrated massive demand but struggled with sustained funding after the initial endowment. - MicroObservatory (Harvard-Smithsonian CfA): Free robotic telescope access for educators. Grant-funded. - Slooh: Subscription-based commercial telescope access. Charged $5-25/month for individual access, higher for institutional. Went through financial difficulties; demonstrates the challenge of monetizing educational telescope access.
Pricing model: - Individual educator license: $200-500/year (access to live data feed, pre-packaged lesson plans, student project templates) - Institutional license (school/department): $500-2,000/year (multiple user accounts, API access, branded portal) - University research course license: $1,000-5,000/year (raw FITS access, integration with coursework, student co-authorship pathway)
Revenue projection:
| Year | Institutions | Revenue/Institution | Total |
|---|---|---|---|
| Year 1 | 5-10 | $500-1,000 | $2,500-10,000 |
| Year 2 | 15-30 | $500-1,500 | $7,500-45,000 |
| Year 3 | 30-60 | $500-2,000 | $15,000-120,000 |
Volume needed for meaningful revenue: 50-100 institutions at $1,000/year average = $50K-100K/year. This is achievable but requires a dedicated education outreach person (part-time or volunteer) to develop curriculum materials and maintain institutional relationships.
Assessment: Low revenue per customer, slow ramp, but highly aligned with mission and consistent with nonprofit status. Education revenue also strengthens grant applications ("broader impacts" section of NSF proposals). This is a complementary revenue stream, not a primary one. The main risk is that educators expect free access (because competing offerings like MicroObservatory are free), and the sales cycle for institutional purchases is 6-12 months.
Revenue Model Summary¶
| Model | Year 1 Revenue | Year 3 Revenue (25 nodes) | Year 3 Revenue (50 nodes) | Difficulty | Mission Alignment |
|---|---|---|---|---|---|
| A: Grants | $0-50K | $50K-200K | $100K-300K | High (competitive, uncertain) | High |
| B: Data licensing | $0-5K | $10-30K | $20-80K | Medium | Medium |
| C: Science-as-a-service | $2K-5K | $10K-25K | $25K-50K | Medium | High |
| D: SSA | $0 | $5K-15K | $10K-30K | Very High | Low-Medium |
| E: Education | $2K-10K | $15K-50K | $25K-100K | Medium | Very High |
Part 4: Three-Scenario Financial Model¶
Assumptions¶
- "Year 1" = first year of Stage 3 operations (owned hardware deployed)
- All costs and revenues cumulative over 3 years unless noted as annual
- Node capital costs assume TTV-quality tier ($2,000/node average) for the realistic and optimistic scenarios; low-cost tier ($1,200/node) for pessimistic
- Operating costs include both per-node OpEx and central infrastructure
- Grant funding listed as total received over 3 years (not annual)
- Revenue figures for data licensing, SaaS, and education are Year 3 annual run rates; cumulative revenue over 3 years is approximately 2x the Year 3 rate (ramp-up in earlier years)
Three-Year Financial Model¶
| Parameter | Pessimistic | Realistic | Optimistic |
|---|---|---|---|
| Network size (Year 3) | 10 nodes | 25 nodes | 50 nodes |
| Node cost (per node, avg) | $1,200 | $2,000 | $2,500 |
| Total capital cost (nodes) | $12,000 | $50,000 | $125,000 |
| Assembly/shipping/deployment | $3,000 | $10,000 | $25,000 |
| Total CapEx | $15,000 | $60,000 | $150,000 |
| Per-node OpEx (annual avg) | $200 | $250 | $300 |
| Node OpEx (annual, Year 3) | $2,000 | $6,250 | $15,000 |
| Central infra (annual, Year 3) | $600 | $2,000 | $25,000 |
| Central infra note | All volunteer labor | All volunteer labor | Part-time paid DevOps |
| Total OpEx (annual, Year 3) | $2,600 | $8,250 | $40,000 |
| Cumulative OpEx (3 years) | $6,000 | $18,000 | $80,000 |
| Total costs (3-year) | $21,000 | $78,000 | $230,000 |
| Revenue (3-year cumulative): | |||
| Grant funding | $0 | $100,000 | $250,000 |
| Data licensing (cumulative) | $0 | $15,000 | $80,000 |
| Science-as-a-service (cumul.) | $0 | $15,000 | $50,000 |
| Education (cumulative) | $0 | $20,000 | $60,000 |
| Crowdfunding/donations | $5,000 | $15,000 | $30,000 |
| Total revenue (3-year) | $5,000 | $165,000 | $470,000 |
| Net cash position (Year 3) | -$16,000 | +$87,000 | +$240,000 |
| Annual burn rate (Year 3) | $2,600 | $8,250 | $40,000 |
| Annual revenue run rate (Yr 3) | $0-2,000 | $35,000-50,000 | $120,000-170,000 |
| Self-sustaining? | No | Yes (if grants renew) | Yes |
Break-Even Analysis¶
Break-even condition (annual): Annual recurring revenue >= Annual operating cost + (Capital replacement reserve / average node lifetime)
Assuming 7-year average node lifetime before full replacement: - At 25 nodes: Annual OpEx ($8,250) + Capital reserve ($50,000/7 = $7,100) = $15,350/year needed - At 50 nodes: Annual OpEx ($40,000) + Capital reserve ($125,000/7 = $17,900) = $57,900/year needed
Break-even without grants: - 25-node network: Needs ~$15K/year non-grant revenue. Achievable in Year 3 of the realistic scenario (data licensing $10K + SaaS $10K + education $15K = $35K). - 50-node network with paid DevOps: Needs ~$58K/year non-grant revenue. Achievable only in the optimistic scenario. The critical dependency is whether education and SaaS revenue scale as projected.
The uncomfortable truth: In the pessimistic scenario, OpenAstro burns $16K over 3 years with minimal revenue. This is survivable if the initial capital comes from a single grant or a small number of donations, but the network would likely contract. In the realistic scenario, grant funding bridges the gap while non-grant revenue ramps up, and the network becomes self-sustaining around Year 3-4. In the optimistic scenario, the surplus funds further expansion.
Part 5: Non-Profit vs. For-Profit¶
Option 1: US 501(c)(3) Non-Profit¶
Advantages: - Eligible for NSF, NASA, and all foundation grants without a university intermediary (no 50-60% overhead) - Tax-deductible donations from individuals and corporations - More credible to the scientific community (perceived as mission-driven, not profit-driven) - NumFOCUS fiscal sponsorship as a bridge to full 501(c)(3) status - Eligible for cloud research credits, hardware donations, and educational discounts
Disadvantages: - Cannot distribute profits to founders or investors - Board governance requirements (3+ board members, annual meetings, IRS Form 990 filing) - Difficult to pivot to commercial SSA contracts without creating a separate entity - Startup cost: ~$500-1,000 in filing fees, 3-6 months processing - Cannot raise equity investment
Option 2: UK Charitable Incorporated Organisation (CIO) or Charity¶
Advantages: - Similar to US 501(c)(3) but simpler governance for small organizations - Eligible for UK Research Council grants, Wellcome Trust, ESA Prodex - Gift Aid (UK tax relief on donations — adds 25% to donation value) - Strong credibility in UK/European academic circles
Disadvantages: - Same limitations as US nonprofit on commercial activity - Less relevant for NSF/NASA grants (requires US partner regardless) - Charity Commission reporting requirements
Option 3: For-Profit LLC/Ltd¶
Advantages: - Simplest to establish (can be done in a day for ~$50-200) - Can freely pursue commercial contracts (SSA, data licensing, consulting) - Can raise equity investment from angel investors or VCs - No board governance requirements for single-member LLC - Can pay founders
Disadvantages: - Ineligible for most foundation grants (NSF, NASA, Heising-Simons, Sloan all require nonprofit or university) - Donations are not tax-deductible for donors - Less credible to scientific community (perceived as extractive) - Investors may push for commercial focus that conflicts with science mission - Revenue is taxed
Option 4: Social Enterprise / B-Corp / CIC¶
Advantages: - Legal commitment to mission alongside commercial operations - UK Community Interest Company (CIC) is cheap to set up and can receive some grants - B-Corp certification signals mission alignment to investors and partners - Can pursue commercial revenue while maintaining science credibility
Disadvantages: - B-Corp certification is a marketing label, not a legal structure (must still be LLC or Corp underneath) - CIC cannot distribute more than 35% of profits to non-asset-locked entities - Less well-understood by grant agencies — may still be treated as "for-profit" for grant eligibility
Recommendation¶
Phase 1 (Now through Stage 2): Use NumFOCUS fiscal sponsorship or Open Collective. This costs 5-10% of incoming funds but provides immediate 501(c)(3) equivalence for grant applications and tax-deductible donations, with zero governance overhead. Apply to NumFOCUS as an Affiliated Project once the pipeline is open-source and has a user base. This is the lowest-friction path and is exactly what projects like Astropy, NumPy, and other scientific software use.
Phase 2 (Stage 2-3 transition, ~12-24 months after first paper): Incorporate as a US 501(c)(3) nonprofit. The trigger point is when incoming funds exceed ~$50K/year, at which point the 5-10% fiscal sponsorship fee exceeds the cost of running your own 501(c)(3). The nonprofit structure maximizes grant eligibility and scientific credibility, which are the primary funding sources in the realistic scenario.
Phase 3 (If SSA or commercial revenue becomes significant, Year 3+): If commercial revenue (SSA, data licensing to industry) exceeds ~$50K/year and is growing, consider creating a for-profit subsidiary (LLC) wholly owned by the 501(c)(3). This is a well-established structure used by organizations like Mozilla (Mozilla Foundation owns Mozilla Corporation). The nonprofit retains the science mission, grant eligibility, and tax-exempt status. The for-profit subsidiary handles commercial contracts, can hire employees at market rates, and remits profits to the parent nonprofit.
Do not start as a for-profit. The realistic revenue path for the first 3-5 years is dominated by grants and donations, both of which require nonprofit status. Starting as a for-profit and converting later is much harder than starting as a nonprofit and adding a commercial subsidiary.
Part 6: The Chicken-and-Egg Problem¶
The Core Dilemma¶
To generate revenue, OpenAstro needs hardware producing data. To buy hardware, OpenAstro needs revenue. To get grants, OpenAstro needs published results from hardware. To publish results, OpenAstro needs hardware.
This is the standard problem for every hardware-dependent citizen science project. The projects that solved it (GMN, PANOPTES, SatNOGS) all used the same basic strategy: start with someone else's hardware, prove the concept, then deploy your own.
Proposed Bridge Strategy¶
Step 0: Pre-Stage-3 (Now — Parallel with Stage 1-2) - Build and publish the pipeline using archival data (Stage 1). Cost: $0 (time only). - Recruit 10-50 volunteer sites using their own hardware (Stage 2). Cost: $0-500 (server hosting). - Produce 1-2 publications. Cost: $0. - This is the credibility engine. It costs almost nothing and is the prerequisite for everything that follows.
Step 1: Proof-of-Concept Owned Node (Month 0-3 of Stage 3) - Build ONE owned node at the founder's location or a partner institution, using the low-cost science tier ($1,200-1,500). Fund from personal savings or a small donation. - Demonstrate full autonomous operation: auto-open, auto-observe, auto-close, auto-upload, auto-calibrate. - Publish a technical note (RNAAS or arXiv) describing the autonomous node design and its performance benchmarks. - Cost: $1,200-1,500. This is the most important $1,500 in the entire project.
Step 2: Seed Funding Round (Month 3-9) - With a working prototype and Stage 2 publications, apply for: - Sloan Scientific Software grant ($100K-500K) — pitch the calibration pipeline - AAS Small Research Grant ($5K) — fund travel and 2-3 more nodes - Raspberry Pi Foundation community fund ($1K-5K) — fund the RPi-based node design - AWS/Google/Azure research credits ($10K-50K in cloud compute) - Simultaneously launch a small crowdfunding campaign (see below). - Target: $15,000-30,000 in combined funding within 12 months of prototype deployment.
Step 3: First Owned Network (Month 9-18) - Deploy 6-8 nodes at volunteer host sites using seed funding. - Leverage volunteer hosts: offer free hardware in exchange for hosting, power, and internet. The host gets co-authorship and a cool telescope in their yard/club. This is the PANOPTES model and it works. - Target the first multi-site result from owned hardware. - Cost: $9,000-20,000 in hardware. Funded by seed round.
Step 4: Publication and Major Grant Application (Month 18-30) - Publish the first paper using owned hardware data. - Submit NSF ATI application (PI at US university, OpenAstro as sub-award or direct applicant if 501(c)(3) established). - Submit Heising-Simons letter of inquiry. - Begin science-as-a-service beta program (offer free monitoring to 3-5 professional collaborators, then convert to paid). - Cost: $0 (existing operations). Revenue begins from SaaS and education.
Step 5: Scale (Month 30-48) - If major grant funded: deploy 15-25 additional nodes. - If not funded: sustain existing 6-8 nodes from operating revenue + small grants + donations. This is the pessimistic scenario — survivable but not growing. - The decision point: If by Month 36, total annual revenue (all sources) is <$10K, the owned hardware network is not economically viable at scale and should be maintained at 6-10 nodes indefinitely, supplemented by the volunteer network.
Is Crowdfunding Realistic?¶
Precedent from comparable projects:
| Project | Platform | Amount Raised | Year | Offering |
|---|---|---|---|---|
| PANOPTES | Indiegogo | ~$10,000 | 2016 | Build units for schools |
| SatNOGS | Hackaday Prize | $10,000 (prize) | 2014 | Open-source ground station |
| Planetary Resources (Arkyd) | Kickstarter | $1.5M | 2013 | Space telescope (failed to deliver) |
| ChipStar (amateur radio sat) | Various | ~$5,000 | 2020 | Fund a cubesat component |
| Open Source Ecology | Various | $100K+ cumulative | 2011-present | Open-source hardware designs |
| Pale Blue Dot campaign | Kickstarter | $40K | 2019 | Citizen science image analysis |
Assessment: Crowdfunding for science hardware typically raises $5,000-30,000 for projects with strong community engagement and a clear "buy a node" or "adopt a telescope" offering. OpenAstro could realistically raise $10,000-25,000 on a first Kickstarter/Indiegogo campaign with the pitch: "Fund a telescope node. It will bear your name. Your data contributes to real science. You are a co-author."
The Patreon model: Monthly recurring donations of $5-25/month from 50-200 supporters = $3,000-60,000/year. This is more sustainable than one-off crowdfunding but requires ongoing community engagement (monthly updates, data highlights, "your telescope saw this" reports). At 100 patrons averaging $10/month = $12,000/year, which covers operating costs for a 10-25 node network. This is realistic based on Patreon data for science/engineering projects.
Recommended crowdfunding strategy: 1. Launch a Kickstarter after the prototype node is working (Step 2 above). Target: $15,000. Reward tiers: $50 = name on the website, $150 = named acknowledgment on data from a specific node, $500 = "adopt a telescope" — your name on a physical node + co-authorship eligibility, $1,500 = fund an entire node at a location of your choosing. 2. After Kickstarter, launch a Patreon for ongoing support. Target: 100+ patrons. Monthly content: data highlights, node status updates, "what the network saw this month." 3. These are supplements to, not replacements for, grant funding and earned revenue.
Part 7: Risk Register¶
| Risk | Probability | Impact | Mitigation |
|---|---|---|---|
| No major grant funded in first 3 years | 40-50% | High | Multiple small grants + crowdfunding + personal funding to sustain minimum network. Keep CapEx low. |
| Node failure rate higher than estimated | 20-30% | Medium | Standardized design with cheap replaceable parts. Raspberry Pi and cameras are commodity items. Keep $100/node/year maintenance reserve. |
| Volunteer node hosts lose interest or move | 30-40% | Medium | Written hosting agreements. Nodes designed for easy relocation. Spare hosts identified in advance. |
| Science-as-a-service revenue fails to materialize | 40-50% | Medium | SaaS is a supplement, not a primary revenue source. Network survives on grants + donations at 10-node scale. |
| ExoAnalytic or similar company enters the amateur science space | 10-20% | High | OpenAstro's open-data, open-source model serves a different market (science) than commercial SSA. Differentiate on mission, not capability. |
| Key person risk (founder leaves/burns out) | 20-30% | Very High | Document everything. Open-source all software. Build a core team of 3-5 committed people before deploying hardware. The project must survive losing any one person. |
| Data quality insufficient for professional publication | 15-25% | High | The Stage 1 pipeline and Stage 2 volunteer network validate data quality BEFORE spending money on owned hardware. Do not skip stages. |
Part 8: Decision Framework¶
Deploy owned hardware when ALL of the following are true:¶
- Stage 1 paper is published (pipeline validated)
- Stage 2 network has >=10 active volunteer sites (community exists)
- At least one multi-site result has been published (network produces science)
- A working prototype owned node has been built and tested for >=3 months
- At least $10,000 in committed funding (grants, crowdfunding, or donations) is available for hardware
- A legal entity (fiscal sponsorship or 501(c)(3)) is in place to receive and manage funds
- At least 3 committed host sites have been identified (club, institution, or individual with power/internet/outdoor space)
Do NOT deploy owned hardware if:¶
- The pipeline is not validated on real data
- The volunteer network has fewer than 5 active sites
- No publication exists
- Total available funding is less than $5,000
- The founder is the only person working on the project
The single most important principle: Stage 3 is not the goal. The goal is publishable science. If Stage 2 (volunteer network) produces excellent science without owned hardware, that is success. Owned hardware is a means to fill specific coverage gaps, not an end in itself. Do not build hardware to solve a problem that does not yet exist.
Summary¶
The economic path for OpenAstro's owned hardware stage is narrow but viable:
- Capital costs are manageable: $1,200-2,500 per science-quality node. A minimum viable network of 6-8 nodes costs $10,000-20,000.
- Operating costs are low: $200-300/node/year, plus $500-2,000/year central infrastructure with volunteer labor. A 25-node network costs ~$8,000/year to run.
- Grant funding bridges the gap: $100K-200K over 3 years from NSF/Heising-Simons/Sloan is realistic but not guaranteed.
- Non-grant revenue is possible but slow: Science-as-a-service + education + data licensing can reach $30-50K/year at 25 nodes by Year 3. This is sufficient to cover operating costs but not rapid expansion.
- SSA is a long-term opportunity, not a near-term revenue source. Do not distract from the science mission to chase SSA revenue before the network has 50+ nodes.
- Non-profit structure is correct. Use NumFOCUS fiscal sponsorship now; incorporate as 501(c)(3) when funds exceed $50K/year; add a for-profit subsidiary later if commercial revenue warrants it.
- The bridge strategy is: prove with volunteers, prototype one owned node cheaply, use the prototype to raise seed funding, deploy 6-8 nodes, publish, and then apply for major grants.
The break-even point for a self-sustaining 25-node network (without reliance on grant renewals) requires approximately $15,000/year in recurring non-grant revenue. This is achievable by Year 3-4 in the realistic scenario but requires deliberate effort to develop paying customers in the science-as-a-service and education markets.
The biggest risk is not financial — it is premature scaling. Deploying 50 nodes before the pipeline is proven, the community is built, and the revenue model is validated wastes money. Deploy 1 node. Then 6. Then 25. Each step must justify the next.