A useful structure can be hard to reproduce.
The right crystal form, connected pore network or ordered particle lattice may depend on a narrow processing window.
CAL HACKS 13.0 · SPACEXAI CHALLENGE CONCEPT
The next
factory floor
is in orbit.
A network of specialized materials factories above Earth. Shared infrastructure. Intelligent production. Extraordinary possibilities for life below.
THE PROPOSAL
Make orbital manufacturing a repeatable service that independent labs can use.
THE AMBITION
Concept illustrationFrom one successful experiment to an industrial network.
Not an operational fleet
01 / THE PROBLEMThe molecule is only
The molecule is only
half the material.
How a material forms can determine how it works. Structure, uniformity and stability can become bottlenecks between a promising substance and a useful product.
A PROCESS VARIABLE WE USUALLY CANNOT SWITCH OFFSchematic
↓
On EarthSettling and buoyant flow can redistribute material during formation.
µgIn microgravity
Reduced gravity-driven transport changes the available process conditions.
Illustrative physics, not experiment imagery. Microgravity does not remove every flow or defect, and it does not improve every material.
One flight does not make a supply chain.
A result still needs matched comparisons, repeatable recipes, return handling and functional qualification.
Every lab faces a second engineering problem.
Containment, power, heat rejection, sensing, control, scheduling and traceability sit between its chemistry and a usable orbital experiment.
THE OPPORTUNITY
Build expertise in the manufacturing process—and let many material programs benefit from it.
02 / THE EVIDENCEThe experiments exist.
The experiments exist.
The industries are real.
Six process programs connect pharmaceuticals, cooling, resource recovery, chemical production and sensing. Select a program to separate its demonstrated result from our proposed next step.
Flight result Flight research / Earth function Our untested extension
CURRENT EXAMPLE · 20 OCT 2026
Varda’s W-7 adds a specific oncology crystallization mission. It complements the earlier ritonavir recovery result and Merck’s distinct protein-crystal work. A launch announcement is not yet a material-performance result.
Organizations are shown as research participants, operators or application examples. They are not announced customers or partners of this concept.
03 / THE CONNECTIONDifferent chemistry.
Different chemistry.
A shared operating pattern.
Every program defines an initial state, creates a formation environment, observes change, preserves the result and measures quality. This is an engineering synthesis—not a claim that one recipe works for every material.
01PrepareFeedstock + target
02ConditionControlled environment
03FormGrow or assemble
04ObserveEstimate the state
05PreserveFinish + return
06QualifyFunction + next recipe
| Program | Prepare | Condition | Form | Observe | Preserve | Qualify |
|---|
PROTEIN CRYSTALS / PREPARE
Contained protein solution
SHARED
Interfaces, decisions and records.
Power and thermal budgets, mounting, telemetry, recipe execution, scheduling and batch identity can follow common conventions.
SPECIALIZED
Chemistry, equipment and acceptance.
Temperature envelopes, fluid paths, optics, containment and final quality measurements remain specific to the process and product.
04 / THE SOLUTIONA common platform.
A common platform.
Specialized factories inside.
Independent labs bring formulations and specifications. The foundry supplies process development, execution and a traceable path toward qualified production.
Conceptual module architecture. Geometry and capacity are illustrative.
Thermal processing
Controlled gradients, solidification and annealing.
NPG · selected small moleculesSolution growth
Contained dosing, concentration and substrate growth.
Protein crystals · MOF membranesParticle assembly
Mixing, evolving topology and timed stabilization.
Bijels · photonic colloidsThe network can use different launch, hosted-lab and return providers. The proposed manufacturing layer should remain portable across qualified vehicle interfaces.
05 / THE PRODUCTION INTELLIGENCERun the process
Run the process
by its state.
A recipe becomes more useful when it responds to what the material is doing. Calibrated measurements and validated controls sit between an observation and a physical action.
REAL NASA EXPERIMENT IMAGEDECLIC–DSI
Illustrative control logicNOT LIVE TELEMETRY
What happens when the process changes?
Observe→Estimate→Validate→Act
Choose the next batch.
Balance information gain, likely usefulness and scarce mission capacity. Early campaigns should reduce uncertainty about the mechanism.
Coordinate the capsule.
Schedule around power, heat rejection, imaging and vibration-sensitive intervals. More simultaneous jobs are not always more productive.
Learn from accepted output.
Connect every process history to Earth characterization. A structural proxy becomes useful only when linked to the property the customer needs.
06 / THE INDUSTRIAL LOGICShare the overhead.
Share the overhead.
Keep the processes precise.
Scale can come from pooled infrastructure, reusable interfaces and accumulated process expertise. Those are mechanisms to test—not a promised cost curve.
Spread common costs across compatible batches.
Platform engineering, integration and mission operations can serve more than one useful payload.
Apply shared capabilities across industries.
A common recipe format, telemetry system and scheduling engine can support multiple cartridge families.
Make the next run more informed.
Repeatable records help improve a validated process family. Cross-material transfer and cost savings still require evidence.
Interactive thought experimentUNITLESS ASSUMPTIONS
What does sharing
the fixed cost change?
1 batch12 hypothetical slots
Dedicated
Shared
lower modeled cost per accepted batch
under these equal-yield assumptions
under these equal-yield assumptions
Show the arithmetic and assumptions
Fixed mission cost F = 100 units; variable cost V = 20 per batch. Dedicated: (F + V) / q. Shared: (F / n + V) / q, where n is occupied compatible slots and q is accepted-batch yield.
The comparison assumes identical yield, no added coordination cost and sufficient power, thermal and return capacity. It excludes financing, regulation and additional infrastructure. It is not a quote, price estimate or forecast.
Durable componentsA small membrane or reactor core performs work repeatedly.
Process knowledgeA flight can improve Earth production without recurring orbital manufacturing.
Seeds or mastersUseful only when the relevant advantage survives proven replication.
07 / THE LONG VIEWOne recipe becomes a service.
One recipe becomes a service.
Many services become an industry.
The long-term vision is a network of specialized orbital production capacity, connected to laboratories, manufacturers and qualification facilities on Earth.
EarthFormulations · qualification · finishing · customers
OrbitSpecialized formation processes · recorded batches
The connectionLaunch · return · data · shared operating standards
Prove a process
One family. A matched Earth baseline. A useful measurement.
Gate: a reproducible advantageQualify the cartridge
Repeat runs, define acceptance and verify return survival.
Gate: consistent accepted outputOffer capacity
Run compatible customer campaigns through a common interface.
Gate: viable service economicsSpecialize the fleet
Allocate vehicles and schedules to validated process families.
Gate: demand and reliable operationsFuture-state concept. No operational fleet, production customers or capacity commitments are claimed.
08 / BUILT FOR THIS CHALLENGEThe route to orbit.
The route to orbit.
The intelligence to use it.
SpaceX provides relevant launch and return infrastructure. Grok can help turn research into structured, tool-backed decisions. Our proposed layer connects those capabilities to repeatable materials processes.
TRANSPORT & RETURN
SpaceX
Shared launch manifests and documented payload interfaces provide a concrete integration reference. Our extension: schedule compatible manufacturing payloads around a booked mission.
ISS research cargo return supplies an established logistics example for compatible hosted experiments. It is mission-specific integration, not pickup of arbitrary independent capsules.
The June 2026 demonstration directly connects SpaceX to microgravity access and return. Cartridges and production software could sit above emerging infrastructure; routine service availability is not assumed.
A future capacity scenario consistent with its reusable transportation design. The first experiment and near-term business case should stand without a particular Starship schedule or price.
RESEARCH & ANALYTICAL TOOLS
xAI / Grok
Extract materials, conditions, measurements and evidence into typed records. A valid schema helps integration; scientific content still requires verification.
Connect a research request to a data loader, numerical optimizer or simulation. Developer-controlled tools validate and execute the requested operations.
Help an operator interpret experimental imagery and investigate potential anomalies, alongside calibrated vision measurements. Scientific accuracy must be evaluated.
Paper → evidence record → candidate experiment → resource schedule → source-linked batch report. This presentation is an interactive concept, not a live Grok or spacecraft integration.
REUSESTANDARD INTERFACESTELEMETRYITERATIONVALIDATED AUTOMATION
These are design parallels and proposed integrations. No endorsement, affiliation or partnership with SpaceX or xAI is implied. Physical control remains inside validated limits.
09 / REAL DATA, BETTER DECISIONSAn archive of experiments.
An archive of experiments.
A starting point for a foundry.
NASA PSI contains images, videos, process records and supporting documents. The opportunity is to turn selected observations into reliable measurements and better experimental choices.
133investigation records returned
by the checked catalogue query
by the checked catalogue query
42.27 GBreported DECLIC–DSI archive
with a raw sample verified
with a raw sample verified
5.52 TBreported OASIS archive
for smectic-film dynamics
for smectic-film dynamics
Sizes are repository-reported totals in decimal units, not amounts downloaded or counts of independent observations. Snapshot checked 24 October 2026. Different runs, ground controls, metadata and raw imagery need separate treatment.
AWS can get the demo started.
DSI, OASIS and pk4 prefixes were visible in the public bucket. The full PSI catalogue should not be assumed to live in that mirror.
aws s3 ls s3://nasa-psi/ --no-sign-requestConnect structure to candidate selection.
CoRE MOF 2024 can support terrestrial structure and descriptor screening. Combine it with processing evidence without treating it as labeled orbital-performance data.
Let unexpected outcomes count.
ACE-T4 studied gels after a hardware-related delay changed its samples. Such records can help identify alternative uses and prevent an “orbit always wins” model.
How to build a defensible analysis pipeline
1. Preserve provenance. Keep accession, run, sample, conditions, file keys and calibration together. Distinguish version changes from a full file manifest.
2. Measure a narrow target. Begin with interface motion, particle order or assembly timing that the imagery actually supports.
3. Hold out entire runs. Adjacent video frames are correlated. Random frame splits can make a model look far more capable than it is.
4. Link to function. Morphology is a proxy until linked to cooling, separation, optics or another measured outcome.
5. Compare the strongest Earth route. Chemistry, apparatus and processing history can confound apparent gravity effects.
6. Choose the next informative experiment. Use uncertainty to allocate scarce orbital capacity; avoid a universal “space benefit” score trained on heterogeneous studies.
10 / FROM RESEARCH TO A SERVICELabs bring the material.
Labs bring the material.
The foundry builds the process.
Customers should buy a clearly defined outcome at each stage. Process development and repeat production are related services with different promises.
Process development
Design and run experiments around a specific gravity-sensitive mechanism.
- Customer receives
- Measurements, a better-supported recipe, or a reason to stop.
- Value unit
- An informative campaign.
Qualification
Repeat the process, preserve the useful structure and define acceptance.
- Customer receives
- A process window and evidence of repeatability.
- Value unit
- A qualified manufacturing route.
Recurring capacity
Manufacture against agreed specifications within a proven process family.
- Customer receives
- Traceable batches with material-specific qualification.
- Value unit
- Accepted output.
THE ECOSYSTEM IS ALREADY FORMING
Build a clear role within it.
Modular platforms and orbital manufacturing are already being pursued. Our proposed differentiation is cross-material recipe development, experiment/production scheduling, bounded control and evidence-linked batch qualification.
Space Cargo Unlimited / BentoBoxModular monitored payload platform; substantial overlap with shared services.
Redwire / SpaceMDAnnounced a 2028 Starfall mission with expected capacity for up to 32 PIL-BOX systems. A plan, not a completed mission.
G-SPACEExisting PSI image analytics and space/Earth comparison tools.
Space ForgeOrbital seeds followed by terrestrial scale-up; a useful architecture already under development.
Protect the customer’s IP. Separate private formulations and batch data. Share process knowledge only where authorized.
Earn the economic claim. Count integration, hardware, power, return, failures and qualification—not only launch cost per kilogram.
Start narrowly, design broadly. One cartridge family and a matched Earth baseline are enough to test the platform thesis.
11 / MAKE IT TANGIBLEFrom a scientist’s request
From a scientist’s request
to a qualified batch.
The hackathon demonstration should make the industrial future visible through one honest, working slice of it.
Three lab requests
Cooling structure, selective membrane and photonic scaffold.
One mission plan
Assign compatible cartridges, power and quiet processing windows.
Real experimental data
Extract one defensible measurement from a documented NASA sequence.
A bounded decision
Show a validated-rule response and its effect on the schedule.
A batch record
Trace evidence, assumptions, measurements and the next qualification test.
A 90-SECOND NARRATIVE
Some of the materials we need on Earth are limited by how they form. Gravity moves particles and fluids while crystals, pores and networks are taking shape.
Researchers have already used orbit to study cancer-drug formulations, pharmaceutical solid forms and ordered materials. But a successful experiment is only the beginning. Industry needs a process it can repeat.
Orbital Foundry is our proposal for that next layer: specialized manufacturing cartridges, a shared operating system and a network that connects laboratories on Earth to production capacity in space.
A lab brings a formulation and a target. We help select the experiment, schedule the process, observe its progress and connect the returned material to a complete batch record. SpaceX provides relevant routes to orbit and back; Grok can help organize the evidence and coordinate validated analytical tools.
At Cal Hacks, we make that future concrete with real NASA data and a transparent production-planning prototype. The ambition is a new industrial base above Earth—whose value is measured in better products and processes below.
A LONG-TERM AMBITION
Extraordinary materials.
Everyday impact.
12 / EVIDENCE LIBRARYDream ambitiously.
Dream ambitiously.
Keep the claims traceable.
Primary papers, official investigations, company announcements and API documentation. Sources establish the cited result or plan; the foundry and its proposed integrations remain a concept.
No sources match this search.
About this presentation
Prepared for Cal Hacks 13.0. Evidence checked 24 October 2026. PSI sizes and catalogue counts are an access-check snapshot, not a completed analysis. Scientific extensions, network illustrations, scheduling examples and the cost calculator are explicitly proposed or illustrative. No performance improvement, customer relationship or sponsor endorsement is implied.
This HTML includes its styles, scripts, data and graphics. It works offline; external citations require internet access. Use P for presentation mode, arrow keys to move between chapters, and Esc to return to the full story.