Inspiration
Most financial systems assume communication is effectively instant. But across the solar system, a message can take minutes or hours to arrive, routes change as planets move, the Sun can block links, and packets can be lost. We wanted to see whether an exchange could still be fair, solvent, and useful when latency is a fundamental physical constraint rather than a networking inconvenience.
What it does
MultiPlanetary Exchange (MPEX) is a market and clearing system designed for interplanetary communication.
MPEX combines:
- Orbital-aware routing through two relay nodes
- Moving-receiver light-time calculations
- Sun-exclusion and maintenance-aware routes
- Local planetary exchanges with a central clearing house on Ceres
- Fair batch auctions that do not reward traders simply for being physically closer to Ceres
- Reserve-before-send asset protection
- Delivery-versus-payment settlement
- Packet retries, deduplication, and idempotent financial operations
- Futures, margin calls, funded default handling, and a guarantee fund
- Automated scenario testing, ledger traces, and long-horizon orbital analysis
The result is a market where hours of communication delay do not automatically create double spending, unfair priority, or inconsistent financial state.
How we built it
We built the project in layers.
First, we modeled the solar-system communication network using orbital data. Routes are calculated using where a receiver will be when a transmission arrives, rather than treating planets as stationary. Links can become unavailable because of the Sun, scheduled maintenance, or communication constraints.
On top of that network, we built the exchange. Each settlement has a local market and ledger. Assets are reserved before cross-planet messages leave their home exchange. Orders are then cleared through a global batch auction on Ceres using deterministic price/time/pro-rata rules, followed by delivery-versus-payment settlement.
We also built futures, collateral, margin calls, default handling, retries, duplicate protection, and conservation checks.
Finally, we created an evidence framework that generates traces, ledgers, orbital scans, shifted-epoch tests, routing results, and visualizations so our claims can be reproduced rather than simply asserted.
Challenges we ran into
The hardest part was realizing that networking and finance could not be designed independently.
A packet being delayed is not just a networking problem when that packet represents money, shares, an order, or a margin call. We had to track what happens financially while information is still traveling.
Choosing the clearing-house location was another challenge. Earth initially seemed like the natural choice, but our simulations showed that latency alone did not settle the question. Long-horizon reliability favored Ceres, so we changed the design based on our results.
We also had to handle retries without creating duplicate financial effects, preserve fair auction priority despite unequal communication delays, and test the system as orbital geometry changed over time.
Accomplishments that we're proud of
We are especially proud that MPEX is more than a static design or visualization: the market, network model, settlement rules, and evidence system run together.
Our test suite passes 17/17 scenarios, while our expanded evidence set contains 18 runs with cash and shares conserved at every step and no over-encumbrance.
We also scanned 200 years of orbital conditions, refined thousands of Sun-blockage boundaries, tested shifted epochs, modeled packet loss and duplicate delivery, and demonstrated funded default handling.
Our simulations also changed our own design: the evidence led us to move the central clearing house from Earth to Ceres.
What we learned
The biggest lesson was that distance alone does not determine the best architecture.
A slightly faster route can be a worse system if it creates greater outage exposure, unfair priority, or weaker financial guarantees. Reliability, capital usage, packet cost, settlement finality, and fairness all have to be evaluated together.
We also learned the importance of making distributed financial operations idempotent. When a message can be lost, retried, or delivered twice, the financial system must still produce exactly one effect.
Most importantly, testing was not something we added after designing MPEX. The evidence changed the design itself.
What's next for MultiPlanetary Exchange
The next step is improving automatic recovery from long communication outages.
MPEX currently implements retries, deduplication, routing around ordinary failures, and safe financial recovery. A major future improvement is proactive relay re-pinning: predicting a long blockage and moving an active session to the alternate relay before connectivity is lost.
We would also expand the orbital simulations, optimize communication and capital efficiency, add more market participants and products, and continue turning the simulator into a richer interactive environment for exploring how markets might operate when the speed of light becomes part of the market structure.
Log in or sign up for Devpost to join the conversation.