Euwyn Poon built Spin into one of the largest micro-mobility companies in the United States — 250,000 scooters deployed across dozens of cities, a distributed hardware operation running in real time across multiple jurisdictions. Then Ford acquired Spin, and Poon started thinking about orbit.
His new company is called Orbital. The seed round: $5 million. The stated goal: 10,000 space-based data centers. The reaction from most people who hear this for the first time is somewhere between skeptical and confused. That reaction is probably the wrong one.
Space data centers are computing infrastructure hosted on satellites or orbital platforms instead of ground-based facilities. Rather than a warehouse full of servers in Virginia or São Paulo consuming enormous amounts of power and water for cooling, you deploy smaller modular units into low Earth orbit.
The pitch addresses several real constraints at once. Space solar energy is abundant and doesn't require connection to terrestrial grids. LEO satellites can serve large geographic areas without costly cable infrastructure buildout. There's no land to acquire, no grid capacity to negotiate, no water-intensive cooling systems to engineer.
$5 million is a seed round, not a deployment budget — Orbital is still in early stages. But the direction it signals, and the founder behind it, deserve more than a passing glance.
Poon's background in micro-mobility isn't as unrelated to data infrastructure as it looks. Running a network of 250,000 connected physical devices across cities in real time is, in practical terms, a distributed systems problem. It demands serious thinking about connectivity, edge computing, operational data pipelines, and what breaks when things at the edge go wrong. That kind of operational experience at scale teaches things that a software-only background doesn't.
The broader context also matters. Data center demand is growing faster than legacy infrastructure can accommodate, driven primarily by AI workloads. Energy grids in key markets are constrained. Land in urban corridors is expensive and scarce. Water availability for cooling is a growing concern in climate-stressed regions. Traditional hyperscaler infrastructure is hitting real physical limits, and serious money is beginning to look for different answers.
Orbital is one bet in that direction. It won't be the last.
The Orbital story is a useful lens for a larger infrastructure shift that business leaders should be tracking — not because space data centers are coming for your cloud bill next year, but because of what the bet reveals about where enterprise compute infrastructure is heading.
Decentralization is accelerating. The model of three or four hyperscalers owning all meaningful compute is under pressure — from sovereign cloud initiatives, from edge computing buildouts, from distributed AI inference, and now from orbital alternatives. The concentration isn't going to reverse overnight, but the trajectory is clear.
Infrastructure is becoming a strategic variable. A few years ago, a CEO's data infrastructure question was simple: which cloud vendor? Increasingly, it means understanding where your data physically lives, who controls the underlying infrastructure, what your cost exposure looks like in five years, and what happens when any of that changes.
Alternative models gain relevance when constraints compound. Energy costs, connectivity gaps, data sovereignty regulations, and FX exposure on dollar-denominated cloud bills don't each individually force a rethink. Together, they do.
For companies in Colombia, Mexico, Argentina, Chile, and across the region, this isn't a distant technology story. The infrastructure constraints that make space-based compute an interesting long-term bet are versions of problems that already show up in operational decisions today.
Cloud costs and currency exposure. Most LATAM companies pay for cloud infrastructure in U.S. dollars. As AI workloads grow — and they will — that cost line grows with them. Any structural shift in where compute lives and how it's priced is directly relevant to the economics of running AI-driven operations in the region.
Connectivity gaps. Logistics companies, agricultural operations, mining, energy, and infrastructure operators across the region face real connectivity limitations in field operations. The long-run potential of space-based compute to address geographic coverage gaps isn't hypothetical — it's the exact use case that makes LEO infrastructure investment interesting to enterprise buyers.
Data sovereignty and compliance. Regulatory frameworks around data residency are tightening across LATAM — Brazil's LGPD, Colombia's evolving data protection landscape, Mexico's regulatory trajectory. Where your data physically resides is becoming a compliance question, not just a technical preference. Space data centers add a genuinely novel dimension to that conversation, one that regulators haven't fully addressed yet.
None of this means your infrastructure planning should wait on Orbital. It means the infrastructure layer of your AI and data strategy deserves more active attention than it typically gets in a quarterly business review.
For business and operations leaders who want to track this intelligently:
Orbital is early. Space data centers won't replace your cloud infrastructure this planning cycle. But the signal embedded in this story — that serious operators are rethinking where computing infrastructure belongs and why — is worth internalizing before the shift becomes obvious.
The companies that navigate the next infrastructure transition well are the ones whose data and AI strategies aren't built on the assumption that today's cloud economics and today's providers will look the same in five years. That kind of strategic flexibility starts with a clear-eyed view of where you have dependencies today.
If your AI, automation, or BI roadmap doesn't include a perspective on infrastructure cost trajectory and lock-in risk, it's missing a layer. The decisions you make about your data architecture today compound — in both directions.
Schedule a free consultation with our team and discover how AI can transform your operations.
Schedule a consultation
TrendsAITwo ex-Goldman and Meta founders built voice AI for Africa and the Middle East — now at 17K+ calls/day. What LATAM operators should take from this.
