The Lyceum: Space Economy Weekly — Mar 19, 2026

The biggest rocket in the world is finally crawling toward its launch pad, but the week's real story is plumbing — not pyrotechnics. Governments on three continents wrote nine-figure checks for the unglamorous systems that modern economies and militaries actually depend on: Arctic weather satellites, orbital traffic rules, sovereign launch pads, and cryogenic fuel management in space. Meanwhile, filings for tens of thousands of orbital data-center satellites and China's formal designation of aerospace as a strategic industry made clear that the contest for who controls orbital capacity — and the rules governing it — is accelerating faster than most people realize.

NASA began rolling the Space Launch System rocket to Launch Pad 39B at Kennedy Space Center on March 19 — the final trip before Artemis II sends four astronauts around the Moon, targeted for launch as early as April 1. The four-mile crawl takes about 12 hours at one mile per hour. Simultaneously, crew members Reid Wiseman, Victor Glover, Christina Koch, and Canadian Space Agency astronaut Jeremy Hansen entered pre-flight quarantine.

This is the first time since Apollo that humans will venture to lunar space. A completed Flight Readiness Review gave teams the "go" to proceed, strengthening the April 1 target. But context matters: a hydrogen leak during a wet dress rehearsal earlier this year forced a scrub, and helium and upper-stage issues required extra troubleshooting during rollback. The program's schedule has been fragile enough that seeing the full stack vertical and moving is, itself, the most tangible proof of progress available. All eyes now shift to final pad checkouts. If those go cleanly, early April is real. If they don't, the pattern of delays that has defined Artemis continues.

Rocket Lab announced a $190 million contract to fly 20 hypersonic test missions for the U.S. Department of Defense using its HASTE rocket — a modified version of the Electron vehicle designed specifically for suborbital flight testing. The company says the first mission could fly within months.

This matters for two reasons beyond the dollar figure. First, hypersonic weapons testing has been bottlenecked by a lack of available flight opportunities, and Rocket Lab is effectively selling "test range as a service" at a tempo legacy providers can't match. Second, a block buy of 20 launches pushes the company's total backlog past $2 billion and provides the kind of stable, recurring revenue that could change how investors price the stock. The Pentagon is moving toward a commercial model for some R&D, amid which Rocket Lab is a primary beneficiary.

Separately, Rocket Lab secured $120 million to develop a dedicated Neutron launch pad at Wallops Island, easing East Coast range congestion for its medium-lift vehicle and signaling domestic infrastructure investment that could reshape competitive pricing for constellation deployments.

Multiple Falcon 9 missions this week pushed the Starlink network past 10,000 active broadband satellites in low Earth orbit. One reused first stage completed its 14th flight. If you sell satellite capacity, compete for spectrum, or worry about orbital coexistence, the math just changed — one operator now has five figures of active spacecraft overhead.

The FCC granted conditional approval for a new Starlink LEO shell but attached a tightened debris-mitigation requirement: a three-year post-end-of-life deorbit window, shorter than previous rules. That decision signals regulators are getting stricter on lifecycle rules even as they approve expansion — a dynamic that raises replacement and operational costs for any high-cadence constellation. Meanwhile, a Falcon 9 rideshare slip in mid-March impacted Planet and Capella Space manifests, temporarily creating EO coverage gaps — a reminder that when one company controls the dominant launch cadence, its schedule hiccups ripple through the entire data economy.

EUMETSAT signed a €248 million contract with OHB Sweden to build EPS-Sterna, a constellation of 20 microsatellites carrying microwave instruments to fill critical observation gaps over the Arctic. The first six satellites are slated to launch in 2029, with the full system operating over a 13-year mission life.

Why the Arctic matters to you even if you live nowhere near it: polar data is the single most important input for modern numerical weather prediction across all of Europe. EUMETSAT projects the constellation will generate more than €30 billion in economic value by improving forecasts that protect lives, infrastructure, and economic activity. This is a textbook example of the mature space economy at work — identifying a specific, high-value data gap on Earth and building the precise orbital tool to fill it. The design is based on a demonstrator satellite OHB built in a record three years, proving that Europe's "New Space" approach can deliver operational systems, not just prototypes.

Earth observation customers have always faced a painful tradeoff: high resolution of a few places, or low resolution everywhere. Satellogic's new Merlin constellation aims to eliminate it — daily remapping of the entire planet at one-meter resolution, with the first satellite launching in October 2026 and full operational capability targeted for the first half of 2027.

If it works, the shift is fundamental: from tasking a satellite to look at a specific port after something happens, to having a persistent, always-on view of everywhere. That enables a move from reactive snapshots to continuous global intelligence, with implications for logistics, insurance, commodities trading, and defense. Satellogic is betting the market is ready to subscribe to a planetary feed rather than buy individual images. Backing that ambition with near-term revenue, the company also closed a $30 million contract for crop-monitoring imagery across 10 million hectares in Brazil — the kind of targeted commercial deal that subsidizes broader constellation buildout.

Starcloud filed with the FCC for permission to operate up to 88,000 satellites aimed at hosting AI compute workloads in orbit. The company has one experimental satellite flying — carrying an Nvidia H100 accelerator — so the concept is technically plausible at small scale. But the filing's real significance is what it reveals about the escalating race to claim spectrum and orbital slots for a new class of space infrastructure: orbital data centers.

SpaceX has filed separately for up to one million orbital compute satellites, making 88,000 look modest. China's Guowang filed with the ITU to add 648 satellites to its planned LEO broadband network — a move as much about reserving spectrum priority as near-term hardware. These filings force hard questions about spectrum management, thermal and debris risks from compute-heavy satellites, and which markets actually benefit from moving processing off Earth. Expect ITU coordination discussions to intensify as domestic filings convert into global claims.

Canada announced $200 million for a commercial spaceport near Canso, Nova Scotia, plus $105 million in grants to three companies — NordSpace, Canada Rocket Company, and Reaction Dynamics — to develop homegrown small-satellite launch vehicles. The goal: a Canadian rocket launching a Canadian satellite from Canadian soil by 2028. Separately, Canada's Defence Investment Agency awarded $32 million to MDA Space for three ground-based optical observatories to track objects in deep space.

Taken together, it's the full stack of space sovereignty: the ability to see what's in orbit and the ability to put your own assets there without depending on a foreign partner. It's a pragmatic response to a world where access to space and space-based intelligence has shifted from luxury to necessity — and it mirrors similar sovereign-infrastructure moves by Ukraine, Taiwan, and the EU this same week.

A 5.75-million-pound rocket crawling to the pad at one mile per hour; 88,000 imaginary data centers claiming spectrum that doesn't exist yet; and a half-humanoid robot named Vyommitra waiting patiently in a capsule for India to press the button. The future of space infrastructure is apparently being built by the world's most patient entities — bureaucracies, crawlers, and robots that don't need to eat. Keep looking up, but maybe also keep looking at the filings.

If someone you know tracks orbits, trades commodities, or just wonders why their GPS keeps lying to them — forward this to them.

The FTC just declared "unfair" AI an enforcement target under Section 5 of the Federal Trade Commission Act — which matters if you're selling AI-derived EO analytics or autonomous mission outputs to downstream customers. Read → FTC Draws a Line: "Unfair" AI Is Now an Enforcement Target