The Lunar Winter
Why We Stopped Going to the Moon — and What It Reveals About How Change Actually Happens
On April 10, 2026, four astronauts splashed down in the Pacific after the first crewed mission beyond Earth orbit since Apollo 17 in December 1972.
Fifty-four years.
Not because the technology wasn’t there. The Apollo Guidance Computer — the machine that actually flew humans to the Moon — ran at 2.048 MHz with 4KB of erasable RAM. A Nest thermostat has more processing power. The gap between 1972 and 2026 was not a capability gap. It was an activation gap.
The Ratchet That Never Stopped Turning
The technology ratchet didn’t pause when the Apollo program ended. It just stopped being aimed at the Moon.
In 1981, the Space Shuttle flew for the first time, controlled by four redundant IBM AP-101B computers running simultaneously — a fly-by-wire system covering seven distinct flight segments from first-stage ascent to runway landing. The computer problem was solved again, better, and then filed away.
By the mid-2000s, NASA’s White Flight Control Room at Johnson Space Center had replaced the Apollo-era analog consoles entirely. Commercial off-the-shelf digital workstations. The mechanical spacecraft display on the front screen gave way to a digitally generated representation of the flight. Mission control infrastructure, rebuilt.
Then Lockheed Martin’s engineers developed a digital twin for the Orion crew capsule — a full system model that, according to the program’s own analysis, could reduce the time to answer engineering questions by days and required human resources by an order of magnitude over historical approaches.
Mainframe-era guidance computers. Internet-era distributed mission control. Cloud-era digital twins. Each wave advanced the capability stack. None of them triggered a return to the Moon.
The ratchet turned. The door stayed shut.
What Finally Opened It
Four conditions aligned — and they didn’t align gradually. They aligned fast, once each one crossed its threshold.
First: cost.
On December 21, 2015, a Falcon 9 first-stage booster — 156 feet tall — executed three engine burns over the Atlantic and landed vertically at Landing Zone 1, Cape Canaveral. Solid ground. Not a barge, not a concept, not a simulation. The booster that had just put eleven ORBCOMM satellites into orbit was standing upright on a Florida launchpad eleven minutes after liftoff.
This single event broke an assumption the entire launch industry had treated as physics: that rockets were expendable. They weren’t. They were just expensive to recover, until someone cared enough to solve it. By February 2026, a single Falcon 9 booster had flown thirty-three times.
Artemis II itself did not benefit from this directly — it launched on SLS, NASA’s own expendable heavy-lift rocket, which NASA’s auditors price at $4.1 billion per flight. But SpaceX’s demonstration changed the broader context: it proved that access to space could be fundamentally cheaper, it attracted commercial and international investment into the sector, and it changed what Congress and NASA leadership believed was worth funding. The $309 billion (in 2025 dollars) spent on Apollo stopped the program cold because the cost was indefensible once the political urgency evaporated. The commercial launch revolution made the case that a permanent return — at lower cost, over time — was no longer a fantasy. SLS is the bridge. What comes after it looks different because SpaceX exists.
Second: competitive pressure.
China’s lunar program made the cost of inaction geopolitical rather than merely budgetary. This isn’t a subtle dynamic — it’s the same one that created Apollo in the first place. Sputnik didn’t make Americans want to go to the Moon. It made them afraid of what it meant if they didn’t.
Third: the ownership model changed.
Artemis is not a NASA mission. That sentence deserves to sit on its own.
NASA provides the mission architecture and the Space Launch System — the rocket that carried the Artemis II crew on their lunar flyby. ESA owns the European Service Module program. Airbus, in Bremen, builds the ESM as prime contractor: propulsion, power generation, thermal control, and life support, all integrated into a single module that makes the Orion capable of deep space operations. Each ESM contains more than 20,000 parts, approximately 12 kilometers of cable, and 8.6 tonnes of fuel. Lockheed Martin builds the Orion crew capsule itself. SpaceX holds the contract for the Starship Human Landing System — the vehicle that will put boots on the lunar surface for Artemis III.
No single organization owns this. No single organization could.
During the Artemis I uncrewed test flight in November 2022, the ESM generated 20% more power than expected while consuming 25% less than predicted across a 25-day, 2.25-million-kilometer mission. That is what happens when you let specialists do what they’re actually good at, instead of insisting on controlling every system internally.
Fourth: someone made a decision.
In August 2006, NASA Administrator Mike Griffin signed a Space Act Agreement with SpaceX under the Commercial Orbital Transportation Services program. SpaceX had not yet reached orbit. They wouldn’t for another two years, and only on their fourth attempt after three failures. The aerospace establishment considered them a curiosity at best.
Griffin signed anyway.
That decision — made by people in a room, not by a capability assessment or a market analysis — is the hinge point of the entire subsequent story. It created the commercial launch industry that changed what space access could cost. It gave SpaceX the revenue and credibility to develop Falcon 9. And it established the model — NASA as customer, private sector as provider — that Artemis would later extend to its European and industrial partners.
The whole chain runs back to a procurement decision that required someone to trust what wasn’t yet proven.
What the Failure Case Looks Like
The Space Launch System — NASA’s own rocket, the one that actually launches Artemis missions — is the counter-narrative running alongside the success story.
SLS total development costs have exceeded $100 billion. The program burns approximately $12 million per day. A single contract for the Mobile Launcher 2 — originally awarded at $383 million — has ballooned to an estimated $1.8 billion, with delivery pushed to September 2027. Block 1B costs are projected to reach $5.7 billion before its first flight, $700 million over the baseline NASA established just two years prior.
The SLS exists because Congress wanted jobs in specific districts, because NASA wanted sovereign launch capability, and because the organizations involved were structured to own rather than partner. It works — it successfully launched Artemis I and II. But it works at a cost that makes it structurally unsustainable without continuous political support, and it works in permanent tension with the commercial alternatives that the broader Artemis coalition has proven are possible.
The SLS is what the old model looks like when it succeeds. The price is instructive.
The Question for the Room You’re Actually In
The ratchet has been turning inside your organization too.
Cloud infrastructure since 2010. Enterprise data platforms since 2012. Machine learning capabilities since 2015. AI agents since 2022. Each wave has advanced what was possible. The question is what you’ve actually activated — and what’s still sitting in the portfolio, capable on paper, dormant in practice.
This is the organizational lunar winter. Not the failure to adopt a technology that’s three years old. The failure to activate a capability that’s been accumulating for a decade or more.
Look at your technology portfolio from the last ten years. The data warehouse that was built to enable analytics and still runs a handful of reports. The automation platform that went through a successful pilot and never scaled. The ML model that the data science team built and the business never operationalized. These are not implementation failures. They are activation failures — and the four conditions that ended the space program’s winter are the same ones worth examining here.
Is it cost? The economics of activation may not yet have crossed the threshold where the return justifies the disruption. That threshold moves — sometimes quickly, once a competitor crosses it first.
Is it competitive pressure? Sometimes organizations need external shock before internal will crystallizes. The capability existed. The urgency didn’t. China’s lunar program didn’t create NASA’s capability to return to the Moon. It created the will.
Is it the ownership model? The most underdiagnosed activation blocker is the assumption that the organization running the initiative needs to own all of it. Artemis didn’t own the service module, the crew capsule, or the lunar lander. It owned the mission. That’s a different thing. Many capability programs stall because the team trying to activate them insists on building what they could assemble — and collapses under the weight of being excellent at too many things simultaneously.
Is it the decision in the room? The 2006 NASA-SpaceX agreement didn’t follow from a risk assessment that concluded SpaceX was a safe bet. It followed because someone decided that the cost of inaction — remaining dependent on the existing model — was higher than the cost of betting on something unproven. That kind of decision doesn’t emerge from analysis. It requires a leader willing to treat waiting for more certainty as a choice, with consequences.
The Artemis II crew didn’t return from the Moon because the technology finally caught up.
The technology caught up in 1995.
They returned because four conditions aligned — and because, at the critical moment in 2006, someone in a room made a bet that wasn’t safe.
Who in your organization has the authority — and the appetite — to make that bet?