Previous analyses likely anchored on mass drivers(costly, heavy) rather than sling launchers or ignored the possibility of launching and storing projectiles in parking orbits for later strikes. Both of these changes make lunar bombardment much more practical.
centrifuge/sling based launch system rather than a mass driver
non-suspicious materials (composites, electric motors)
mass driver weighs more than the solar panels to power it (capacitors, coils, etc.), centrifuge parts weigh much less
store projectiles in parking orbits building up "strike all at once" arsenal.
Stored projectiles can strike even if lunar launcher is inoperable/destroyed.
Stored projectiles have approximately weekly strike window, can strike any longitude, latitude is less flexible.
Strike decision/initiation occurs days before impact.
Northern/southern hemisphere chosen at launch.
This is a representative good tradeoff, other options might be possible but require more projectile delta-v.
Otherwise, conventional analysis applies. Others have numbers for tonnes launched /MW of solar, proposals for deploying near the poles to get nearly continuous power, proposals for generating oxygen (reaction mass).
Data centers at the lunar poles are feasible but uneconomical. Building the infrastructure is worth it if there's a roadmap to turning it into nation state nuclear power level destructive capability.
Okay, so I saw ‘Elon Musk wants to build a mass driver on the Moon’ in another context earlier, and my first thought was to ask Claude ‘what would be the military impact of Elon Musk having a mass driver on the Moon’ because we all know who first came up with putting a mass driver on the moon (good news is that Claude said it probably wouldn’t accomplish anything because of physics), but it’s maybe the kind of thing I didn’t quite expect to have him point out first.
I disagree about feasibility/threat.
This has been sitting in my drafts ever since Zvi's reaction to the podcast but doing all the analysis to validate parking orbit dynamics was awful.
Plausible scenario
A company builds a moon base including a data center and a lot of solar at the lunar south pole around Shackleton crater, this is widely understood to be uneconomical but they do it anyways. In situ resource utilisation (ISRU) equipment is built to gather lunar materials, nickel iron from regolith, water from cold polar craters.
Less than 100 tonnes of materials shipped from earth are used to build sling based launchers. ISRU nickel iron, oxygen and some additional shipped parts (cold gas thrusters, electronics) are used to build 250kg projectiles. Each one is a 250kg oxygen canister with 100m/s delta-v budget. They are launched into a parking orbit. Near zero launch signature. Each of these projectiles can deliver 3 tonnes TNT equivalent energy to a target on earth.
10MW of solar power is sufficient to launch 30,000 impactors per month.
This is peak WW2 levels of sustained bombardment except each of these 3 tonne equivalent bombs is a precision guided munition.
There's some additional supporting equipment, MEO/GEO/L1 satellites for communication and trajectory guidance. Once the projectiles are parked, they're a primed weapon waiting for a launch command.
Rough numbers
1MW of electrical power feeding a launch system at 90% efficiency gives
24.5 T/day (launched mass)
360 T/day (TNT equivalent to earth)
Peak WW2 allied bombing dropped 3000T/day of explosives (bombs were ~50% explosives by weight)
10MWe gets you this, sustained, so long as supplies hold out
10MWe of solar might cost <$100B for 1000 tonnes(very pessimistic) at $100M/Tonne. Likely much less than 1000T though. A few hundred tonnes for solar is plausible.
Spinlaunch
Spins a payload at the end of a long arm, then lets it go.
Moon is ideal
no vacuum chamber required
2.5km/s escape velocity requires no exotic materials (carbon fiber arm)
8000km/h = 2.2km/s (proposed launch speed)
+30% material strength to reach 2.5km/s and lunar escape velocity
Can make the arm much longer
reduce G-forces/release mechanism forces
improve directional accuracy
compare a mass driver which
is much bigger and more expensive
requires storing launch energy for a payload in large capacitors
Parking orbits for projectile storage
If launched projectiles are sent directly to earth, each projectile hits 2-4 days after launch.
I've run the trajectory calcs for a specific family of parking orbits. They have a 1 week period with a large strike initiation window. The strike command triggers a burn that gives a similar ~3 day travel time. There's some subtlety regarding trajectories, re-entry angles (>30° below the horizon is feasible) and timing but the whole thing is quite practical.
Materials required
carbon fiber and some very big bearings for spin-launch. Big mass imbalance when payload is released.
Per impactor hardware
impactor body
sintered nickel iron
magnetically separated from regolith
acts as a pressure vessel for O2 cold gas propellant
cold gas propellant (~15% projectile mass, 100m/s delta V)
Electronics, communications
satellite phone equivalent
gives position/velocity tracking as well
Star tracker (image sensor + lens)
Smartphone grade IMU
Cold gas thrusters (Main high thrust + attitude control thrusters)
Previous analyses likely anchored on mass drivers(costly, heavy) rather than sling launchers or ignored the possibility of launching and storing projectiles in parking orbits for later strikes. Both of these changes make lunar bombardment much more practical.
Otherwise, conventional analysis applies. Others have numbers for tonnes launched /MW of solar, proposals for deploying near the poles to get nearly continuous power, proposals for generating oxygen (reaction mass).
Data centers at the lunar poles are feasible but uneconomical. Building the infrastructure is worth it if there's a roadmap to turning it into nation state nuclear power level destructive capability.
Why this post?
To quote Zvi (On Dwarkesh Patel’s 2026 Podcast With Elon Musk and Other Recent Elon Musk Things)
I disagree about feasibility/threat.
This has been sitting in my drafts ever since Zvi's reaction to the podcast but doing all the analysis to validate parking orbit dynamics was awful.
Plausible scenario
A company builds a moon base including a data center and a lot of solar at the lunar south pole around Shackleton crater, this is widely understood to be uneconomical but they do it anyways. In situ resource utilisation (ISRU) equipment is built to gather lunar materials, nickel iron from regolith, water from cold polar craters.
Less than 100 tonnes of materials shipped from earth are used to build sling based launchers. ISRU nickel iron, oxygen and some additional shipped parts (cold gas thrusters, electronics) are used to build 250kg projectiles. Each one is a 250kg oxygen canister with 100m/s delta-v budget. They are launched into a parking orbit. Near zero launch signature. Each of these projectiles can deliver 3 tonnes TNT equivalent energy to a target on earth.
10MW of solar power is sufficient to launch 30,000 impactors per month.
This is peak WW2 levels of sustained bombardment except each of these 3 tonne equivalent bombs is a precision guided munition.
There's some additional supporting equipment, MEO/GEO/L1 satellites for communication and trajectory guidance. Once the projectiles are parked, they're a primed weapon waiting for a launch command.
Rough numbers
Spinlaunch
Parking orbits for projectile storage
If launched projectiles are sent directly to earth, each projectile hits 2-4 days after launch.
I've run the trajectory calcs for a specific family of parking orbits. They have a 1 week period with a large strike initiation window. The strike command triggers a burn that gives a similar ~3 day travel time. There's some subtlety regarding trajectories, re-entry angles (>30° below the horizon is feasible) and timing but the whole thing is quite practical.
Materials required