Why Starship Cannot Go Anywhere Without Refueling in Orbit?

Added: 2026-05-08

[00:00:05]SpaceX has launched over 400 rockets.

[00:00:08]They've landed boosters on drone ships and caught a super heavy with mechanical arms, but they have never refueled a rocket >> [music] >> in orbit.

[00:00:18]And until they figure it out, Starship is stuck. It has enough fuel to reach low Earth [music] orbit, a few hundred kilometers up, but not enough to go anywhere else. Not the moon, not Mars, nowhere. To get Starship to the lunar surface, its tanks need 1,200 tons of cryogenic propellant, roughly the weight of 800 cars loaded in orbit in zero gravity at minus 183° C, cold enough to turn the oxygen you breathe into [music] a liquid. The obvious question is how? How do you move 1,200 tons of cryogenic propellant between two vehicles in [music] orbit?

[00:01:00]But that's actually the second question.

[00:01:02]The first one is more fundamental. Why does liquid stop behaving like liquid the moment you remove gravity? To understand, we need to go back to a physics problem that has haunted rocket engineers since the 1960s.

[00:01:23]When NASA [music] was planning the Apollo missions in the 1960s, engineers ran into a constraint that had been sitting in the physics textbooks since 1903, the Tsiolkovsky rocket equation. And 60 years later, it's still the single biggest obstacle standing between Starship and the moon.

[00:01:44]The equation says that to accelerate in space, you need propellant. And the more propellant you carry, the heavier the vehicle gets. So, you need even more propellant to push that extra weight.

[00:01:55]It's a loop that punishes scale. And here's what SpaceX built to take on that equation. It has two parts. At the bottom is super heavy, the booster. It packs 33 engines with one single job, get everything off the ground.

[00:02:12]On top sits Starship, the spacecraft.

[00:02:15]That's the part that actually reaches orbit, flies to the moon, or one day makes it to Mars.

[00:02:22]Together, they form the tallest, most powerful rocket ever built. For the full stack, the numbers look like this. About 5,000 tons on the launchpad. 92% of that, roughly 4,600 tons, is propellant.

[00:02:38]To put that in perspective, imagine driving across the country in a car where 92% of the vehicle's weight is gasoline. By the time you arrive, the tank is empty. The super heavy booster burns through about 3,400 tons of that just getting off the pad, then separates and flies back. That leaves Starship alone with roughly 1,200 tons of its own propellant. And by the time it reaches that low Earth orbit, or LEO, where the International Space Station flies, nearly all of that is gone. Without refueling, the only thing Starship can do at that point is turn around and come home. But here's the catch. Starship is big, and big means heavy. The stainless steel hull, six Raptor engines, internal tank walls, the payload bay, all of that structure adds up to somewhere between 100 and 165 tons before a single drop of propellant goes in. Every one of those components is essential. The engines provide thrust, the hull protects the cargo, the tank walls hold the fuel.

[00:03:49]But the rocket equation doesn't care. It sees all of it as mass that has to be accelerated. And every kilogram of structure is a kilogram that can't be payload.

[00:04:00]So, the heavier the vehicle, the more fuel it needs just to move itself, and the less it can deliver to the moon.

[00:04:08]To go from LEO to the lunar surface, Starship needs a lot more acceleration, roughly 5.5 to 6 km/s of delta V.

[00:04:18]But the tanks are empty, and the vehicle itself weighs over 100 tons. So, it can deliver exactly zero tons of payload to the moon.

[00:04:27]Zero.

[00:04:29]But if Starship gets refueled in orbit, its tanks filled back up before leaving LEO, that number jumps to over 100 tons.

[00:04:38]That's not an incremental improvement.

[00:04:40]That's the difference between a mission that doesn't exist and one that carries more cargo to the moon than the entire Apollo program combined.

[00:04:49]So, why not just build a bigger rocket and skip the refueling?

[00:04:53]Because the rocket equation scales exponentially.

[00:04:56]Doubling the delta V doesn't just double the fuel. It multiplies it many times over.

[00:05:02]A rocket large enough to fly directly to the moon without refueling would need to be five to 10 times the size of Starship.

[00:05:10]The cost, the infrastructure, the engineering risk, all of it becomes unmanageable.

[00:05:16]The smarter path is to use a reusable vehicle to reach LEO, which is already halfway to anywhere in terms of energy.

[00:05:24]Refuel it there and then fly on.

[00:05:27]So, the answer sounds straightforward.

[00:05:29]Launch a second Starship, a tanker, loaded with propellant. Dock the two vehicles together in orbit, transfer the fuel over. Simple, right? It's not.

[00:05:40]Because the moment you try to move cryogenic liquid in zero gravity, you run into a physics problem that no engineer dares ignore.

[00:05:55]But before any fuel can move, two of the largest spacecraft ever built need to [music] find each other in orbit and lock together.

[00:06:05]That alone is an engineering challenge on a scale nobody has ever attempted.

[00:06:10]The two Starships dock nose to nose, and there's a specific reason for that. The engines sit at the tail. That's where most of the mass is concentrated. By connecting at the nose, the lighter end, the combined center of mass stays manageable, and the vehicle pair is far more stable during docking. It also puts the propellant transfer lines right at the connection point, so fuel flows through the shortest possible path.

[00:06:37]SpaceX doesn't start from scratch here.

[00:06:39]Their Dragon capsule has docked with the International Space Station over 100 times fully autonomously. The system uses lidar to measure distance down to the centimeter, optical star trackers to lock onto fixed reference points, [music] and cold gas nitrogen thrusters for precise corrections. That heritage transfers directly to Starship. Inside each nose cone is the docking mechanism itself, a probe and drogue system. One Starship extends a probe, the other receives it in a cone-shaped drogue.

[00:07:14]Once they lock, a ring of quick-disconnect couplers seals the propellant transfer lines. These couplers have to hold at minus 183° C, cold enough that most metals become brittle. And they have to seal tight against vacuum on one side and pressurized cryogenic liquid on the other. When the two lock together, the combined structure weighs somewhere between 300 and 400 tons, more than 15 times the mass of the first module ever attached to the International Space Station. Nothing this large has ever been docked in orbit before. And that's worth being honest about. Dragon has proven the software works, the lidar works, the autonomous approach works.

[00:07:59]But Dragon weighs about 12 tons.

[00:08:01]Starship weighs over 100. Docking a Dragon is like parallel parking a sedan.

[00:08:07]Docking two Starships is like pushing two freight trains together in the dark with no brakes. This is uncharted territory. If this is the kind of breakdown you come here for, hit subscribe, The Cosmic Rush. Now, back to the problem. But suppose they solve it.

[00:08:24]Suppose the two Starships dock perfectly, the seals hold, the structure is stable. At that point, a completely different problem shows up, one that has nothing to do with docking or software or structural loads. This one is hiding inside the fuel tanks, and it starts the moment anyone tries to move cryogenic liquid in an environment with absolutely no gravity.

[00:08:52]>> [music] >> On Earth, refueling takes about 3 minutes. Gravity pulls the liquid down, the pump pushes it through, done. Every gas station on the planet works because gravity does the hard part for free. But take gravity away, and everything that felt obvious suddenly stops working. In zero gravity, there is no up and no down. The liquid inside the tank doesn't settle at the bottom because there is no bottom. Instead, it drifts. It forms floating blobs. It clings to the tank walls, wraps around internal structures, and spreads itself across every surface it can find. Surface tension, a force so weak it's barely noticeable on Earth, takes over completely. Think of it this way. Try pouring water out of a bottle, but the water doesn't fall. It pulls itself into a floating ball inside the bottle and just stays there. Now, imagine that same problem with 1,200 tons of liquid at minus 183° C, colder than anywhere on Earth, including Antarctica. That's the problem SpaceX engineers have to solve, and here's the critical part. At the base of each tank is a sump where the feed line draws propellant out. On Earth, gravity keeps that sump flooded with liquid. In zero gravity, it could be surrounded entirely by gas. And when the feed line opens, it doesn't pull clean liquid, it pulls a mix of liquid and gas. Even a small amount of gas in that flow is enough to shut the whole system down. The obvious fix is a mechanical pump. Just force the liquid through the line. But, here's the problem. Cryogenic propellant is already extremely close to boiling. It doesn't take much to push it over the edge, and pumps work by creating a low-pressure zone at the inlet to pull fluid in. That tiny pressure drop is enough. The liquid boils on the spot. Vapor bubbles form instantly at the pump inlet, a phenomenon called cavitation. On Earth, those bubbles rise and collapse harmlessly. In zero gravity, they have nowhere to go. They pile up, block the flow, and the pump chokes. It spins, but moves nothing.

[00:11:11]There's a second problem hiding in the same tank. Mixed in with the liquid is a pocket of gas. Engineers call it ullage.

[00:11:19]That gas is warmer than the liquid around it, and in zero gravity, the propellant sloshes constantly. There's nothing holding it still.

[00:11:28]When the cold liquid splashes into the warm gas, the gas condenses back into liquid almost instantly.

[00:11:35]And when gas disappears, the pressure inside the tank drops suddenly, without warning. [music] Engineers call this ullage collapse. One second, the tank pressure reads normal.

[00:11:47]The next second, it craters.

[00:11:49]And without stable pressure, there's no force to push propellant through the transfer lines. The system stops. And the worst-case scenario ties both problems together. If vapor bubbles from cavitation travel through the transfer lines and reach the engine, the Raptor doesn't just lose power, it can tear itself apart. The engine is designed for a steady flow of dense, cold liquid.

[00:12:13]Replace that with gas, even partially, and the combustion becomes chaotic.

[00:12:17]That's not a malfunction, that's a catastrophe. So, the liquid won't settle. Mechanical pumps cause cavitation, and the pressure inside the tank can collapse without warning. On paper, this should be impossible to solve, but SpaceX believes they have an answer, and it's elegant in a way that might be surprising.

[00:12:45]The first problem was that the liquid wouldn't settle.

[00:12:49]Without gravity, there's no force pulling it to the bottom of the tank.

[00:12:53]SpaceX's solution is to create their own gravity. Just barely.

[00:12:58]Small thrusters on the vehicle fire during the transfer, giving the entire structure a gentle, sustained push.

[00:13:05]The acceleration they produce is tiny, about 0.001 G, roughly 1/1000 of Earth's gravity. To put that in perspective, it's about the force of a single ant pushing a soda can.

[00:13:20]But, in microgravity, that's enough.

[00:13:22]That faint push settles 1,200 tons of cryogenic propellant to the correct end of the tank. Liquid at one end, gas at the other. Engineers call this ullage acceleration. It's the same principle NASA used on the Saturn V in the 1960s. [music] The physics hasn't changed, the scale has. The second problem was cavitation.

[00:13:46]Pumps destroying themselves by boiling the liquid they're trying to move.

[00:13:50]SpaceX's answer is radical. Don't use a pump at all. Instead, the tanker's tank is pressurized higher than the receiver's tank.

[00:13:59]The propellant flows from high pressure to low pressure, the same way air rushes out of a balloon.

[00:14:05]No mechanical pump, no moving parts, no cavitation risk. Just a pressure difference doing all the work. The transfer line itself is about 20 to 50 cm in diameter, roughly the width of a dinner plate, designed to move over 1,000 tons in a few hours. And the third problem, ullage collapse, where pressure in the tank drops without warning, gets solved in what might be the most clever way of all. Remember, cryogenic propellant naturally boils off over time. A tiny fraction is constantly evaporating.

[00:14:40]Most engineers treat boil off as a problem, wasted fuel. SpaceX treats it as a resource. That evaporated gas is fed back into the top of the tank, where it builds up pressure.

[00:14:53]The warmer the gas, the higher the pressure. So, the system is self-regulating.

[00:14:58]Boil off creates pressure, pressure pushes liquid, liquid flows through the transfer line. No external gas tanks needed, no helium, [music] no additional weight. The problem becomes the solution.

[00:15:12]Engineers call this autogenous pressurization. The transfer lines themselves are engineered to survive the environment. Each pipe is vacuum jacketed. An inner line carrying cryogenic fluid surrounded by a vacuum gap that blocks heat transfer, wrapped in layers of reflective insulation called MLI, the same principle as a thermos bottle, but built to operate in space at minus 183° C carrying pressurized liquid between two vehicles moving at 28,000 km/h.

[00:15:45]So, on paper, the transfer system is brilliantly simple. Settle with thrust, push with pressure, use boil off as the pressure in. Three problems, three solutions, zero pumps. But, there's one more enemy that never sleeps. One that's been attacking since the moment the propellant was loaded. That enemy is heat.

[00:16:12]Liquid oxygen boils at minus 183° C. Liquid Liquid methane boils at minus 161° C. In In orbit, the sun hits one side of the vehicle at full intensity, while the other side faces the void of space.

[00:16:30]Heat doesn't need air to travel, it radiates straight through. And no matter how good the insulation is, some of that heat always gets in.

[00:16:39]The cryogenic propellant absorbs it and slowly boils away. Engineers call this boil off.

[00:16:46]The target is to lose less than 0.1% per day.

[00:16:51]That sounds like almost nothing, but run the numbers. 0.1% of 1,200 tons is 1.2 tons lost every day.

[00:17:02]If the full refueling campaign takes 16 tanker flights spread over roughly 48 days, the total loss comes to about 58 tons.

[00:17:11]That's the weight of an M1 Abrams battle tank just evaporating into space. To fight that, every surface of the tank is wrapped in multi-layer insulation, or MLI.

[00:17:24]Dozens of thin, reflective foil sheets stacked together with vacuum gaps between them, like stacking 50 mirrors with nothing but empty space in between.

[00:17:34]Each layer reflects radiant heat back.

[00:17:36]The vehicle can also rotate slowly around its long axis during coast, a technique called passive thermal control roll, spreading solar heating evenly instead of baking one side. The more heat they block, the more propellant they keep. And that brings up the logistics. One moon mission doesn't require one launch, it requires 10 to 16 successful launches back to back. Each tanker flight has seven critical steps: launch, reach orbit, rendezvous, dock, transfer, undock, deorbit. That's seven steps repeated 16 times. 112 steps that all have to go right in sequence.

[00:18:20]Think of it like flipping a coin 16 times in a row and needing heads every single time. One tails, and everything goes up in smoke. That is an unprecedented operational chain. No space program in history has attempted anything close to this launch cadence for a single mission.

[00:18:38]The skeptics who call this unrealistic aren't wrong to ask hard questions. The engineering is real, the physics works, but the execution at this scale has never been done. So, the technical challenges are real. The skeptics have valid points. But, here's where the story takes a turn, because the economics of Starship tell a completely different story.

[00:19:04]>> [music] >> Here's the number that changes the conversation. NASA's SLS rocket costs roughly $2 billion per launch, and it's expendable. It flies once, and it's gone. SpaceX is targeting $10 million per Starship launch. Even if a single moon mission requires 16 tanker flights, the total cost comes to about $160 million.

[00:19:31]The rocket that needs 16 launches is still 12 times cheaper than the rocket that needs one. And SpaceX isn't the only one working on this. ESA is developing the Odyssey orbital depot, targeting 2028. China's Tianzhou cargo ships have already transferred storable propellant to Tiangong. OrbitFab is building commercial refueling ports designed to become the first orbital gas stations. The race for orbital refueling is global, but SpaceX is the only one attempting it at 1,200 tons scale. And this is where the math becomes a flywheel.

[00:20:09]Reusability drives the cost down. Lower cost allows more launches. More launches generate more flight data.

[00:20:18]More data improves reliability. And better reliability makes refueling more predictable. That's not just a cost advantage. That's a compounding loop.

[00:20:27]And once it starts spinning, it pulls away from everyone.

[00:20:36]But the skeptics aren't wrong. The Government Accountability Office has flagged orbital refueling as a top risk for the Artemis program.

[00:20:45]The technology, they say, isn't mature.

[00:20:48]Timelines have slipped. And that's a fair assessment. But data is catching up.

[00:20:54]On the IFT-3 mission in March 2024, SpaceX successfully transferred cryogenic propellant between internal tanks on a single Starship while in orbit. The first step toward proving the concept works in space. A full ship-to-ship transfer demo between two Starships is currently targeting mid-2026.

[00:21:15]Theory is becoming reality. And there's one counterargument that deserves complete honesty. No one has ever transferred 1,200 tons of cryogenic propellant in zero gravity. That is true.

[00:21:30]That is the biggest unknown in the entire mission architecture.

[00:21:34]But it's worth remembering that no one had ever caught a 230-ft rocket booster with a pair of mechanical arms, either.

[00:21:42]Until October 2024. [music] So, back to where we started. 1,200 tons of propellant at minus 183° C.

[00:21:59]It won't [music] settle. It boils the moment a pump touches it. And the pressure holding it in place can collapse without warning. The rocket equation says Starship can't reach the moon without that fuel. The physics says moving it in zero gravity should be nearly impossible.

[00:22:17]But at $160 million for 16 flights versus $2 billion for one, the economics say it doesn't matter how hard it is. It has to work.

[00:22:29]And so, they're building it.

[00:22:32]What SpaceX is building goes far beyond a single refueling test. [music] It's an orbital supply chain. 250-m vehicles docking autonomously, settling propellant with milli-gravity thrust, pushing it across with nothing but pressure, and [music] keeping it cold in an environment that never stops trying to boil it away.

[00:22:52]The architecture doesn't depend on a single breakthrough. [music] It's built on known physics, proven hardware, and a cost structure that rewards repetition.

[00:23:01]Whether it takes one attempt or 10, the system is designed to get better every time it flies.

[00:23:07]And that race is already underway.

[00:23:10]Here's the question worth debating.

[00:23:12]If SpaceX needs 16 perfect launches in a row for a single moon mission, what's the acceptable failure rate? Drop your number in the comments.

[00:23:22]If this kind of engineering deep dive is what you come here for, subscribe to the Cosmic Rush. This is where we go deep.