SpaceX finally launched it. Starship Version 3 lifted off last week. It was a historic debut, sort of. The wait had been long. Over seven months. Then whoosh. South Texas pad, new rocket, same ambition.
Did it go perfectly? No. The Super Heavy booster didn’t splash down gently like planned. Engine glitches appeared. But SpaceX called it a success. They tend to do that. And they have reasons.
V3 isn’t just another test. It’s the muscle SpaceX needs for what’s next.
Built Heavy, Built Faster
Starship V3 is taller now. 408 feet. That’s 124.4 meters of steel and fire. It uses Raptor 3 engines. Lighter, sleeker, stronger than before.
The upgrades aren’t cosmetic. The Super Heavy stage pumps fuel faster. All 33 engines fire quicker. The upper stage, called Ship, got bigger tanks and docking ports. Those ports matter. Because rockets need gas. A lot of it.
To go anywhere far—Moon, Mars, beyond—you can’t lift everything at once. Physics hates that. So SpaceX plans to send tankers up first. Then meet the main Ship in orbit. Top off the tanks. Go dark.
Experts say dozens of launches might be needed. Exact numbers are fuzzy. But the goal remains clear: refuel in space, go further.
NASA is Watching
Remember Artemis? The program to put humans back on the Moon. NASA picked SpaceX in 2021 to build the lander. But it’s not a done deal. Not yet.
Artemis 3 is supposed to dock in Earth orbit around mid-2027. Artemis 4 lands astronauts near the South Pole by late 2028? Maybe. But there’s competition. Blue Origin built Blue Moon. It’s in the running too.
NASA has said both landers could fly on Artemis 3? Unlikely, but possible. For Artemis 4 though, only one gets the nod. The clock is ticking. SpaceX needs to move fast.
The Grounded Rocket Problem
Right now? Starship is grounded. The FAA declared the May 22 failure a “mishap.” The booster tried to steer home and missed. Badly. An investigation is ongoing. No more launches until SpaceX explains why it went wrong.
After the clearance? The work really begins.
SpaceX needs to reach orbit. Then stay there. Longer missions mean better data on heat, propellant boil-off, system stress. One Ship launches first. Sits in space. Gathers intel. A second Ship follows. Meets up. Transfers fuel.
“It will start with a Starship… gathering data… including long duration propellant storage,” the company said back in October. That was their plan. Still is, likely.
Musk says they have stock. Roughly 10 more Ships coming this year. Plus half that number of boosters. The production line is churning. He posted it on X. He believes it. Do you?
Life Support and Ladders
Fuel isn’t enough for crew. People need air. Water. Temperature control. SpaceX has flown Crew Dragon since 2020. But Starship is bigger. Harder to manage.
In an update from late 2025, the company claimed progress. “Lunar environmental control and life support… demonstrations complete.” They used a full-size cabin module. Multiple people inside. They injected oxygen and nitrogen. Tested sanitation. Humidity control. Sounds complex. Is it ready? We don’t know yet.
And the height issue? Starship’s cabin sits up top. 171 feet up. Astronauts need a way down to the grey dirt.
Elevators. Airlocks. SpaceX tested the concept in Hawthorne, California, back in 2024. Partnered with Axiom Space for the spacesuit integration. The tech exists in principle. Now comes flight certification.
What Happens Next?
An uncrewed flight to the surface is critical. Both Blue Moon and Starship must ace a landing test before humans board. No timeline announced for Starship’s version. But one must fly before late 2028 if Artemis stays on schedule. Blue Moon sends a robotic prototype this fall? If plans hold. That won’t count for crew certification. Just a step.
The cadence should increase. Test more. Break more. Fix it faster. That’s SpaceX’s model. Iterative. Relentless.
Elon Musk sets high bars. Maybe impossibly high. Does that motivate? Or blind?
Starship flies again soon? Probably. The stockpile sits ready. The FAA might unclench.
Until then, the question lingers: Can a megarocket truly learn in real time without crashing every few steps?
