Starship

The V3 stack now flying stands 124 metres, which makes it the tallest launch vehicle ever flown, nine metres wide the whole way up. Thirty-three Raptor 3 engines on the Super Heavy booster produce about 80,800 kilonewtons, roughly three and a half times a Falcon Heavy. Earlier versions were shorter: 121.3 metres for V1, 123.1 for V2.

The 100 tonne payload figure is a design target, not a demonstrated capability, and the reason to be careful about it is that every one of the thirteen flights so far has been deliberately suborbital. Flight 13 in July 2026 deployed twenty real Starlink V3 satellites and they burned up about twenty minutes later, because the stack was never travelling fast enough for them to stay up. That is a deployment test, not an operational mission, and reporting it as one would be wrong.

The recovery record has gone backwards, which is not the story the cadence graphs suggest. Three tower catches happened on flights 5, 7 and 8, all in 2024 and early 2025 with the older booster. Neither V3 flight has caught anything: both ended in a hard splashdown after a boostback burn that cut short. The ship has never been caught or landed at all, and every ship recovery to date is a water splashdown in the Indian Ocean.

There was a seven month gap between flight 11 in October 2025 and flight 12 in May 2026 while V3 and a second Starbase pad came online. Three and a half years in, the programme has produced eight successes and five failures, no orbital flight, and no in-space propellant transfer, which is the demonstration the lunar lander architecture actually depends on.

SpaceX has published a price for what this is all supposed to buy: Mars cargo at 100 million dollars per tonne, no earlier than 2028. That is 100,000 dollars a kilogram, against 7,000 for a Falcon 9 rideshare to sun-synchronous orbit. It is worth quoting precisely because it is the first time anyone has put a number on interplanetary freight, and the number is enormous.