SpaceX launched the first batch of its newest Starlink internet satellites on July 24, 2026. The top part of the rocket, called Starship, did its job: it released 20 satellites into space, survived the fiery trip back through the atmosphere, and landed in the Indian Ocean about an hour after liftoff from SpaceX's launch site in Texas (SpaceX).

But the lower part of the rocket — the booster called Super Heavy, which lifts Starship off the ground — failed for the second time on this newer version of the vehicle. The booster did not fire all the engines it needed to slow itself down for a landing in the Gulf of Mexico. It hit the water faster than expected and exploded (TechCrunch).

This is the second Super Heavy failure on this version, known as V3. During the first V3 flight in May 2026, the booster failed during the moment when it separates from the upper stage. That May flight also had the upper stage lose an engine (TechCrunch). That first V3 flight was the twelfth Starship test flight overall, on May 22, 2026 (SpaceX).

The July 24 launch came after a cancelled attempt a little over a week earlier. That earlier try aborted right after the engines ignited because several engines failed. SpaceX replaced six engines before the successful flight (TechCrunch).

The upper stage had a much better day than the booster. Starship released all 20 satellites, and SpaceX stayed in communication with every one of them while they were in space. However, because Starship cannot yet travel fast enough to circle the Earth in orbit, the satellites burned up in the atmosphere about 20 minutes after they were released. SpaceX labeled the flight as an "Expended" mission on its launch schedule, meaning the payloads were never meant to stay in space (SpaceX).

One thing was different from earlier Starship missions. When the upper stage tipped over into the water, it did not explode. It floated. That let SpaceX use a drone to inspect the heat shield tiles that protect the spacecraft during reentry (TechCrunch).

SpaceX has said that launching 60 of these new V3 satellites on Starship could bring a "potential twenty-fold increase" in how much data the satellites can beam down to Earth compared to one launch on SpaceX's older, smaller rocket, the Falcon 9 (TechCrunch). SpaceX's filing to become a public company also noted that without a fully reusable Starship, progress on Starlink "would be at a slower pace and higher cost" (TechCrunch).

This was the first Starship launch since SpaceX went public in June 2026, in the largest stock market debut in history. The company's stock had peaked above $200 per share. On July 24, it closed at $115 and then dropped another 2% after hours when news of the booster failure spread (TechCrunch).

The stock drop is worth a closer look. A 42% fall from the peak, plus additional after-hours losses, shows investors are rethinking how risky the Starship program is. The upper stage performed well — surviving reentry and floating intact so SpaceX could study it afterward. But the booster has now failed twice in a row on this version. Two failures back-to-back raise the question of whether the booster's design, or the way its engines start up for landing, needs deeper engineering work than the usual test-and-fix approach has delivered.

The satellite deployment was a partial success. Staying in contact with all 20 satellites confirms their systems work in space, even though they were always going to burn up on this flight. It is essentially a test run for the satellites themselves — checking that their power, computing, and communications systems work — done before the rocket can actually place them in a lasting orbit. SpaceX has done this kind of short test before across its programs, and even 20 minutes of data in space has real value for improving the satellite design.

The bigger picture is about reusability. SpaceX's own public filing says it plainly: a fully reusable Starship is how the company plans to launch Starlink satellites more often and at lower cost. The upper stage floating intact is a good sign for the heat shield. But the booster still cannot land itself, and until it can, the full system that makes Starlink cheaper to build out is not yet working.

Where things stand: SpaceX now has satellites proven to work in space, an upper stage that can survive the trip back to Earth, and a booster that has failed twice. The company's approach has always been to learn from failure and fix things for the next flight. But the gap between how well the upper stage is performing and how poorly the booster is performing has grown over two V3 flights. Closing that gap will likely decide how soon Starship can stop being a test vehicle and start launching Starlink satellites for real.