And after 334 flights, it happened: the first Falcon 9 failure since 2016.

It seemed like it would never happen, but eight years later, the SpaceX Falcon 9 failed. On July 12, 2024, at 02:35 UTC, the Starlink Group 9.3 mission lifted off from SLC-4E at Vandenberg Air Force Base, California. The mission was to place twenty Starlink v2 Mini satellites into orbit, thirteen of which were D2C (direct cell access) for direct communication with mobile terminals. The launch could not have been more routine. It was the 70th Falcon 9 launch of 2024 and the 354th overall – numbers that make your head spin. Indeed, the liftoff was uneventful, and the first stage B1063 performed its function and landed on the OCISLY barge without problems, completing its 19th flight (it was the 329th successful landing of the first stage and the 255th overall).

But during the second stage’s ascent to orbit and after fairing separation, frost and a leak of liquid, later confirmed to be liquid oxygen, began to appear in the traditional chambers that show us the Merlin 1DVac engine. Frost and liquid oxygen ice formation on the second stage is a common occurrence on Falcon 9 launches and usually attracts attention, but in this case the volume of ice was much greater. In addition, the liquid leak and excessive inflation of the insulating material layers (MLI) during ignition were clearly abnormal. Despite everything, the second stage’s launch was completed, placing a payload of twenty satellites into an initial orbit of approximately 135 by 280 kilometers with an inclination of 53.2 degrees.

Starlink missions use two second-stage burns: the first to leave the payload in an orbit of about 150 x 300 kilometers, and a second, very short-duration burn to raise the perigee until the orbit is circularized to about 300 kilometers. (Originally, a single burn was used, but it was changed to two due to the larger size of the v2 Mini and the greater atmospheric friction at these altitudes as a result of the expansion of the atmosphere near solar maximum.) From here, the satellites use their own argon-based ion thrusters to circularize the orbit and raise the perigee to 300 kilometers. However, the second burn did not occur. According to Elon Musk in X, the engine “survived a RUD (Rapid Unscheduled Disassembly),” meaning it disintegrated before the second burn. However, the stage survived, was able to deploy the satellites, and could then be passivated to prevent it from exploding in orbit. Obviously, entry into the atmosphere was impossible, and the stage remained in orbit, albeit for a short time.

Without this second ignition, the Starlink satellites were left in orbit with a perigee that was too low (most were 135, but some were 115 kilometers, although one appears to have reached 190 kilometers). At that altitude, atmospheric friction is very high, and while Musk indicated that they had fired their ion engines at maximum power…deformation 9“— these engines produce very little thrust. This has yet to be confirmed, it appears that all the satellites reentered the atmosphere between July 12 and 13 (at least one of the twenty Starlink satellites was observed reentering the atmosphere over Chile on July 12 at around 11 UTC). The loss of twenty Starlink satellites is not particularly relevant in a mega-constellation with more than 6,250 objects in orbit (although the failure is compounded by the re-entry of about eight defective units from Group 9.1, launched on June 19). But it is relevant to the operations of Falcon 9, a launch vehicle that wants to cement its position as the fastest-flying vector in history.

As usual, in the event of a launcher failure resulting in the loss of the payload, the FAU (Federal Aviation Administration) ordered an investigation into the incident, and so Falcon 9 missions are on hold until a decision is made otherwise. On another launcher flying at a slower speed, the FAA’s decision probably won’t have a very significant impact. After all, all indications are that the second stage failure has a simple solution (unless it’s a design flaw), and no one doubts that SpaceX will fix the problem in a matter of weeks. But no matter how simple the solution, there’s a time for research. For a launch vehicle that was on track to break the record of 96 launches in a year, this setback is bad news. And let’s not forget that Falcon 9 is a rocket certified for crewed launches with Crew Dragon, so this shutdown will affect future Polaris Dawn (July 31) and Crew 9 (mid-August) missions, as well as other uncrewed missions related to the ISS, such as the Cygnus NG-21 cargo ship (August 3). Of course, it is also possible that the problem will be delayed and that other key missions could be affected, such as the Europa Clipper probe, which is due to launch in October (reports have emerged in recent days of a possible delay in the mission due to a malfunction of an instrument that may not be able to withstand the huge dose of radiation expected in Jupiter orbit, but it is unclear to what extent this problem will lead to a serious delay in the mission).

That this failure occurred on a “nondescript” Starlink mission and not on a billion-dollar crewed launch or probe (yes, it only takes one second-stage ignition to launch a Crew Dragon, but that’s not the point). Regardless, this is the first Falcon 9 failure since Falcon 9 second stage F9-029 exploded on the SLC-40 ramp on December 1, 2019, destroying Israel’s Amos 6 satellite due to a pressurization system failure. The helium stage is what destroyed the launch vehicle. To find a Falcon 9 failure during a launch, we have to go back to June 28, 2015, when another second stage also failed due to a pressurization system failure. In that case, a second-stage explosion caused Falcon 9 v1.1 to disintegrate before first stage separation, resulting in the death of the Dragon cargo ship CRS-7. Since the Amos 6 disaster, Falcon 9 has flown a record 334 missions without a single failure.

And while it is well known that the Falcon 9 is the only reusable launcher currently in service in the world, the truth is that what is equally or more remarkable is its reliability. When SpaceX announced that it would reuse the Falcon 9 first stage, few doubted that it would be able to do so, but many wondered whether it would be able to do so with the minimum reliability that would allow it to compete in the same market as other launch vehicles such as the Delta IV, Atlas V or Ariane 5. Now we will have to wait to find out the exact reasons for the failure. If it is proven that the incident was the result of pressure to launch more missions and surpass the record of 96 annual missions, the matter could become more serious. If not, then we will surely see the Falcon 9 in the air again in a few weeks. Either way, the Falcon 9 failure occurred while the Starliner spacecraft was still docked to the ISS, and reminds us why NASA wants a second crewed spacecraft capable of accessing the ISS, despite Crew Dragon’s excellent performance so far.