How Do SpaceX's "Chopstick" Arms Catch A Falling Rocket Booster?

Table of Contents (click to expand)

SpaceX does not land its Super Heavy booster, it catches it. After the booster separates from the ship it flips, fires its engines to fly back toward the pad, falls engine-off for about three minutes while waffle-shaped grid fins steer it, then relights thirteen of its thirty-three engines to brake and keeps three running to hover beside the tower, where two steel arms close under small catch points below the fins. The catch removes the need for landing legs, which saves weight and lets the booster be set straight back on the launch mount; by November 2026 the arms had caught three boosters and one upper stage, and automatic checks had sent several other boosters to the sea instead.

Twenty-five seconds. That is how long Super Heavy's landing burn lasts. It starts when the engines relight and ends with the booster resting on two steel arms. In those seconds a stainless-steel tube 72 meters (236 feet) tall stops falling. It slides sideways and parks itself against the tower it left seven minutes earlier.

Watch the clip and your brain files it under special effects. A building is not supposed to come back. The usual follow-up is some version of "how does that not wreck the tower, the rocket, or both?"

The answer is a chain of ordinary physics, and every link has to hold on a rocket that gets one try.

Why Is Catching A Falling Rocket Booster So Hard?

Drop a set of keys and gravity speeds them up the whole way down. But a rocket booster is a sleek metal cylinder, and air alone will never stop it. Left to fall, it would reach a terminal velocity far too high to survive. The only brake it carries is the set of engines on its tail.

Those engines work the same way on the way down as on the way up. They throw hot gas out one end, and the gas shoves the rocket the other way. That is Newton's third law doing the braking. Point the engines at the ground and the shove slows the fall. Match the booster's weight and it hovers.

In a video game that settles it. Real rocket engines hate running at a trickle. A booster coming home is nearly empty, so even a gentle push from a big engine is too much. Landing engineers put it in one line: "every second spent hovering is wasted propellant". That means the landing burn has to reach zero speed and zero height at the same instant. There are no spare seconds to fix a mistake.

Then there is the aim. Wind shoves the booster on the way up and again on the way down, scattering where it ends up. For Falcon 9, a precise landing meant squeezing that scatter to about ten meters (33 feet) at sea. A pair of arms leaves even less room.

Three frames from the first catch, October 13, 2024: the booster comes in from the left, slows beside the tower, and the arms close. (Photo Credit: Bech, Wikimedia Commons, CC BY 2.0)
Three frames from the first catch, October 13, 2024: the booster comes in from the left, slows beside the tower, and the arms close. (Photo Credit: Bech, Wikimedia Commons, CC BY 2.0)

What Happens To The Super Heavy Booster Between Launch And Catch?

Rockets do not go straight up. They tip over soon after liftoff and build sideways speed, because orbit is mostly about going sideways. By the time the booster lets go of the ship, it is far downrange and moving away fast.

Super Heavy's answer is to turn around and fly back. SpaceX's own timeline for the November 2026 flight runs like this:

  1. Most engines cut off two minutes and nineteen seconds after liftoff. The ship lights its own engines while still attached and pulls away. SpaceX calls this hot-staging.
  2. The flip and boostback burn. Seconds later the booster swings around. It relights its engines pointing the wrong way, which first kills its sideways speed and then pushes it back toward Texas.
  3. The coast. For roughly three minutes the engines are off. The booster is a huge thrown object on a long arc. The grid fins near its top do the steering.
  4. The landing burn, from about six minutes ten seconds to six minutes thirty-five. Thirteen engines light and brake hard. Then they shut down to three, which carry the booster the last stretch into the arms.
Super Heavy's round trip, schematic and not to scale: ascent, flip and boostback, a three-minute unpowered coast, then a 25-second landing burn into the arms. Timings from SpaceX's published Flight 15 timeline.
Super Heavy's round trip, schematic and not to scale: ascent, flip and boostback, a three-minute unpowered coast, then a 25-second landing burn into the arms. Timings from SpaceX's published Flight 15 timeline.

How Does The Booster Slow Down Without Landing Legs?

Falcon 9 taught SpaceX the braking part. Its single landing engine cannot throttle low enough to let the nearly empty stage hover. Instead the burn is timed to reach zero speed right at the deck, a maneuver fans nicknamed the "hoverslam".

Super Heavy has more engines to play with, and it uses them in steps. Thirty-three Raptors push with about 8,240 tonnes-force at liftoff. Thirteen sit in the center and can swivel to steer. For the landing burn the booster lights those thirteen and dumps most of its speed in seconds. Then it shuts down to the inner three. Three engines are still a lot of push for a mostly empty tank, so they run throttled back. The booster can slow to a near hover and slide sideways into position. On the first catch, observers described it hovering beside the tower as the arms closed.

Lighting thirteen and finishing on three is the same logic as braking hard early and gently late in a car. Thirteen engines shed speed while there is still height to spare. Three give fine control at the end, when a small error means a dent in the tower. It is also insurance. On the March 2025 flight only twelve of the thirteen relit. The catch still worked, because the plan never depended on every engine.

The same booster, three very different amounts of push. Full-throttle figures from SpaceX; each engine can also throttle down, so the push at the catch is lower still.
The same booster, three very different amounts of push. Full-throttle figures from SpaceX; each engine can also throttle down, so the push at the catch is lower still.

What Are Grid Fins, And What Part Of The Rocket Do The Arms Grab?

For the three minutes with the engines off, the booster still has to steer. It does that with grid fins: flat lattices of small surfaces in a frame. They look like an oversized waffle iron bolted near the top of the rocket. Tilt one and it deflects the air rushing past, nudging the booster the way a rudder nudges a boat.

A Falcon 9 grid fin, folded flat against the booster before launch. Tilted into the airflow on the way down, the lattice steers the stage. (Photo Credit: SpaceX, Flickr via Wikimedia Commons, CC0)
A Falcon 9 grid fin, folded flat against the booster before launch. Tilted into the airflow on the way down, the lattice steers the stage. (Photo Credit: SpaceX, Flickr via Wikimedia Commons, CC0)

Super Heavy's fins carry a second job. In December 2020, before any tower existed, Elon Musk announced the plan on Twitter. SpaceX would catch the booster with the launch tower arm, "using the grid fins to take the load". In practice the catch points are small hardpoints on the booster's body just below the fins. The arms close under those points. The booster's weight passes into the tower the moment the engines stop. The newest booster moved from four fins to three, each about half again as large. Its catch point also moved lower on the body to line up with the arms.

The arms ride a carriage that a winch hauls up and down the tower. Hydraulic rams swing them open and shut. Tracks on top of the arms let the tower shuffle the booster sideways and set it back on the launch mount. Musk's 2021 name for them was "robot chopsticks", and it stuck. A 72-meter rocket now gets handled like a dumpling.

Final approach on the first catch: the booster, still tilted, comes in above the open arms. The catch points sit just below the grid fins at the top of the booster. (Photo Credit: Steve Jurvetson, Flickr via Wikimedia Commons, CC BY 2.0; cropped)
Final approach on the first catch: the booster, still tilted, comes in above the open arms. The catch points sit just below the grid fins at the top of the booster. (Photo Credit: Steve Jurvetson, Flickr via Wikimedia Commons, CC BY 2.0; cropped)

Why Does SpaceX Use Chopsticks Instead Of Landing Legs?

Falcon 9 lands on four fold-out legs, and it does so well. Its most-flown booster launched and landed for the thirty-sixth time in July 2026. So why change a winning formula?

The other way to do it: a Falcon 9 first stage on its legs at Cape Canaveral after the first-ever orbital booster landing, December 2015. (Photo Credit: SpaceX, Flickr via Wikimedia Commons, CC0)
The other way to do it: a Falcon 9 first stage on its legs at Cape Canaveral after the first-ever orbital booster landing, December 2015. (Photo Credit: SpaceX, Flickr via Wikimedia Commons, CC0)

Musk gave the reasons in the same December 2020 thread. Legs are dead weight on the way up. Catching it instead "saves mass and cost of legs", he wrote. It also lets the tower put the booster straight back on the launch mount, "ready to refly in under an hour".

Every kilogram of leg is a kilogram of payload that never reaches orbit. Time matters more. A Falcon 9 booster that lands at sea has to be sailed back, lifted off the ship and checked over. A booster already in the arms of its own launch tower is most of the way to its next flight. SpaceX's design goal is to fly again without refurbishment. That part has not been demonstrated yet. But the direction is clear: move the hardware off the rocket and onto the ground, where weight is free.

How Many Booster Catches Has SpaceX Done, And Why Were Some Called Off?

As of mid-November 2026, the tower has caught three boosters and one ship. The first booster catch came on October 13, 2024, on the fifth test flight. It happened about seven minutes after liftoff, on the first try. Thousands of checks on the vehicle and the pad had to pass first. Two more boosters were caught in January and March 2025.

On the sixth flight, in November 2024, health checks on the tower itself failed. The booster flew a pre-planned detour to a soft splashdown in the Gulf of Mexico. On the most recent flight, in November 2026, the booster ran low on fuel and its boostback burn ended early. Its own guidance called off the catch during the coast. It hovered over the water, shut down, and was destroyed on purpose. That is the design working. The tower is the expensive half of the pair, so the booster only comes near it if every check says yes.

The same night, the ship came back from orbit after 22.5 hours and fourteen laps of the planet. A matching pair of arms caught it, a first. The upper stage is the harder target, since it comes in higher, farther and faster than the booster. SpaceX's commentator called it "the holy grail of spaceflight" on the webcast. The control room played "Baby Come Back". Musk's forecast from 2021 still holds: "Success is not guaranteed, but excitement is."

So, How Do The Chopstick Arms Catch A Falling Rocket Booster?

Put the chain together. The booster cannot hover for long and cannot steer with wings. It has to arrive at the right place at the right speed in one continuous move. The boostback burn sends it home. The grid fins trim the aim through three minutes of falling. Thirteen engines bleed off most of the speed and three bring it to a crawl. The swiveling center engines slide it sideways until the catch points sit over the arms. The arms close, the engines stop, and the tower takes the weight.

Notice what is missing from that list: legs, parachutes, a landing pad. The hard parts moved from the rocket to the ground, where they can be heavy and bolted to concrete. The rocket's job shrank to flying to the right spot.

The unsolved part is the boring one: doing it every time. Three catches from three boosters that were cleared to come in, two sent to the sea, and no booster yet caught twice. The next test is not whether the arms can catch a rocket. The next test is whether a caught rocket can be fueled, stacked and flown again in days rather than months. That cadence is the only reason anyone built the arms.

References (click to expand)
  1. Starship vehicle page (height, propellant, thrust, engine layout, reusability goal) — SpaceX
  2. Starship Flight 15 mission summary and timeline — SpaceX
  3. Starship's Fifth Flight Test — SpaceX
  4. Starship's Sixth Flight Test — SpaceX
  5. Starship's Eighth Flight Test — SpaceX
  6. SpaceX Catches a Falling Starship—A First in Spaceflight History — Scientific American
  7. 'Now everything changes': SpaceX catches Starship upper stage for 1st time ever — Space.com
  8. SpaceX targets bold new 'catch' strategy for landing Super Heavy rockets — Space.com
  9. SpaceX's Super Heavy booster will fly with fresh 'grid fins' — Space.com
  10. Elon Musk says SpaceX will attempt to recover Super Heavy rocket by catching it with launch tower — TechCrunch
  11. SpaceX will use 'robot chopsticks' to catch massive Starship rocket, Elon Musk says — The Independent
  12. Autonomous Precision Landing of Space Rockets (Blackmore), Frontiers of Engineering 2016 — National Academies Press
  13. Advancing Supersonic Retropropulsion Using Mars-Relevant Flight Data (Braun and Sforzo) — NASA Technical Reports Server
  14. SpaceX launches Falcon 9 rocket on record-breaking 36th flight — Spaceflight Now
  15. SpaceX Super Heavy (supplementary: grid fin count, hardpoints, flight sequence) — Wikipedia
  16. SpaceX Starbase (supplementary: tower arm carriage, winch and hydraulics) — Wikipedia
  17. List of Starship launches (supplementary: flight-by-flight outcomes) — Wikipedia