SpaceX Drone Ships: OCISLY, ASOG, and JRtI Explained
You’ve seen the footage: a 47-meter Falcon 9 booster descends through a column of fire onto a tiny platform bobbing in the Atlantic Ocean. The crowd cheers. The webcast cuts to a camera on the deck shaking violently from engine exhaust. And somewhere in the frame, you catch the name painted in huge letters: Of Course I Still Love You.
SpaceX’s autonomous spaceport drone ships (ASDS) are some of the most remarkable vessels in maritime history — not because they’re big or fast, but because they hold position with meter-level accuracy in open ocean while a 25-ton rocket lands on them at 2 m/s. They’re the key that unlocked high-energy mission recovery, and without them, SpaceX’s reuse economics wouldn’t work.
This guide covers all three drone ships, how they work, why they exist, and the robotic systems that secure a freshly landed booster before it can tip over.
The Three Ships (And Why They’re Named That)
All three drone ship names come from sentient starships in Iain M. Banks’ Culture science fiction series. Elon Musk is a fan of the books, and the tradition of naming the ships after Culture vessels has become one of SpaceX’s most beloved quirks. In the novels, these were massive, AI-controlled spacecraft. In reality, they’re converted barges. The irony is intentional.
Just Read the Instructions (JRtI)
The original. JRtI entered service in January 2015 at SpaceX’s Pacific launch operations out of Vandenberg Space Force Base in California. It supported west-coast polar and sun-synchronous orbit missions, which fly south over the Pacific. After SpaceX consolidated most launches to Cape Canaveral, JRtI saw less frequent use but remains the primary Pacific recovery asset. It’s the smallest of the three, based on the Marmac 300 barge hull.
Of Course I Still Love You (OCISLY)
The workhorse. OCISLY has supported more booster landings than any other drone ship, stationed in the Atlantic for Cape Canaveral launches. It entered service in June 2015 and quickly became the star of SpaceX landing attempts — including the early failures that produced spectacular fireball footage. OCISLY was originally based at Port Canaveral but has also operated from Jacksonville, Florida. After years of service, OCISLY was retired from active landing operations in mid-2024 following over 50 successful catches, with its duties transferred to the newer ASOG.
A Shortfall of Gravitas (ASOG)
The newest and most capable. ASOG was purpose-built (rather than converted from an existing barge) and entered service in 2021. It’s stationed in the Atlantic and has gradually taken over primary landing duties from OCISLY. ASOG features an upgraded stationkeeping system and structural reinforcements based on lessons learned from hundreds of landing attempts on the earlier ships. It’s the drone ship you’re most likely to see on current SpaceX webcasts.
Dimensions and Physical Specs
All three drone ships share similar dimensions, though ASOG is slightly updated in construction. They’re essentially flat steel decks with thrusters underneath — imagine a floating parking lot that can hold itself perfectly still.
| Specification | JRtI / OCISLY | ASOG |
|---|---|---|
| Length | 91 m (300 ft) | 91 m (300 ft) |
| Width | 52 m (170 ft) | 52 m (170 ft) |
| Landing Zone | ~50 × 50 m | ~50 × 50 m |
| Displacement | ~9,500 tonnes | ~10,000 tonnes |
| Deck Marking | SpaceX “X” logo | SpaceX “X” logo |
| Base Barge | Marmac 300/304 | Purpose-built |
| Home Port | Port Canaveral / Vandenberg | Port Canaveral |
For perspective, the landing zone is about the size of a football field’s end zone. The Falcon 9 booster’s landing legs span about 18 meters. So the booster needs to hit a 50-meter target after traveling ~600 kilometers downrange at up to 2 km/s. The margin isn’t as tight as it looks — but it’s not generous either.
How They Hold Position: Stationkeeping
Here’s the engineering problem: you need a 9,500-tonne barge to hold position in open ocean — often 300-600 km offshore — with enough accuracy that a rocket can target it. The ocean has currents, waves, and wind. The barge has no crew aboard during landing (for obvious safety reasons). Everything is automated.
Four Azimuth Thrusters
Each drone ship is equipped with four diesel-powered azimuth thrusters that can rotate 360 degrees. Unlike a ship’s rudder (which only works when the ship is moving forward), azimuth thrusters can push in any direction at any time, giving the barge the ability to translate sideways, rotate in place, or hold against any combination of current and wind.
The thrusters are controlled by a dynamic positioning (DP) system — the same technology used on offshore oil drilling platforms that must hold position over a wellhead in rough seas. The DP system takes input from GPS receivers (typically getting 3-meter accuracy), motion sensors, gyrocompasses, and wind sensors, then continuously adjusts thruster output to maintain position.
GPS and Position Accuracy
The drone ships use differential GPS (DGPS) augmented by additional correction sources to achieve position accuracy in the range of 1-3 meters. This is more than accurate enough — the landing guidance algorithm on the booster targets the center of the deck, and the booster’s own landing accuracy (sub-2-meter) is the tighter constraint.
During the final seconds before landing, the ship doesn’t try to “catch” the booster or adjust position. It simply holds as steady as possible and lets the booster do the precision work. The ship is the target; the booster is the guided munition.
Sea State Limitations
Drone ships can operate in surprisingly rough conditions — SpaceX has landed boosters in Sea State 4-5 (waves of 1.25-2.5 meters). But there are limits. In very heavy seas, the deck motion exceeds what the booster’s landing legs can absorb, and the risk of the booster toppling after touchdown becomes unacceptable. SpaceX has scrubbed landing attempts (while still launching the mission) when sea conditions deteriorate beyond the safe envelope. The payload still reaches orbit; the booster just doesn’t attempt recovery.
Why Ocean Landings? (Why Not Fly Back Every Time?)
SpaceX does fly boosters back to the launch site when the mission profile allows it — these are called Return to Launch Site (RTLS) landings, and they touch down at Landing Zone 1 or 2 at Cape Canaveral. So why bother with drone ships at all?
The answer is delta-v. After stage separation, the first stage is flying downrange at 6,000-8,000 km/h. To fly all the way back to the launch site, the booster needs to reverse its direction entirely — a boostback burn that consumes a large amount of propellant. This propellant has to be reserved before launch, which means less propellant available for the ascent, which means less payload to orbit.
For heavy payloads or high-energy orbits (like geostationary transfer), there’s simply not enough propellant margin for a boostback burn. The booster separates at higher velocity, coasts farther downrange, and lands on a drone ship positioned 300-600 km from the launch site. The booster still performs an entry burn (to survive reentry heating) and a landing burn (to touch down gently), but it doesn’t need to reverse course.
The trade-off math is straightforward. An RTLS landing might reduce payload capacity by 30-40% compared to an expendable mission. A drone ship landing reduces it by only 15-20%. For Starlink missions (where SpaceX controls the payload mass), RTLS often works. For heavy commercial payloads, drone ships are essential for making reuse economically viable.
Octagrabber: The Robot That Saves the Booster
Landing a 25-tonne, 47-meter-tall rocket on a barge is impressive. Keeping it upright afterward is a separate challenge entirely.
A landed Falcon 9 booster is standing on four landing legs with a very high center of gravity. In calm seas, this is fine — the legs are wide enough to keep it stable. But as the drone ship transits back to port (an 8-12 hour journey), ocean swells can rock the barge enough to topple the booster. Early on, SpaceX ground crews would board the drone ship by helicopter and manually weld steel shoes over the landing legs to secure the booster — a dangerous, time-consuming process.
Enter Octagrabber: a low-profile, autonomous robot that lives on the drone ship deck. After a booster lands, Octagrabber rolls under the engine section on tank treads and clamps onto the base of the rocket, physically securing it to the deck. The whole process takes roughly 45-60 minutes and requires zero human presence.
How Octagrabber Works
Octagrabber is essentially a remote-controlled (with autonomous capability) platform on tracks. It carries steel gripper arms that interface with specific hard points on the Falcon 9’s octaweb — the structural cage that holds the nine Merlin engines. By clamping onto the octaweb and anchoring itself to the deck, Octagrabber creates a rigid connection between the booster and the ship that can withstand significant sea motion.
The robot is custom-built for the Falcon 9’s geometry. SpaceX has built updated versions for different booster variants, and there’s presumably a version in development (or already built) for Falcon Heavy side boosters and potentially for Super Heavy (Starship’s first stage), though Super Heavy’s current recovery plan involves catching it with the launch tower’s “chopstick” arms rather than landing on drone ships.
Landing Accuracy and Success Rate
SpaceX’s drone ship landing record is remarkable, especially considering they started from zero experience in 2015. The first several attempts ended in spectacular failures — too much lateral velocity, insufficient throttle response, a leg failing to lock, hydraulic fluid running out for the grid fins. Each failure taught lessons that were incorporated into software and hardware updates.
By 2020, drone ship landings had become so routine that a failure was newsworthy. Current landing accuracy is consistently within 2 meters of the target center on the deck. The booster’s G-FOLD guidance algorithm (the same convex optimization system that computes fuel-optimal trajectories in real time) has been refined through hundreds of flights to the point where it’s essentially a solved problem.
As of early 2026, SpaceX has completed over 300 successful booster landings across drone ships and land pads combined, with the vast majority on drone ships. The success rate over the last 100+ attempts exceeds 98%, with the rare failure typically attributed to hardware issues (engine problems, leg mechanism failures) rather than guidance errors.
The Journey Home: From Landing to Reuse
After Octagrabber secures the booster, the drone ship begins the transit back to port, typically taking 8-12 hours depending on distance and sea conditions. Support vessels (usually the tugboat that towed the drone ship to position, plus SpaceX’s recovery support ships like GO Searcher or GO Navigator) accompany the convoy.
At port, a crane lifts the booster off the drone ship and places it horizontally on a transporter. It’s then trucked to SpaceX’s refurbishment facility, where it’s inspected, serviced, and prepared for its next flight. Turnaround time has decreased dramatically — SpaceX has reflown a booster as quickly as 21 days after its previous mission, though typical turnaround is 30-60 days.
The drone ship itself requires minimal maintenance between missions. The deck gets inspected for heat damage (Merlin exhaust burns at ~3,300°C, though the exposure time is measured in seconds), the thrusters get checked, and the DP system gets calibrated. SpaceX has turned drone ships around for the next landing in as little as a few days.
Drone Ships vs. The Starship Catch Tower
With Starship, SpaceX is moving toward a different recovery concept: catching the Super Heavy booster with the launch tower’s mechanical arms (“Mechazilla” / “chopsticks”) rather than landing it on a drone ship. The booster hovers near the tower, and the arms close around hard points on the vehicle.
Why the change? Super Heavy is significantly larger and heavier than Falcon 9’s first stage, and the landing legs needed for a drone ship landing would add substantial mass. By catching at the tower, SpaceX eliminates landing legs entirely, saves the mass for payload or propellant, and can immediately reposition the booster for stacking and reflight. The dream is a booster that launches, gets caught, gets restacked, and launches again within hours.
But drone ships aren’t going away. Falcon 9 will continue flying for years alongside Starship, and the drone ship infrastructure (ships, port facilities, Octagrabber, support vessels) represents a mature, reliable system. Drone ships might also serve as emergency landing options for Starship missions where a tower catch isn’t feasible.
Full Drone Ship Fleet Overview
| Ship Name | Entered Service | Ocean | Status (2026) | Notable Milestone |
|---|---|---|---|---|
| Just Read the Instructions | Jan 2015 | Pacific | Active (low frequency) | First Pacific drone ship landing |
| Of Course I Still Love You | Jun 2015 | Atlantic | Retired from landings (2024) | 50+ successful catches, most of any ASDS |
| A Shortfall of Gravitas | 2021 | Atlantic | Primary active ship | Purpose-built (not a barge conversion) |
Frequently Asked Questions
Is there anyone on the drone ship when the rocket lands?
No. The drone ship is completely uncrewed during landing operations. All personnel are evacuated to support vessels stationed several kilometers away. The stationkeeping system operates autonomously, and the landing cameras are remote. SpaceX ground crews only board the drone ship after the booster is confirmed stable and Octagrabber has secured it — typically several hours after landing.
What happens if the booster misses the drone ship?
It goes into the ocean. The booster’s flight termination system and guidance algorithm include safety constraints that would divert the booster away from the drone ship if something goes catastrophically wrong during the landing burn. In early attempts, several boosters hit the drone ship but failed to land successfully — tipping over or exploding on deck. The resulting damage to the drone ship was repairable, though dramatic. Since those early days, the guidance system has been refined to the point where complete misses are essentially nonexistent.
How far from shore do drone ships park for a landing?
Typically 300-600 kilometers offshore, depending on the mission profile. Higher-energy missions (heavier payloads, higher orbits) result in the booster separating at higher velocity and coasting farther downrange, pushing the landing zone further from shore. Starlink missions, which are relatively light, tend to have closer landing zones. The exact position is computed mission-by-mission based on the booster’s predicted trajectory after stage separation.
Could drone ships be used for Starship booster landings?
Theoretically yes, but it’s not the current plan. Super Heavy boosters are designed to fly back to the launch site and be caught by the tower arms, eliminating the need for landing legs and ocean recovery logistics. However, SpaceX could develop a drone ship landing capability for Super Heavy as a backup or for missions where a return-to-launch-site profile isn’t feasible. No such development has been publicly announced.
Why are the names so weird?
They’re from Iain M. Banks’ Culture novel series, where hyper-intelligent AI starships choose their own ironic, philosophical, or humorous names. Elon Musk is a fan of the series and adopted the tradition for SpaceX’s drone ships. The names are intentionally playful — a contrast to the deadly serious engineering underneath. Other Culture ship names that might show up on future SpaceX vessels include So Much For Subtlety and Experiencing A Significant Gravitas Shortfall.
How much does a drone ship cost?
SpaceX hasn’t disclosed exact figures, but estimates based on barge conversion costs, thruster systems, and DP equipment put each drone ship at roughly $20-40 million. That sounds like a lot until you consider that each successful landing recovers a Falcon 9 booster worth an estimated $30-35 million in manufacturing cost. A single successful landing essentially pays for the drone ship. With over 300 successful landings and counting, the fleet has been one of SpaceX’s most cost-effective investments.