Thirty-three of these engines fire at once to lift the most powerful rocket ever built. They run at pressures no other flying engine has ever reached. And SpaceX wants to build them for the price of a pickup truck. The Raptor engine is a full-flow staged combustion methane engine producing up to 2,256 kN of thrust — and it’s the engine designed to put humans on Mars.

That’s not marketing. Every major design decision behind Raptor — the fuel, the thermodynamic cycle, the manufacturing approach, the cost target — traces directly to one architecture: landing people on Mars and keeping them there. It’s simultaneously the most advanced liquid rocket engine ever mass-produced and, if SpaceX hits its targets, one of the cheapest. That contradiction is the entire point.

Raptor Engine Specifications

Specification	Raptor 1	Raptor 2	Raptor 3	RVac
Thrust (sea level)	1,813 kN (407,000 lbf)	2,256 kN (507,000 lbf)	~2,747 kN (617,000 lbf)*	N/A (vacuum-optimized)
Thrust (vacuum)	~1,960 kN (440,000 lbf)	~2,450 kN (551,000 lbf)	~2,940 kN (661,000 lbf)*	~2,550 kN (573,000 lbf)
ISP (sea level)	~310 s	~327 s	~330 s*	N/A
ISP (vacuum)	~340 s	~350 s	~355 s*	~380 s
Chamber pressure	~250 bar (3,626 psi)	~300 bar (4,351 psi)	~350 bar (5,076 psi)*	~300 bar (4,351 psi)
Expansion ratio	~40:1	~40:1	~40:1	~80:1
Propellants	Sub-cooled liquid methane (CH₄) / sub-cooled liquid oxygen (LOX)
Cycle	Full-flow staged combustion (FFSC)
Mixture ratio (O/F)	~3.6:1
Cost target	~$1M	~$500K	~$250K*	TBD
Nozzle extension	Regeneratively cooled	Regeneratively cooled	Filmless (radiative)	Niobium alloy extension

* Target / projected figures for Raptor 3.

The Engine Cycle Nobody Could Build

For decades, full-flow staged combustion was a textbook fantasy. Thermodynamically perfect, practically impossible. The Soviets tried with the RD-270 in the 1960s — extensive test campaigns, never flew. After that, nobody tried again for 40 years. SpaceX didn’t just try. They mass-produced it.

How Engine Cycles Work (The Quick Version)

Every large rocket engine needs turbopumps to push propellant into a high-pressure combustion chamber. Turbines need hot gas to spin. Where that gas comes from defines the engine cycle:

• Gas generator (open): Burn a small amount of propellant in a side chamber, spin the turbines, dump the exhaust overboard. Simple and reliable, but you’re throwing away 2-3% of your propellant. Merlin and the F-1 use this.
• Staged combustion (closed): Route the turbine exhaust into the main chamber for a second burn. Nothing wasted. This is how the RS-25 and RD-180 work. The preburner runs either fuel-rich or oxidizer-rich, but not both.
• Full-flow staged combustion: Two preburners — one fuel-rich, one oxidizer-rich. All fuel passes through one as gas. All oxidizer passes through the other as gas. Both streams enter the main chamber already gasified. Nothing wasted, no liquid injection, no mixing inefficiencies.

Why FFSC Is Better

Because all propellant enters the main chamber as gas, combustion is more complete and uniform. This allows higher chamber pressures without combustion instability — Raptor runs above 300 bar, compared to 207 for RS-25 and 107 for Merlin. Higher pressure means more thrust from a smaller engine.

The dual-preburner setup also means each turbine runs cooler than in a single-preburner design. Cooler turbines last longer. Longer-lasting turbines enable reusability.

And there’s a huge bonus for landing: because both propellant streams are already gasified, Raptor can throttle down below 40% without risking combustion instability. That’s critical when a nearly empty Starship is settling onto a landing pad.

Why Nobody Could Build It Before

The oxidizer-rich preburner. Fuel-rich preburners are manageable — excess methane is relatively gentle on metal. But an oxygen-rich preburner produces superheated, high-pressure gas that is mostly oxygen. Hot oxygen destroys almost everything it touches.

SpaceX solved it with an in-house superalloy called SX500, specific coatings, and modern manufacturing (3D printing, advanced casting) that simply didn’t exist in the 1960s. Plus computational fluid dynamics that let them model combustion at resolutions Soviet engineers couldn’t dream of.

Why Methane? Because Mars.

SpaceX could have used hydrogen. They could have stuck with kerosene. They chose methane, and the reason lives on the surface of Mars.

You Can Make It on Mars

Mars has CO₂ in its atmosphere (95%) and water ice in its soil. The Sabatier reaction combines them to make methane:

CO₂ + 4H₂ → CH₄ + 2H₂O

A solar-powered chemical plant on Mars could manufacture return fuel from local resources. No other high-performance rocket propellant offers this. Without methalox, every return trip requires shipping fuel from Earth — a non-starter for any permanent presence.

No Coking

Kerosene leaves carbon deposits (coke) inside cooling channels and on injector faces. SpaceX learned this the hard way with Falcon 9 — inspecting and cleaning after flights. Methane burns clean. One carbon atom, four hydrogens, minimal soot. For an engine designed to fly hundreds of times, that’s a big deal.

Denser Than Hydrogen

Hydrogen gives the best specific impulse, but it’s so light (71 kg/m³) that you need enormous tanks. Liquid methane is six times denser at 422 kg/m³. Smaller tanks, lighter structure, partially offsetting the ISP gap. For a vehicle Starship’s size, methane’s density advantage is decisive.

No Helium Required

Most rockets use helium to pressurize tanks. Helium is expensive, the bottles are heavy, and they add failure modes. Methane and oxygen can pressurize their own tanks — warm methane gas pressurizes the fuel tank, warm oxygen gas pressurizes the oxidizer tank. Starship has zero helium in its propulsion system. Every component that isn’t there can’t fail.

Temperature Compatibility

Liquid methane (-161°C) and liquid oxygen (-183°C) store at similar temperatures. Liquid hydrogen boils at -253°C, creating severe thermal gradients when paired with LOX. The methane-oxygen pair can share a common bulkhead between tanks — which Starship does — saving significant structural mass.

Raptor’s Evolution

Raptor 1 (2016-2022): Make It Work

First fired at McGregor, Texas in September 2016. Proved FFSC worked at scale — 250 bar, 1,813 kN. Large, complex, expensive, covered in sensors and shielding. These were the learning engines. They powered Starship prototypes through the hop program (SN5, SN6, SN8 through SN15) and early orbital vehicles.

Raptor 2 (2022-2025): Make It Simple

Classic SpaceX simplification. Cut ~30% of the parts. Reduced mass. Increased thrust 25% to 2,256 kN. Pushed chamber pressure to ~300 bar — the highest of any operational engine in history. Fewer sensors (fewer failure points), more 3D-printed components, fewer bolted interfaces. Powered IFT-1 through IFT-7.

Raptor 3 (2025-): Make It Cheap

The headline: SpaceX removed the engine heat shield entirely. Previous Raptors had a metal shield to protect from recirculating exhaust. Raptor 3 uses filmless cooling — the outer nozzle wall radiates heat directly, glowing cherry-red during operation. No shielding. No film cooling.

Think about how bold that is. Film cooling — injecting cooler propellant along nozzle walls — has been standard since the V-2. Removing it requires absolute confidence in your thermal models, materials, and regenerative cooling. SpaceX has that confidence from testing thousands of Raptors.

Target: ~2,747 kN thrust, ~350 bar chamber pressure, and approximately $250,000 per engine. For context, each RS-25 on the Space Launch System costs NASA about $146 million. Even Merlin costs $1-2 million. At $250K, Raptor 3 costs less than many luxury cars.

RVac (Raptor Vacuum)

Same core powerpack as standard Raptor, but with a massive niobium alloy (C-103) nozzle extension. Expansion ratio jumps to 80:1 (vs. 40:1), boosting vacuum ISP to ~380 seconds. The nozzle glows white-hot in operation, shedding heat into space. Three RVac engines sit on the outer ring of Starship Ship, complementing three sea-level Raptors in the center.

39 Engines Working as One

Super Heavy: 33 Raptors

Thirty-three engines in concentric rings: 13 outer (gimbaling for steering), 20 inner (fixed). Combined liftoff thrust: approximately 74,500 kN — roughly twice Saturn V and the most powerful rocket stage ever built.

The coordination is staggering. Each Raptor has its own dual-preburner system, turbopumps, and throttle control. The flight computer manages 33 independent powerplants simultaneously, balancing thrust, compensating for underperformers, and steering through differential throttling and gimbaling. The propellant feed must deliver thousands of kilograms per second split evenly across 33 engines.

For landing, Super Heavy lights a subset of inner engines and throttles deep. SpaceX has already demonstrated the booster returning to the launch tower and being caught by mechanical “chopstick” arms.

Starship Ship: 3 + 3

Six engines total: three sea-level Raptors in the center (for landing burns where extended nozzles would be damaged by ground proximity) and three RVac on the outer ring (for peak orbital performance). All six can fire together during ascent. Only the center three light for landing.

Hot Staging

Starting with IFT-2, SpaceX ignites the Ship’s engines before Super Heavy separates. The Ship fires through a vented interstage while still attached. Soviet/Russian rockets have done this for decades, but it was new to SpaceX. It costs some booster structural life but improves payload by roughly 10% — a massive gain from a procedural change, not a hardware redesign.

IFT Flight History: Learning Fast

IFT-1 (April 2023)

Only 27 of 33 Raptors lit at liftoff. More failed during ascent. Vehicle lost control at ~4 minutes, self-destructed. The launch pad itself was destroyed by unmitigated exhaust. But the data was invaluable — SpaceX identified failure modes, feed issues, and thermal gaps. They also built a massive steel-and-water flame deflector they’d initially hoped to skip.

IFT-2 (November 2023)

All 33 engines lit and burned through the full boost phase. Hot staging worked. The booster broke up during boostback (propellant feed issue, not an engine problem). The Ship reached near-orbital velocity before flight termination. Engine reliability jumped from ~82% to near 100% in one iteration.

IFT-3 (March 2024)

All 33 booster engines nominal. Ship reached orbital velocity for the first time. Demonstrated propellant transfer and payload bay door operations in space. Ship broke up during reentry — valuable thermal protection data. Engine reliability: effectively 100% during powered phases.

IFT-4 (June 2024)

The flight where everything worked. All 39 engines — 39 for 39. Booster executed a controlled splashdown in the Gulf of Mexico. Ship survived reentry, proved the heat shield tiles, and performed a controlled landing burn before splashing down in the Indian Ocean. Engine performance was no longer the limiting factor.

IFT-5 (October 2024)

The historic one. All 39 nominal. Super Heavy returned to the launch site and was caught by the tower arms — broadcast live, one of the most-watched moments in spaceflight history. Ship survived reentry and splashed down in the Indian Ocean.

IFT-6 (November 2024)

Full engine reliability. Booster splashdown. Ship deployed simulated Starlink satellite dispensers in space before reentry and splashdown.

IFT-7 (January 2025)

All 39 nominal again. Second successful tower catch, confirming IFT-5 was no fluke. Ship reached orbit, deployed test payloads, and completed controlled reentry.

The progression: from 27/33 on IFT-1 to 39/39 routinely by IFT-4. SpaceX achieved in seven flights what most programs take decades to accomplish.

Raptor vs. the Competition

Parameter	Raptor 2	Merlin 1D+	BE-4	RS-25	RD-180
Manufacturer	SpaceX	SpaceX	Blue Origin	Aerojet Rocketdyne	NPO Energomash
Cycle	Full-flow staged combustion	Gas generator (open)	Ox-rich staged combustion	Fuel-rich staged combustion	Ox-rich staged combustion
Propellants	CH₄/LOX	RP-1/LOX	CH₄/LOX	LH₂/LOX	RP-1/LOX
SL thrust	2,256 kN (507K lbf)	845 kN (190K lbf)	2,400 kN (540K lbf)	1,860 kN (418K lbf)	3,830 kN (861K lbf)
Vac ISP	~350 s	~311 s	~340 s	~452 s	~338 s
Chamber pressure	~300 bar	~107 bar	~134 bar	~207 bar	~267 bar
Reusable?	Yes (designed for 100+ flights)	Yes (demonstrated 20+ flights)	Yes (designed for 25+ flights)	Yes (designed for 55 flights)*	No (expendable)
Approx. cost per engine	~$500K	~$1–2M	Not public	~$146M	~$25M (historical)
First flight	2023 (IFT-1)	2013	2024 (Vulcan)	1981 (STS-1)	2000 (Atlas III)

* RS-25 was designed for reuse on the Space Shuttle but is used expendably on SLS.

A few things jump out. Raptor’s chamber pressure is the highest of any operational engine — and it isn’t close. RS-25 wins on vacuum ISP because hydrogen is simply a higher-energy fuel, but RS-25 costs roughly 290 times more per engine. When your architecture needs 33 engines per booster and you plan thousands of missions, cost per engine isn’t secondary — it’s primary.

Blue Origin’s BE-4 is Raptor’s closest analog: methane-LOX, designed for reuse. But BE-4 uses simpler ox-rich staged combustion (not full-flow), runs at less than half Raptor’s chamber pressure, and achieves lower vacuum ISP.

Manufacturing: How Do You Build This for $250K?

SpaceX builds Raptors at a pace no liquid rocket engine has matched — reportedly exceeding one per day during peak production. Every Starship stack needs 39. If SpaceX hits its goal of multiple launches per week, they need hundreds per year.

Three strategies drive the cost down:

• Parts elimination: Raptor 3 has dramatically fewer components than Raptor 1. Every eliminated part is one you don’t machine, inspect, or assemble. SpaceX’s culture relentlessly questions whether each component is necessary. If it might not be, they remove it and test without it.
• 3D printing: Critical components that would traditionally need dozens of machined parts and welds are printed as single pieces in nickel superalloys. This collapses assembly time, eliminates weld joints (common failure points), and enables geometries impossible with conventional machining — like optimized internal cooling channels.
• Vertical integration: SpaceX makes nearly every component in-house. No vendor markup, no supply chain risk, and engineers can optimize the full system rather than individual parts in isolation.

What’s Next

Raptor’s development is far from done. SpaceX is pushing toward 330+ tons-force from a single sea-level engine — which would rival the F-1 that powered Saturn V. The real test isn’t any single performance metric. It’s reliability at scale: 39 engines at a time, flight after flight, minimal refurbishment.

The IFT campaign proved it’s achievable. The question is whether it’s sustainable at the flight rates SpaceX envisions — dozens of launches per year, scaling to hundreds.

Every Raptor that fires is part of the most aggressive engine development program since the Shuttle Main Engine. The difference: SpaceX is doing it at 1/290th the per-unit cost, five times the production volume, and with reuse as a non-negotiable requirement rather than an afterthought.

Frequently Asked Questions