There are many sources like Our World in Data describe the cost to launch with SpaceX Falcon 9 as $2600 per kilogram or $1200 per pound but the cost to SpaceX is $629 per kilogram or $285 per pound. The ~$67–70 million external selling price is for about 30 out of the 170 missions in 2025 while the sub -$15 million. This error is like using the $50,000 Nvidia GPU selling price instead of the $10,000 Nvidia internal cost for GPUs. Nvidia has made its own data centers. Multiple sources, including old 2020-2023 Elon Musk’s statements in interviews, peg the marginal cost of a reused F9 launch at ~$15 million. I think this is $11 million. The Upper stage (expendable, must be built new each time): ~$10 million public estimates but I think this is $7 million. Booster refurbishment: ~$1 million (per Musk in 2020; this has likely decreased with experience, as boosters now routinely fly 30+ times). Propellant (LOX/RP-1): ~$0.2–0.3 million. Screenshot The Super Heavy Starship aspirational long-term marginal cost (high reuse, 100+ flights per vehicle, rapid cadence like 100+ launches/year) is $2–million per launch or even under $1 million in ultra-optimized scenarios (like 10,000-30,000 launches per year for the fleet). The Falcon 9 upperstage is at extreme scale in 2025 with 170 built. Marginal manufacturing might dip to $7 million through further automation, cheaper materials, and engine cost reductions. The merlin enignes might be below $1 million. SpaceX acquired the aerospace division of Hexagon Purus, known as Masterworks, a specialist in Type 4 composite overwrapped pressure vessels (COPVs), for approximately $15 million in late 2025. This vertical integration move primarily targets in-house production for Starship’s high-pressure systems but could also apply to Falcon 9’s COPV needs, reducing reliance on external suppliers. Starship was having problems with the quality of COPVs. Bad COPVs had caused loss of prototype Starships. Aerospace-grade COPVs (like those in Falcon 9) are estimated at $100,000–500,000 each based on industry analogs (high-pressure carbon-fiber composites for space). Assuming ~$200,000–300,000 per COPV (mid-range for Type 4 or COPV 2.0 variants), a 20% margin savings equates to ~$50,000 per unit. With 2–4 COPVs per upper stage then total savings ~$150,000 per launch. Merlin 1D Vacuum engine (1x). Elon Musk has implied sub-$1 million unit costs with mass production. Industry estimates range from $750,000–2 million. SpaceX is making over 300+ engines/year. They make expended upper stage and a few new boosters each year. I think the cost is $500k-750k. This is ~5% of total upper stage cost. The vacuum variant includes a larger extendable nozzle, adding ~20–30% to a standard Merlin 1D’s price. Tanks and structure for the upper stage ~$4–6 million (50–60% of total). This includes LOX/RP-1 tanks (aluminum-lithium alloys, cryogenic insulation), interstage, and pressurization systems. The COPVs contribute ~10–20% of this subcategory or about $0.5 to 1.0 million total with the new inhouse maker). Avionics and guidance are estimated at ~$1–2 million (10–20%). Flight computers, sensors, batteries, and communication hardware. This was likely brought down to the $1 million level. I think the upper stage costs are now in the $7-9 million range. SpaceX would get avionics and guidance to $1 million and I think the tanks can be brought to $4-5 million. F9 Booster build cost remains ~$30 million (mature production. 9 Merlin 1D engines dominate at lower unit costs with mass production, plus structure/avionics). Total marginal cost per kg: $11 million ÷ 17,500 kg ≈ $629/kg. Total marginal cost per lb: $11 million ÷ 38,580 lb ≈ $285/lb. SpaceX Starship Makes it Even Cheaper and Boosts Volume of Payload by 100X or even 1000s of Times Screenshot