The N-1: How the Soviet Moon Rocket Failed — and How Its Engines Ended Up Flying for America

August 5, 2026

The N-1 was the Soviet Union's entry in the race to land a crew on the Moon: a super-heavy launcher 105 meters tall, powered at liftoff by thirty first-stage engines, and approved by the Kremlin in 1964 with the explicit goal of beating America. It never completed a single successful flight. Four test launches between 1969 and the early 1970s ended in four failures, and the program's legacy was nearly written off as a cautionary tale. But the story has an unusual second act. Engines from the N-1's propulsion family were placed in storage for roughly two decades, then sold to the United States, refurbished by Aerojet, and flown in 2013 as the AJ-26 on the Antares rocket — hardware built to defeat the American space program, ultimately flying for it. Drawing on program histories maintained by RussianSpaceWeb (Anatoly Zak) and technical records from Encyclopedia Astronautica (Mark Wade), this feature reconstructs what the N-1 was, why it failed, and why its engine cluster concept looks far less eccentric in the era of Falcon Heavy and Starship.

I. Conception, Approval, and the L3 Lunar Plan

According to RussianSpaceWeb's program history, work on what became the N-1 began in the late 1950s at the OKB-1 design bureau, led by chief designer Sergei Korolev. The original concept was not a dedicated Moon rocket at all but a multipurpose super-heavy booster intended to support space stations, lunar missions, and eventually Mars expeditions, with an early payload concept of roughly 75 tonnes to low Earth orbit. That changed in 1964, when the Kremlin formally approved the program with a single goal: beat the United States to a crewed lunar landing. The payload requirement was raised to 92–95 tonnes to LEO, and six engines were added to the first stage design to lift it. The launcher was only one element of the plan. The N-1 was to carry the L3 expedition complex, which would send a crew of two toward the Moon, with a single cosmonaut descending to the surface — a notably smaller expedition than the American effort it was racing against. Later L3M studies envisaged three cosmonauts, hydrogen upper stages, and a two-launch mission scenario, but these remained paper plans. On the ground, the program at least reached hardware: in 1967, the 1M1 full-scale mock-up was assembled at the Tyuratam launch site — the cosmodrome known publicly as Baikonur — for fit and interface testing. In its final form, the vehicle stood 105 meters tall with a maximum diameter of 17 meters, a liftoff mass of 2,783–2,825 tonnes, and a dry mass of 277–281 tonnes. Its three lower stages carried 1,780 tonnes of liquid oxygen and 680 tonnes of kerosene, produced a liftoff thrust of 4,500 tonnes, and were rated to deliver 95 tonnes to low Earth orbit. The specifications below are drawn from RussianSpaceWeb's program records.

II. Thirty Engines: The Propulsion Gamble

The defining engineering decision of the N-1 was its engine count. The Block A first stage carried 30 engines — 42 across all three stages — an arrangement chosen in part because of a split within the Soviet propulsion establishment. Initial propulsion studies had involved Valentin Glushko's OKB-456 bureau, but the well-documented rift between Korolev and Glushko meant that development of all three lower stages ultimately went to Nikolai Kuznetsov's OKB-276, an organization without a prior record of building engines for launch vehicles of this class. The engines ran on liquid oxygen and kerosene using staged combustion, an efficient but demanding cycle. Two consequences followed. First, the vehicle's 4,500 tonnes of liftoff thrust had to be generated by a large cluster of comparatively modest units — an average of roughly 150 tonnes of thrust per first-stage engine — all of which had to ignite, throttle, and shut down in coordination. Second, and more consequential, the program flew without ever conducting a full-stage ground test of the 30-engine Block A. RussianSpaceWeb's account is blunt on this point: thirty engines to integrate, and no all-up firing of the first stage before its first flight. In effect, each launch doubled as the first integrated test of the most complex propulsion stage built up to that time. That decision frames everything that happened next.

III. Four Launches, Four Failures

The N-1 flew four times from Site 110 at Tyuratam, and none of the flights succeeded. The first attempt, on 21 February 1969, involved Vehicle No. 3L carrying a 7K-L1A spacecraft; the flight ended 68.7 seconds after liftoff. The second, on 3 July 1969, was the most destructive: Vehicle No. 5L failed at the moment of liftoff in a pad explosion at Site 110 — a failure that occurred just over two weeks before the Apollo 11 mission it had been intended to preempt. The third attempt came nearly two years later, on 27 June 1971, when Vehicle No. 6L, carrying a dummy payload, also failed in flight. A fourth vehicle was launched subsequently and failed as well, closing the program's flight record at zero successes in four attempts. It is worth being precise about what the record does and does not show. The four failures had different immediate circumstances — one in-flight shutdown just over a minute in, one catastrophe on the pad, one later failure with a mass-simulator payload — so no single component can be blamed from the public record alone. What the record does support is a systemic interpretation: a first stage of unprecedented engine count, flown repeatedly without ever having been fired as a complete unit on the ground, was being debugged through full-scale launches. Each attempt consumed a complete vehicle. The table below summarizes the flight record as documented in RussianSpaceWeb's program history.

IV. The NK-33 and Its American Second Life

The N-1's cancellation did not mean the end of its engine lineage. According to Encyclopedia Astronautica's technical records, the Kuznetsov bureau had developed the NK-33 — also designated 11D111 — a liquid-oxygen/kerosene staged-combustion engine belonging to the NK-15 family created for the N-1, but modified with multiple-ignition capability and a longer service life. It was earmarked for the planned N1F follow-on family, which was never flown. Instead, the manufactured engines were placed in storage, where they remained for roughly twenty years. Their re-emergence reads like a post-Cold-War parable. In 1996, Kistler Aerospace selected the NK-33 for the first stage of its K-1 commercial reusable launcher; that project was cancelled before flying. The engines then found a second customer: the Taurus II rocket, later renamed Antares, which adopted the engine — redesignated AJ-26 by Aerojet, with the AJ26-62 variant designation — for its first stage. The AJ-26 made its first flight in 2013, some 38 years after the engines had been fabricated. Engines built to power the rocket that was supposed to beat America to the Moon thus ended their development odyssey lifting cargo for the American launch market. The arrangement had limits, and the record reflects them: Aerojet was developing an alternative engine because the available stock of NK-33s was finite, and the deterioration of US–Russian relations in 2014 threw the broader use of Russian-built engines on American launch vehicles into question. The second life of the N-1's propulsion work was real, but it was always understood to be a finite resource rather than a permanent supply line.

V. The Cluster Lesson: From the N-1 to Falcon Heavy and Starship

For decades, the N-1 was cited as proof that clustering large numbers of engines on a single stage was unworkable. The modern record suggests a different conclusion. SpaceX's Falcon Heavy flies with 27 engines on its combined first stage, and Starship's booster is designed around 33 engines — both figures exceeding the 27–33 range that brackets the N-1's own 30 — and the multi-engine cluster concept has been made to work operationally. What distinguishes the modern cases is not the engine count itself but the treatment of integration risk: the N-1's Block A never underwent a full-stage ground firing before flight, so each of its four launches was simultaneously an attempt at an operational mission and the first test of its propulsion system as an integrated whole. Read that way, the N-1's lesson is less 'thirty engines cannot work' than 'thirty engines cannot be certified by flying them.' The irony of the program is therefore double. Its hardware was vindicated twice over: the NK-33 demonstrated its soundness by flying successfully on Antares decades after manufacture, and the architectural bet behind the N-1 — many smaller engines rather than a few enormous ones — became standard practice on the most-flown heavy launchers of the following century. What the program never recovered from was the combination of a rushed 1964 mandate, a fractured propulsion establishment, and a test philosophy that substituted four expendable super-heavy vehicles for a ground test stand. For programs today that are again flying boosters with thirty-plus engines, the N-1 remains the reference case: not for why engine clusters fail, but for how much depends on proving them before they leave the pad.

VI. FAQ

• Q: What was the N-1 rocket?
• A: The N-1 was the Soviet Union's super-heavy lunar launcher, developed by the OKB-1 design bureau under Sergei Korolev with work beginning in the late 1950s. Approved by the Kremlin in 1964 with the goal of beating America to the Moon, it stood 105 meters tall, had a liftoff mass of up to 2,825 tonnes, and was rated to carry 95 tonnes to low Earth orbit.
• Q: How many times did the N-1 launch, and did any flight succeed?
• A: The N-1 flew four test launches from Site 110 at the Tyuratam (Baikonur) cosmodrome, and all four ended in failure. The first, on 21 February 1969, failed 68.7 seconds after liftoff; the second, on 3 July 1969, exploded on the pad at liftoff; the third, on 27 June 1971, failed in flight; and a fourth attempt also failed.
• Q: Why is the N-1's first stage considered so risky?
• A: The Block A first stage used 30 engines — 24 in the original design plus six added when the payload requirement rose to 92–95 tonnes — and the program never conducted a full-stage ground test of that 30-engine cluster before flight. Each launch therefore doubled as the first integrated test of the stage.
• Q: What is the NK-33 engine?
• A: The NK-33 (also designated 11D111) is a Kuznetsov liquid-oxygen/kerosene staged-combustion engine from the NK-15 family developed for the N-1, modified with multiple-ignition capability and longer life. It was intended for the planned N1F follow-on rocket, which never flew, leaving the manufactured engines in storage for roughly twenty years.
• Q: Did Soviet moon-rocket engines really fly on an American rocket?
• A: Yes. After being selected in 1996 for Kistler Aerospace's cancelled K-1 launcher, the NK-33 was adopted for the first stage of the Taurus II rocket, later renamed Antares. Redesignated AJ-26 by Aerojet, the engine first flew in 2013 — 38 years after the engines were fabricated.
• Q: Which modern rockets use engine clusters comparable to the N-1's?
• A: SpaceX's Falcon Heavy flies with 27 engines and Starship's booster is designed around 33, bracketing the N-1's 30-engine first stage. Their operational record shows that the multi-engine cluster concept itself was sound — the N-1's decisive weakness was flying such a stage without ever test-firing it as a complete unit on the ground.