The 16,500-Foot Recovery: The Loss of K-129 and the Birth of the Glomar Response, 1968 to 1976

On 24 February 1968 the Soviet Project 629A diesel-electric ballistic missile submarine K-129 (NATO Golf-II class) sailed from Rybachiy submarine base at Petropavlovsk-Kamchatsky for a routine seventy-day deterrent patrol in the Central Pacific. She carried ninety-eight men under Captain First Rank Vladimir Ivanovich Kobzar, with three R-21 (SS-N-5) liquid-propellant submarine-launched ballistic missiles in fin silos and nuclear-tipped torpedoes. She made an initial test-dive radio report, then went quiet. The US Navy’s SOSUS network recorded an acoustic event consistent with a submarine implosion in the Central Pacific on or about 8 March 1968 (Sontag 1998, Sharp 2012) or 11 March 1968 (per US Navy and CIA-derived material). On or about 20 August 1968 the modified US Navy submarine USS Halibut (SSGN-587) photographed the wreck at roughly 16,500 feet. Six years later, in August 1974, the purpose-built CIA ship Hughes Glomar Explorer lifted part of her forward section from the seabed. Press disclosure followed in 1975; the next year a federal court turned the agency’s refusal to comment into a doctrine. The open question that still sits at the centre of the case is what actually killed K-129.

The documented account

K-129, the sailing, and the disappearance

K-129 was a Project 629A submarine commissioned in 1960 and assigned to the Soviet Pacific Fleet 15th Submarine Squadron. Her armament on this patrol was three R-21 (SS-N-5 “Serb”) liquid-propellant SLBMs in fin-mounted silos, plus nuclear-tipped Type 53 and conventional torpedoes. The standard complement was about eighty-three; the Russian Federation under Boris Yeltsin posthumously awarded the “For Courage” medal to ninety-eight K-129 sailors, and ninety-eight is the figure used in Russian official commemoration. The commanding officer was Captain First Rank Vladimir Ivanovich Kobzar, the executive officer Captain Second Rank Aleksandr Mikhailovich Zhuravin, with a KGB cypher specialist embarked. She sailed from Rybachiy on 24 February 1968 and transited eastward at snorkel depth.

Soviet doctrine required scheduled radio check-ins when K-129 crossed the 180th meridian and when she entered her patrol area. She missed both. The Pacific Fleet declared her missing in mid-March 1968 and mounted a search with roughly forty surface ships, submarines, and long-range aircraft over thousands of square nautical miles. The Soviet search found nothing. K-129 was officially declared lost with all hands. The US Navy’s SOSUS arrays, cross-correlated across Naval Facility Point Sur and Adak, produced a localisation at roughly 40 deg N, 179 deg E. The US Navy Special Projects Office under Dr John P. Craven analysed the data.

USS Halibut and the AZORIAN decision

In summer 1968 USS Halibut (SSGN-587), modified for deep-ocean search, deployed to the area towing a deep-tow camera vehicle nicknamed “Fish”. On or about 20 August 1968 Halibut photographed the wreck at approximately 4,900 to 5,000 metres (about 16,000 to 16,500 feet). The hull lay in sections, with a forward portion intact enough to be of intelligence value: R-21 SLBMs with thermonuclear warheads, nuclear torpedoes, communications gear, code books, and the cryptographic keying material of a Soviet strategic submarine. The depth was twice anything any then-existing salvage operation had achieved.

In 1970 a recovery proposal was put forward by Secretary of Defense Melvin Laird and National Security Adviser Henry Kissinger, approved by President Nixon, and sustained under President Ford. The CIA’s Special Projects Staff under John Parangosky and DCI Richard Helms ran the project. The codename for the overall programme was AZORIAN; “JENNIFER” was the security compartment designation that became publicly attached to the operation by the 1975 press. Reported total cost is about US$800 million in then-year dollars, with the exact accounting varying between sources.

The Hughes Glomar Explorer cover and the August 1974 recovery

The cover story was that Howard Hughes’s Summa Corporation was developing a one-off ship to harvest manganese nodules from the abyssal Pacific, then a credible novel commercial venture. Hughes Glomar Explorer (HGE) was built by Sun Shipbuilding at Chester, Pennsylvania, with construction beginning 1 November 1972. She was about 619 ft long with a centreline moon pool and a heave-compensated pipe-string lifting system. The capture vehicle, officially “Capture Vehicle” and nicknamed “Clementine”, was fabricated by Lockheed at Redwood City, California: a multi-storey steel grappling structure with eight hydraulic claws shaped to cradle K-129’s forward section. Hughes Mining Barge One (HMB-1), a submersible barge with an opening roof, sheltered Clementine and mated it to the HGE underwater off Catalina Island, shielded from Soviet satellite observation.

The HGE arrived on station in early July 1974. The pipe string was deployed through the moon pool, Clementine was lowered approximately 16,500 ft, the claws were closed around K-129’s forward section, and the lift began. During the lift several of Clementine’s maraging-steel arms fractured. Per the declassified CIA Studies in Intelligence history and the Polmar and Sharp accounts, the intended recovery was the forward roughly 138 ft of the hull; the actually recovered portion was on the order of a 38-ft bow section. The portion that fell back included the sail, the missile silos containing two of the three R-21s, and a large fraction of the cryptographic material and crew remains. Soviet vessels including the tug SB-10 and the missile-range ship Chazhma observed the operation, were uncertain of its purpose, and did not interdict.

The recovered forward section contained, per the declassified history together with Sharp 2012 and Polmar 2010: the remains of six Soviet sailors; at least one R-21 missile and missile components; two nuclear-tipped torpedoes (Sharp and Polmar report nuclear torpedo recovery; the CIA text is redacted on precise nuclear weapons recovered); code books and cryptographic fragments described in the redacted CIA text as “materials of apparent interest”; and equipment including the ship’s bell, later returned to the Russian Federation. The Hersh New York Times headline of 19 March 1975 emphasised that the operation “failed to raise atom missiles”.

The burial at sea and the 1975 disclosure

The six recovered sailors’ remains were buried at sea in a formal ceremony on 4 September 1974 from the deck of the Hughes Glomar Explorer, conducted partly in Russian using elements of the Soviet Naval Code. CIA Chief Warrant Officer 4 (Ret.) Jim Reeb served as the Russian-language reader. The ceremony was filmed on 16 mm. The bodies were committed in a metal casket; the Project Azorian Wikipedia entry notes radiological concerns as the reason for sea burial rather than repatriation. Three of the six identified sailors are publicly reported across multiple secondary sources as Viktor Lokhov, Vladimir Kostyushko, and Valentin Nosachev. Other names appearing in some accounts are reported less consistently, and the Cold File does not assert a definitive full roster.

On or about 7 February 1975 the Los Angeles Times ran an initial story. Jack Anderson aired a more substantive radio account on 18 March 1975. On 19 March 1975 Seymour M. Hersh published in The New York Times under the headline “C.I.A. Salvage Ship Brought Up Part of Soviet Sub Lost in 1968, Failed to Raise Atom Missiles”, with an editor’s note explaining that the Times had held the story at the request of DCI William Colby for about a year. Colby and his successor George H. W. Bush declined to confirm or deny operational details.

Phillippi v. CIA and the Glomar response

Following the disclosure, the journalist Harriet Ann Phillippi filed a FOIA request with the CIA for records of the agency’s contacts with news organisations about the Glomar Explorer story. The CIA refused to confirm or deny the existence of such records, arguing that any answer would itself confirm operational information. Phillippi sued. In Phillippi v. CIA, 546 F.2d 1009 (D.C. Cir. 1976), decided 16 November 1976, the D.C. Circuit (J. Skelly Wright writing) affirmed the legal availability of the “neither confirm nor deny” posture in FOIA litigation, while remanding for the CIA to articulate its rationale on the public record. This is the canonical origin of the “Glomar response”, now used routinely by US intelligence and national security agencies and recognised by every federal circuit to have considered it.

In October 1992 CIA Director Robert Gates met Russian President Boris Yeltsin in Moscow and presented him with a copy of the burial-at-sea film and an accompanying flag. In February 2010, in response to a FOIA request from historian Matthew Aid, the CIA released a heavily redacted version of the 1985 Studies in Intelligence article “Project AZORIAN: The Story of the Hughes Glomar Explorer”. The National Security Archive at George Washington University published it with analysis.

The evidence

Acoustic. SOSUS recorded an event consistent with a submarine implosion in March 1968. The date is reported as on or about 8 March 1968 (Sontag 1998, Sharp 2012) or 11 March 1968 (per US Navy and CIA-derived material). Cross-correlation produced a localisation near 40 deg N, 179 deg E. The CIA Studies in Intelligence text is redacted on the precise time fix.
Wreck position. Two coordinate sets appear: approximately 38 deg N, 178 deg 56’ W in contemporary press, and approximately 40 deg N, 179 deg E in CIA-derived material. Open sources do not cleanly reconcile them.
Halibut photographs (August 1968). Used to identify the boat as K-129 and assess hull condition; the imagery itself remains largely classified, and the open record carries only secondary description.
Recovered forward section (1974). Approximately the forward 38 feet of the hull, against an intended 138 ft. The portion that fell back contained the sail and the silos with two of the three R-21s, plus a large fraction of the cryptographic material and the rest of the crew. The sections most likely to discriminate between a missile-propellant event and a battery event were lost back to the seabed.
CIA Studies in Intelligence article. The 2010 redacted release is the single most important documentary source in the open record. The release is heavily redacted on acoustic detail, the precise count and identity of nuclear weapons recovered, and final disposition of the recovered section.
Phillippi v. CIA, 546 F.2d 1009 (D.C. Cir. 1976). The founding statement of “neither confirm nor deny” in US FOIA practice. Not a source on the cause of loss; a source on the legal architecture the loss generated.
Russian-language scholarship. Nikolai Cherkashin’s The Death of Submarine K-129 (Veche, 2005) and Mikhail Mormul’s The Catastrophes Below (Yauza, 2008) are the principal carriers of the Soviet and Russian institutional reading, cited as carried by English-language scholarship.

The strongest single piece of evidence is acoustic. The most direct physical evidence is a single forward section recovered without the parts that would most cleanly resolve the cause. Every reading is inferred from what is partial, and the Soviet and Western sources do not agree on which inference is the strongest.

The hypotheses

Everything in this section is a hypothesis. None of it is a finding. The hypotheses are ordered by their place in the documented analytical record, not by likelihood.

A. Battery hydrogen explosion during snorkeling. The dominant US analytical reading, per Sontag, Drew and Drew (1998), Craven (2001), and Sharp (2012). For: the SOSUS signature is consistent with a single sharp explosion followed by implosion, and the damage on the recovered forward section is consistent with an internal explosion forward of the sail. Against: it cannot be definitively confirmed without the unrecovered after sections.

B. Collision with USS Swordfish (SSN-579). The dominant Russian-language reading, per Cherkashin (2005) and Mormul (2008). For: USS Swordfish put in to Yokosuka, Japan, in March 1968 with a damaged sail, and Russian analysts argue the timing, geography, and photographed damage are consistent with a submarine-to-submarine collision. Against: the US Navy has consistently and on the record denied any collision involving USS Swordfish and K-129, attributing the Swordfish sail damage to ice contact under Arctic ice earlier in 1968, and no declassified US documentary record supports a collision. The Cold File reports the contested historiographical question, attributes the collision reading strictly to Russian-language scholarship and the denial strictly to the US Navy’s institutional position, and does not adjudicate.

C. R-21 missile propellant explosion in a fin silo. For: the R-21 used UDMH and a nitric-acid-based oxidiser, a combination with a documented history of incidents in Soviet service. Against: the fin silos that would speak most directly to this reading were in the after section that fell back to the seabed.

D. Snorkel induction failure and flooding. For: snorkel-valve failure is a known loss mechanism for diesel-electric submarines and was implicated in other Soviet submarine losses of the era. Against: a slow-flooding sequence does not fit the sharp acoustic event recorded by SOSUS.

E. Combined snorkeling, battery hydrogen, and ignition sequence. Sharp 2012’s most defensible reading, in effect a combination of A and D. For: it accommodates both the acoustic signature and the damage visible on the recovered forward section. Against: it remains an inference from the same partial record that supports each component reading separately.

F. Hostile action by a US Navy unit other than a collision. A minority Russian-language reading not endorsed by the mainstream historiography on either side. For: the Cold War undersea contest is documented context. Against: this is weakly supported and not endorsed by Polmar (2010), Sharp (2012), Craven (2001), or mainstream Russian naval historiography. The Cold File notes the hypothesis as a Russian-conspiracy reading for completeness and does not adopt it.

What remains unknown

The precise cause of K-129’s loss has never been established. Whether any classified or otherwise unreleased US Navy documentary record bears on a USS Swordfish and K-129 collision is not known on the open record. The full and authoritatively sourced names of all six recovered sailors, and the basis for their identification, are not consistently established in open English-language sources. The full inventory of material recovered and of material that fell back to the seabed is not openly stated; the 1985 CIA Studies in Intelligence article is heavily redacted on both. The final disposition of the recovered forward section after intelligence exploitation is not in open sources. The Soviet Pacific Fleet’s contemporaneous internal post-loss investigation has been partially declassified by the Russian side since 1991, but the full record is not openly available.

Sources

Primary

Secondary

Norman Polmar and Michael White, Project Azorian: The CIA and the Raising of the K-129 (Naval Institute Press, 2010).
David H. Sharp, The CIA’s Greatest Covert Operation (University Press of Kansas, 2012). The author was the CIA recovery-phase director.
Sherry Sontag, Christopher Drew and Annette Lawrence Drew, Blind Man’s Bluff (PublicAffairs, 1998).
John P. Craven, The Silent War (Simon & Schuster, 2001).
US Naval Institute, “The Loss and the Mysteries of the K-129,” Naval History, August 2024
Nikolai Cherkashin, The Death of Submarine K-129 (Veche, 2005, Russian-language). Cited as carried by English-language secondary sources.
Mikhail Mormul, The Catastrophes Below (Yauza, 2008, Russian-language). Cited as carried by English-language secondary sources.